iftutt CoIIese of Agriculture Sit Cornell ®niber Experiments. Commenc- Number Aiea, Kumber of ing of years. acres. Plots. Wheat (various mannres) 1843-4 50 11 34 (or 37) Wheat, alternated with fallow 1851 43 1 2 Wheat (varieties) 1867-8 15 4-8 about 20 Barley (various manures) 1852 42 4i 29 Oats (various manures) . 1869 101 o| 6 Beans (various manures) 1847 32 2 l| 10 Beans (various manures) 1852 273 1 5 Beans (alternated with wheat) 1851 28 « 1 10 Clover (various manures) 1848-9 295 3 18 Various leguminous plants . 1878 16 3 18 Turnips (various manures) . 1843 286 8 40 Sugar-beet (various manures) 1870 5 8 41 Mangel-wurzel (various manures) . Total . Potatoes (various manures) . 1876 1876 18 8 2 41 10 51 18 Rotation (various manures) . 1848 46 3 12 Permanent grass (various manures) 1856 38 7 22 ' Including one year faEow. ' Including one year wheat, and five years fallow. " Including four years fallow. ^ Including two years fallow. ^ Clover, twelve times sown (first in 1848) ; only eight crops, four very small ; one year wheat, five years barley, twelve years faUow. ^ Including barley without manure three years, 1853-55. It is obvious that the results of field experiments with the individual crops, conducted as above described, must of them- selves throw much light on the characteristic requirements of the particular crop under investigation, whilst those of the ex- peiiments on the growth of crops in an actual course of rotation will serve to confLrm and control those obtained with the indi- vidual crops, and will in their turn receive elucidation from the results with the individual crops. Then, again, the results of the experiments on the application of different manures to the mixed herbage of grass-land — which includes, among others, members of the botanical families that contribute some of the most important of our rotation crops — may, independently of their value in reference to the special objects for which they were undertaken, be expected to afford interesting collateral evidence in regard to the requirements of individual plants when thus grown in association, instead of separately year after year, or in rotation, as in the other series of experiments. Ob- viously, too, the chemical, and in some cases the botanical. 16 THE EOTHAMSTED EXPERIMENTS. statistics of the crops so variously grown, and the chemical statistics of the soils of the plots upon which they have been grown, must afford very important data for further study and elucidation. An examination of Table I. will show that the individual crops which have been grown separately year after year on the same land include — wheat, barley, and oats, as members of the order Gramineee ; beans, clover, and other plants, of the order Leguminosse ; turnips of the Cruciferae ; sugar-beet and mangel- wurzel of the Chenopodiacese ; and potatoes of the Solanese. Then the experiments on rotation include those with members of three of the above orders — turnips of the Cruciferae, barley and wheat of the Gramineae, and clover and beans of the Legu- minosse. Lastly, there are the experiments on the mixed herbage of permanent grass -land, which includes, besides gramineous and leguminous plants, numerous species of other natural orders. The first experiments undertaken were those with root-crops, which were commenced in June 1843, so that last year (1894) was the fifty -second of their continuance. The second were those on wheat, commenced in the autumn of 1843, so that the crop of the last harvest was the fifty-first grown in succession on the same land. The experiments with beans were com- ■ menced in 1847 ; but, for reasons which will be fully explained, they have not been continued up to the present time. Those with clover were commenced in 1848, and have been succeeded on the same land by others with various leguminous plants, which are still continued. Then of the other more important series, those on barley were commenced in 1852, and are still in progress, the crop of 1894 being, therefore, the forty-third in succession. Experiments with oats were commenced in 1869, and continued for ten years. Others, on the growth of wheat alternated with fallow, but without manure, were commenced in 1851, and are still going on, 1894 being the forty-fourth year; and those on potatoes were commenced in 1876, the crop of 1894 making, therefore, the nineteenth in succession. The ex- periments on an actual course of rotation were commenced in 1848, and are still continued, so that the crop of wheat now growing will complete the twelfth course of four years, and the forty -eighth year of the experiments. Lastly, those on the mixed herbage of permanent grass - land were commenced in 1856, so that 1894 completed the thirty- ninth year of their continuance. It should be observed that the earlier field experiments were commenced without any idea of long continuance, and it was only as the results obtained indicated the importance of such continuance that the plan eventually adopted was gradually developed. It is, however, to long continuance that we owe INTEODUCTION. 17 some of the most interesting and the m.ost valuable of our results, as win be fully illustrated as we proceed. Table I. further shows the area, and the number of plots, under experiment in each case ; and it may be stated that the total area under exact and continuous experiment hias been for some years, and is at the present time, about 40 acres. The next point to consider is — What is the most appropriate selection to make among the field and other results ; and what is the most appropriate order in which to consider them, in attempting to illustrate the objects, plan, and results, of the Eothamsted investigations ? It will be readily understood that our selection of crops for investigation was largely influenced by the actual practice of our own part of the country. The separately grown individual crops were, in fact, the chief of those entering into our rotations ; whilst the rotation selected for study was the well-known " four course " — namely, roots, barley, leguminous crop (or fallow), and wheat. Obviously, therefore, the most natural order of illustration would be that indicated by the ideas and conditions in accordance with which the experiments have been arranged and conducted; and the order so indicated will, we think, be found to be, upon the whole, not only the most convenient but the most instructive. We have, it is true, in different parts of the country a great variety of soil and of climate, and accordingly great variety in crops, and in the order of their rotation. Still, it will be seen that the selection of individual crops experimented upon in- cludes most, and certainly the most typical, of those grown in the varied rotations of different parts of the country ; and it will be admitted that, in some important respects, the characteristic requirements of the individual crops are very similar whether grown in one locality or in another. Indeed, it cannot fail to be recognised that, mutatis mutandis, the results which have been obtained under given conditions at Eothamsted are not without their significance and bearing, under the different con- ditions of other localities. In accordance with what has been said, it is proposed to consider the results obtained, with the selection of the crops experimentally grown, and in the laboratory investigations con- nected with them, as given in the following list. Lastly, it will be seen that the very important complementary subject of the feeding of animals will also be considered. 1. Eoot-crops — Common turnips, Swedish turnips, sugar- beet, and mangel-wurzel ; each grown continuously. 2. Barley — grown continuously. 3. Leguminous crops — Clover, beans, and various other Le- guminosae ; mostly grown continuously. Also the question of the fixation of free nitrogen. VOL. VII. B 18 THE KOTHAMSTED EXPERIMENTS. 4. Wheat — grown continuously. 5. Eotation of crops — Eoot - crops (Swedish turnips), barley, leguminous crops (or fallow), and wheat. 6. Eesults of experiments on the feeding of animals— for the production of meat, milk, and manure, and for the exercise of force. It will be observed that Nos. 1, 2, 3, and 4, refer to the indi- vidual crops grown continuously ; and No. 5 to the same crops grown in rotation. Eeference to the list given in Table I. will show, however, that among the field experiments there enumer- ated there will still remain untouched the following : — The experiments with oats grown continuously ; Those with potatoes grown continuously ; Those on the alternation of wheat and fallow ; The very extensive series on the mixed herbage of permanent grass-land — including results as to the amounts of produce obtained, and those relating to its composition, both botani- cal and chemical. There also remains the extensive series of investigations on rainfall and drainage — their quantity and composition. It seemed, indeed, desirable that as complete a view as practi- cable within the space to be occupied should be given of the investigations selected for illustration ; leaving -the subjects which it was not possible so to include to be studied, by those who desire so to do, in the various papers relating to them which have been published elsewhere, and to which full refer- ence is given in the lists of papers which will be found in the annually issued ' Memoranda of the Origin, Plan, and Eesults of the Field, and other Experiments, conducted on the Farm and in the Laboratory,' at Eothamsted. In the same document will also be found, besides much general information in regard to the experiments, descriptive and numerical details relating not only to the experiments which will be treated of in the following pages, but also to those the consideration of which cannot be included in the present Eeport. SIR JOSEPH HENRY GILBERT, M,A„ PH.D., LL.D,, F.R.S, EOOT-CEOPS. 19 SECTION I. — EXPERIMENTS WITH ROOT -CROPS GROWN CONTINUOUSLY; BARNFIELD, ROTR- AMSTED. Introduction. The Boot-crops, the conditions of growth and the composi- Conditions tion of which we have first to consid.er, include members of ofji'^th more than one natural Order of plants ; and they are grown fr^l for, so to speak, certain intermediate parts and products, which are, by cultivation, very abnormally developed ; whilst the crops are not allowed to ripen, but are taken when in a succulent and immature condition. "We shall thus have in- teresting points of comparison, or contrast, brought out, as to the conditions of growth of these crops, and of those to which we owe ripened products, such as the cereal grains. The crops to which we shall specially direct attention are — some varieties of turnips belonging to the Order Cruciferae, and two varieties of beet, namely, the sugar-beet, and the feeding mangel, of the Order Chenopodiacege. The introduction of turnips into our rotations may be im^ortcmce said to have been one of the most important improvements "Z^"™*? of modern times. The growth of the crop constitutes in- deed an essential element, not only in the ordinary four- course rotation, but in all our varied rotations. From certain characters of the turnip plant, and of other Emt-crops root-crops, especially their abundant leaf-surface, and from ««<^«*<'- certain conditions of their growth, it has frequently been assumed that they are largely dependent on the atmosphere for their nitrogen ; and that they are in fact thus collectors of nitrogen for the crops grown in alternation with them. But we shall see that experimental evidence does not support this conclusion ; and that we must look in other directions for an explanation of the undoubted benefits of the growth of root-crops in rotation. The object to be attained in the cultivation of root-crops is Abnormal to encourage, by artificial means, a quite abnormal develop- ™°^ ^''^^ ment of a particular part of the plant. If, for example, the turnip-plant were grown for its natural seed-product — oil — a heavier soil would be more suitable than when the object is to develop the swollen root. In our climate a biennial habit would be induced, and it would be so grown as to be exposed to the summer temperature at a later stage of the life-history of the plant — that is, at the seed-forming and ripening period. Under these circumstances there would be much less of fibrous root distributed through the surface-soil, the main TOOt would be much more fusiform, tapping rather than 20 THE EOTHAMSTED EXPERIMENTS. Tv/mij^s Common practice of root Old- til/re. spreading laterally, the leaves and stem would be larger, both actually and proportionally to the root, and the enlarged root itself would serve as a store of material for the second or final growth. To obtain the cultivated root, however, as grown as a rota- tion and food crop, the conditions required are very different. The seed is sown at a different period, and the character of the manuring, and of the season of growth chosen, are in their conjoint influence such as to favour a very abnormal accumulation of the store-material in the root, and to secure that this development shall attain a maximum within the limits of the season. It will be seen, however, that the cultivated turnip very soon reverts to its more natural characteristics if the mode of treatment be not such as to favour the artificial development. The first results to be adduced relate to experiments with a variety of the common turnip, or Brassica rwpa. 1. Experimmts with Norfolk White Titrnips. Eoot-crops — whether common turnips, Swedish turnips, or mangel-wurzel — are in ordinary practice grown by the aid of large dressings of farmyard manure, with or without artificial manures in addition. The farmyard manure is in some cases applied for the preceding grain crop, but more generally directly for the root-crop itself. The following table shows the results obtained with Norfolk white turnips, both without manure, and by 12 tons of farmyard manure applied annually for three years in succession. TABLE 2.— Produce of Norfolk White Turnips. Without Tnamwe, With Roots. ■ Leaves. Seasons, Without manure. With farmyard manure. Without manure. With farmyard manure. 1843 . 1844 . 1845 . tons. cwt. 4 3f 2 4} 13| tons. cwt. 9 9A\ 10 15|j 17 Of tons. cwt. Not weighed 14i tons. cwt. Not weighed 7 7f Mean 2 Vi 12 8J Thus, the produce of this assumed restorative crop, when grown without manure, went down in the third year tO' practically nothing — only 13f cwt. per acre ; whilst in the third year with farmyard manure there was more than 17 tons. But the amount varied very much according to EOOT-CEOPS. 21 season, it being nearly twice as great in the third year as in the first. l^ow, the farmyard manure employed would contain much more of nitrogen, and also of most of the mineral constituents, than the crops grown. The fact is that, independently of the great advantage Advcm- accruing from the opportunity for cleaning the land, the *^^otcfop^ value of the root-crop in rotation is mainly to be attributed a rotation. to the large amount of farmyard manure generally applied for its growth ; to the large proportion of the constituents of the manure which remain, and become slowly available to succeeding crops ; to the large amount of the nitrogen and other constituents remaining in the leaf, which serve directly as manure again. Then they are gross feeders, so to speak, converting a large amount of manure into vegetable produce ; whilst, when the edible portion — the root — is consumed by store or fattening stock, a very small proportion of the nitrogen, and of other constituents valuable as manure, is retained by the animal; the remainder, perhaps more than 90 per cent, of the nitrogen, being voided, becoming manure again. When, however, roots are consumed for the produc- tion of milk, a much larger proportion is lost to the manure. The next table (3) shows which constituent, or class of Table 3 ear- constituents, of the complex material farmyard manure, has i'^««»««^- the most characteristic influence on the growth of the root- TABLE 3. — NoBFOLK White Turnips geown tear after year ON THE SAME Land. Eesults stowing tlie effects of exhaustion and manures, four seasons, 1845-48. Manures and produce per acre per annum. Series 4. Series 1. Series 3. Ammonium- Series 6. No nitro- Ammonium- salts and Bape-oake genous salts =45 lb. rape-calte= = 90 lb. manure. nitrogen. 136 lb. nitrogen. nitrogen. WITHOUT MINERAL MANURE (THREE TEARS ONLY, 1846-48). Roots Leaves Total tons. cwt. 1 4 17 tons. cwt. 1 7 1 tons. cwt. 5 10 3 19 tons. cwt. 6 11 3 3 9 9 9 14 WITH VARIOUS MINERAL MANURES. Eoots Leaves 8 4 2 14 9 18 4 6 10 6 5 3 11 4 12 Total 10 18 14 4 16 8 15 12 22 THE EOTHAMSTED EXPEEIMENTS. ArtificiaZ crop. It shows the average yield over four consecutive seasons, 1845-48, of roots, of leaves, and of total produce, of Norfolk white turnips, grown without manure, and with a variety of artificial manures. The upper division shows the produce without mineral manure, and the lower division the mean produce of different mineral manures — ^namely (1), superphosphate of lime (plot 5); (2) superphosphate and potash salt (plot 6) ; (3) superphosphate, and potash, soda, and magnesia salts (plot 4). The first point to notice is, that on some of the manured with rniifi- plots there is an average of about 11 tons of roots, and more ure and than 4J tons of leaves, giving of total product per acre more than 15^ tons. " Without manure," on the other hand, this assumed " restorative crop " yields an average of only 1 ton 4 cwt. of roots, 17 cwt. of leaves, and a total produce of only 2 tons 1 cwt. The character of the unmanured root was, moreover, totally different. It had more the shape of a carrot than of a turnip. Its composition was also totally different from that of the cultivated root, as is strikingly illustrated by the following figures, which relate to the crops of the third season of the experiments, 1845. Gomposi- tion of roots grown with and with- out man- iires. Produce vdthout manure. Without manure Farmyard manure . Superphosphate of lime Roots per acre. tons. cwt. 1.3f 17 1 11 2 Nitrogen per cent in dry matter. per cent. 3.31 1.56 1.52 Efect of mitrogenous ma/nitre. Thus, under the influence of manure there is a very large amount of non-nitrogenous substance accumulated, diluting, so to speak, the high percentage of nitrogen of the natural, uncultivated root. There is indeed also much more nitrogen taken up by the cultivated plant ; but in it there is, in pro- portion to the ■ nitrogen, a large amount of other matters formed, the accumulation of which converts the plant into an important food-crop. Even mineral manures alone, especially those which contain phosphates, have a very marked effect in inducing such accumulation ; and it is pre-eminently by the action of such manures that a great amount of fibrous root is developed in the surface-soil, under the influence of which more nitrogen, and at the same time more mineral matters, are taken up. The results in the other columns of Table 3 (p. 21) show that the addition of nitrogenous manure, whether as ammo- EOOT-CEOPS. 23 nium-salts, or as rape-cake, or both, gives a further increase in the produce of the roots. But the second line of each division of the table shows that a prominent effect of the nitrogenous manures is also largely to increase the produc- tion of leaf. The next Table (4) shows, first, the average proportion of UafanA leaf to 1000 of root under the four characteristically different '"'"'*• TABLE 4. — Norfolk White Tuenips. Grown year after year on the same land. Mean of plots 4, 5, 6 — four years, 1845-1848. Series 1. Mineral manure alone. Series 3. Series 4. Mineral Mineral and Series 5. and. ammonium- Mineral and ammonium- salts and rape-cake salts rape-cake = 90 lb. =47 lb. = 137 lb. nitrogen. nitrogen. nitrogen. LEAF TO 1000 ROOT. 329 434 600 418 PER :ent. Dry ( In root . matter \ In leaf . 8.54 14.56 8.07 13.54 7.66 12.43 7.96 12.94 Nitrogen ( In root in dry ( In leaf 1.60 3.75 2.64 3.68 2.45 (3.68) 1.78 (3.68) Mineral ( In root in dry ( In leaf 7.26 12.24 8.22 11.88 9.03 11.12 8.30 11.87 PER ACRE, LB. Drj- matter fin root . J In leaf . l,Leaf -1- or - root fin root . In leaf . ,Leaf-(- or -root fin root . In leaf . ,Leaf-i- or -root 1581 853 1807 1289 1770 1703 1963 1296 -728 -618 -67 -667 Nitrogen 25 32 48 48 43 63 35 48 -(-7 -f20 4-13 Mineral 118 100 148 151 160 187 165 151 matter -18 -f-3 ■ 4-27 -14 conditions as to manuring. It also shows the percentages of dry matter in the roots and in the leaves respectively, and the percentages of nitrogen and of total mineral matter (ash) in the dry matter. In the lower division of the table are 24 THE EOTHAMSTED BXPEBIMENTS. given the amounts per acre of each of these constituents, in the roots and leaves respectively, and the amounts per acre, more or less, in the leaf than in the root. Effeet of Thus, with the Norfolk white turnip we have less than one- T^Tmd^ third as much leaf as root without nitrogenous manure, but root. nearly two-thirds as much with the largest supply of nitrogen by manure — that is, with the greatest luxuriance of growth. The economic importance of the difference in the propor- tion of leaf to root, under the iniluence of different conditions as to manuring, is illustrated by the other results given in the table ; and similar results given in corresponding tables relating to Swedish turnips, sugar-beet, and mangel-wurzel, will show how great is the difference in this respect between different descriptions of root-crops. In the case of the Norfolk white turnips, not only is there a large proportion of leaf, but the leaf, contains a very much higher percentage of dry matter than the root, and there is a very much higher percentage of both nitrogen and total mineral matter in the dry substance of the leaf than in that of the root. The significance of these facts is more clearly brought out in the lower division of the table, which shows the amounts per acre, in root and in leaf respectively, of dry matter, of nitrogen, and of total mineral matter, under the different conditions of manuring; also the amounts of these in the leaf -)- or — the amounts in the roots. It is seen that there was in one case, that with the highest nitrogenous manuring, nearly as much dry or solid matter per acre in the leaf, which for the most part only becomes manure again, as in the edible part of the crop — the root. In three cases there is actually more of the nitrogen of the crop in the leaf, remaining for manure, than there is in the portion available as food. There is also, in two cases, more of total mineral constituents in the leaf than in the root. 2. Experiments with Swedish Turnips. Swedes. The experiments with the Swedish turnip — Brassica cam- pestris rutabaga — were made in the same field, on the same plots, and with to a great extent similar manures, as in the case of the Norfolk white turnips already considered. The mineral manures were in fact practically the same through- out, and the nitrogenous manures were nearly the same in the first two of the four years, 1849 and 1850, but in the second two no nitrogenous manures were used. Further, the results were obtained in the next succeeding four years to those in which the Norfolk whites were grown. KOOT-CKOPS. 25 Table 5 shows the average amounts of produce — roots, leaves, and total — under the different conditions of manuring over the four years, two with and two without nitrogenous manures. TABLE 5.— Swedish Tuenips. Results showing the effects of ex- haustion and manures, four seasons, 1849-1852. Manures and produce per acre per annum. Series 4. Series 3. Ammonium- Series 5. Series 1. Ammonium- salts and Rape-cake No nitro- salts =41 lb. rape-cake= =98 lb. genous nitrogen 139 lb. nitrogen manure. (1849 and nitrogen (1849 and 1850 only). (1849 and 1850 only). 1850 only). WITHOUT MINERAL MANURE. Boots Leaves i • tons. 2 owt. 6 6 tons. cwt. 3 17 6 tons. 7 cwc. 17 tons. 7 cwt. 14 13 Total 2 12 4 3 7 17 8 7 WITH VARIOUS MINERAL MANURES (PLOTS 4, 5, ANE 6). Roots . . . 1 7 5 Leaves! . . . ' 10 8 18 11 12 2 19 11 9 15 Total . . 1 7 15 9 9 13 1 : 12 4 1 Average of three years only, 1850-52, leaves in 1849 not weighed. Compared with the produce of the white turnip, that of the Swedish turnip shows upon the whole rather less root without nitrogenous manure — that is, with the mineral manure alone — owing to the gradual exhaustion of the nitrogen of the soil where none had been, applied by manure for a number of years. But, on the other hand, there is, with nitrogenous manures, in two cases out of three, more of the Swedish than of the white turnip root. A very important point to notice is that there was, even when there was more root, very much less leaf in the case of the Swedish turnip. Thus, whilst with the highest nitro- genous manure there was, with an average of lOJ tons of the white turnip roots, nearly 6J tons of leaves, there was with the Swedish turnip, with more than 12 tons of roots, not quite 1 ton of leaf. Here, then, the result of growth is that almost the whole of the accumulation is in the food-product, the root, and a very insignificant amount remains in the leaf, most of it simply to become manure again. This point will be more clearly illustrated by the results given in Table 6, which gives the leaf to 1000 root, and the Swedes and white compared. Produce of roots and leaves. Accwm/ula- tion in the root. Taile 6 ex- 26 THE ROTHAMSTED EXPERIMENTS. Propor- tions of leaf amd root. same particulars as before relating to the percentage compo- sition of each, and to the amounts of the selected constituents per acre in each. TABLE 6. — Swedish Turnips. Proportion of leaf to root, and selected constituents in root and leaf, per cent and per acre. Mean of plots 4, 5, and 6 ; iowc years, 1849-52. Series 1. Mineral manure alone. Series 3. Mineral and ammonium- salts =41 lb. nitrogen. Series 4. Mineral and ammonium- salts and rape-cake = 139 lb. nitrogen. Series 6. Mineral and rape-cake = 98 1b. nitrogen. LEAF TO 1000 ROOT. 69.0 61.i 78.5 65.5 PER CENT. Dry f In root , matter ( In leaf . 11.59 13.81 11.51 13.08 10.54 12.97 10.89 13.19 Nitrogen ( In root . in dry \ In leaf . 1.40 3.95 1.69 4.07 2.19 4.11 1.84 4.00 Mineral 1 t . <. matter U^ ™°' • • in dry / 1'' leaf . . 4.38 12.16 4.49 11.85 4.83 10.54 4.66 10.59 PER ACRE, LB. rIn root . Dry J In leaf . matter j iLeaf-l- or -root 1879 154 2245 166 2840 270 2769 227 -1725 -2079 -2570 -2542 f In root . Nitrogen<^I''leaf . .. 26 6 38 7 62 11 51 9 V.Leaf-1- or -root - 20 - 31 - 51 - 42 rIn root . Mineral 1 In leaf . matter 1 83 19 102 20 139 29 130 24 .Leaf-H or -root - 64 - 82 - 110 — 106 It- is seen that instead of 300 to 600 parts of leaf for 1000 of root, as in the white or common turnip, we have, with the Swedish turnip, in no case 100 of leaf to 1000 of root. The highest proportion is 78| to 1000, and this is with the highest nitrogenous manuring, and the most luxuriant crops. It is further seen that the percentage of dry matter in the root ranged from lOJ to 11|-, whilst in the white turnip it averaged only about 8 per cent. "We have, therefore, not EOOT-CEOPS. 27 only a larger proportion of edible root, but that root contains a larger proportion of solid matter or food-material. As with the Norfolk white, however, so also with the Composi- Swedish turnip, the leaf contains a much higher percentage *""i °f , of dry substance than the root, and the dry substance of the Uaves of leaf contains a much higher percentage of both nitrogen and ^"Jf^ .. total mineral matter than does the dry substance of the root, lurrdps. ^ The lower division of the table shows, when compared with the corresponding particulars relating to the Norfolk white turnip, that with the Swedish turnip there was, with the highest manuring, fully one and a-half time as much dry substance per acre in the root — that is, one and a-half time as much food produced per acre as with the common turnip. Further, there is a quite insignificant amount of matter accumulated and remaining in the leaf, for the most part only serving as manure again. Of the nitrogen, again, there is, under all conditions of manuring, even those giving the greatest luxuriance, a very small proportion remaining in the leaf. The same is the case with the total mineral matter. The question obviously suggests itself. If the Swedish tur- Supenm-ity nip has all these advantages over the numerous varieties of °f^^^- the so-called common turnip, why are these ever grown ? why not always the Swedish turnip ? In the first place, soil and season have to be taken into why other account. Then the economy of the farm requires that de- ^"™^. scriptions should be selected that can not only be sown in ' due succession, but which will mature at different periods, so as to supply food for stock in due succession, and also fre- quently to get the crop early off the land, to leave it' free for some other crop. Again, a comparatively large proportion of leaf serves as protection against frost while the crop is still in the field ; and the storing qualities of the root have to be considered in connection with the character of the seasons of the locality. For example, on the light soils of Norfolk, which are very favourable for the development of root, and but little for that of leaf, and where the roots can be largely consumed by sheep on the land without injury to its me- chanical condition, the Swedish turnip is the predominant root. In the north-east and east of Scotland, on the other hand, several varieties of yellow common turnips are grown in much larger proportion, and a large amount of leaf is not recognised as a disadvantage. And here it may be observed ProducHon that, the higher the nitrogenous manuring, and the heavier "^^oJT" the soil, the greater is the tendency to produce a large amount leaf. of leaf. Further, as a rule the larger the amount of leaf re- maining vigorous at the time the crop is taken up, the less 28 THE EOTHAMSTED EXPERIMENTS. Aaawrmla- tionfrom rape-cake. Fwrther trials with fully ripe will be the roots ; and within limits it is desirable, with a view to the storing qualities of the root, that it should not be too ripe. After the four crops of Swedish turnips had been taken from the land, barley was grown for three years in succession without any manure, in order as far as possible to equalise the condition of the various plots, as affected by the previous manuring. It will suffice to say that the results clearly showed that there had been accumulation where rape-cake had been applied. Then for five years in succession (1856-60) Swedish turnips were again grown on the comparatively exhausted plojs, much on the same plan as before, but with smaller amounts of nitrogen supplied. No special interest attaches to the results over these five years for our present purpose. Table 7 shows the average produce per acre over the next ten years, 1861-70, again with Swedish turnips. During this period larger quantities of nitrogen were again applied, but for mineral manure superphosphate of lime was used alone — that is, without any further addition of either potash, soda, or magnesia. TABLE v. — Swedish Turnips. Results showing the effects of ex- haustion and manures. Mean of ten seasons, 1861-70. Manures and produce per acre per annum. Series 1. No nitro- genous manure. Series 2. Sodium nitrate =82 lb. nitrogen. Amimonium- salts= 82 lb. nitrogen. Series 4. Ammonium- salts and rape-cake = 180 lb. nitrogen. Series 5. Rape-cake = 98 lb. nitrogen. WITHOUT MINEBAL MANURE. tons. cwt. tons. owt. tons. cwt. tons. owt. tons. cwt. Roots . 11 1 1 13 4 9 4 15 Leaves . 3 5 3 1 18 Total . 14 1 6 16 5 9 5 13 WITH SUPERPHOSPHATE OF LIME (PLOTS 4, 6, AND S). Roots . Leaves . 2 9 9 5 S 1 4 9 17 7 9 1 14 6 8 1 3 Total . 2 18 6 8 5 6 9 3 7 11 Formeir The results of these experiments are little more than con- re«ife cm- firmatory of those which have gone before, but the amounts of produce are throughout on a lower level. This can only in part be attributed to the exclusion of potash from the man- ures. It is doubtless mainly due to the incidental circum- stance that in growing the same description of crop, with the KOOT-CKOPS. 29 same comparatively limited and superficial root-range, for so Reduction many years in succession, the surface-soil became less easily '^P^''^f<^ worked, and the tilth, so important for turnips, was frequently lontinums unsatisfactory ; whilst for want of variety and depth of root- "o*-""^- range of the crop a somewhat impervious pan was formed below. The fact is, however, of itself of considerable interest, as indicating one important and very beneficial influence of a rotation of crops. Indeed, we shall presently see that even the change to another description of root-crop, with a totally different and much more extended root-range, is accompanied with a much increased production over a given area by the use of the same manures. Looking to the Table (7), it is seen that there are now five mtrau of series of plots instead of only four, nitrate of soda being »<"^«»<^ appiied on beries 2, m amount supplying the same quantity ium-saits of nitrogen as in the ammonium-salts on Series 3. The "^V'l'^- result is a greater produce of both root and leaf than with the ammonium-salts. The superphosphate alone (see lower division of column 1) Superphos- gives much less produce than the mineral manures in the series P^^^- of four years before considered, doubtless to a great extent owing to the still further exhaustion of the available nitrogen of the surface-soil. In fact the surfaee-soUs in question showed, on analysis, lower percentages of nitrogen than those of any other experimental field at Eothamsted — a result which is quite consistent with the fact of the large amount of root distributed through the surface-soil by the growing turnip. Again, consistently with this supposition, and with the mtrogen- results that have gone before, there is still very marked but ^J"'^' somewhat reduced effect from all the nitrogenous manures ; and again, the amount of leaf is very small, but it is the greater the higher the nitrogenous manuring, and the greater the luxuriance of growth. Table 8 shows the proportion of leaf to 1000 of root ; also Table 8 ex- the percentages of dry matter, and of nitrogen and mineral p^"'^'^- matter in the dry matter ; and, as before, the amounts of each per acre, in the roots and in the leaves. With the SOU gradually becoming closer, and less favour- Propor- able for root-development, the proportion of leaf to root is *^°(^ somewhat higher. root. It should be explained that the percentages given in par- enthesis are not the results of direct determinations in each particular case, but are deduced from comparable results. They are, however, undoubtedly near enough to the truth for the purpose of the present illustrations. 30 THE EOTHAMSTED EXPERIMENTS, Again, we see much higher percentage of dry substance in SC'^ *^e leaf t^a^ i^ *^e root; also much higher percentages of nitrogen, and of total mineral matter, in the dry substance of the leaf. TABLE 8. — Swedish Ttjenips. Means of plots 4, 5, and 6 ; ten years, 1861-1870. Series 1. Mineral manure alone. Series 2. Mineral and sodium nitrate = 82 lb. nitrogen. Scries 8. Mineral and ammoni- lun-salts =82 lb. nitrogen. Series 4. Mineral and ammoni- um-salts and rape-cake = 180 lb. nitrogen. Series S. Mineral and rape-cake = 98 lb. nitrogen. LEAP TO 1000 BOOT. 18i 185 191 228 180 PER CENT. Dry matter Nitrogen in dry Mineral mat- ter in dry ( In root \ In leaf ( In root ( In leaf ( In root \ In leaf 12.04 14.93 (1.40) (3.95) 4.55 11.64 11.01 14.46 (1.69) (4.07) 6.38 10.62 11.32 14.24 (1.69) (4.07) 4.71 12.23 10.94 13.78 (2.19) (4,11) 5.10 11.54 10.83 14.66 (1.84) (4.00) 5.03 11.27 PEE AOEE, LB. Dry matter /In root . ) In leaf . (,Leaf-|- or -root { In root . J In leaf . (,Leaf-|- or -root / In root . ) In leaf . (.Leaf-H or -root 629 146 1285 320 1084 268 1777 498 1511 376 -483 -965 -816 -1279 -1135 Nitrogen . 8.8 5.8 21.7 13.0 18.3 10.9 38.9 20.5 27.8 15.1 -3.0 -8.7 -7.4 -18.4 -12.7 Mineral mat- ter 28.9 ,16.8 71.1 33.1 53.6 32.5 94.2 57.5 76.6 41.9 -12.1 -38.0 -21.1 -36.7 -34.7 Looking to the lower division of the table, it is seen that there is here again, under all conditions of manuring, much more solid matter per acre in the root than, in the leaf. There is also more nitrogen, and more total mineral matter, accumulated in the root; though the proportion of the nitrogen which is accumulated in the leaf is higher than in the previous experiments. EOOT-CEOPS. 31 3. Experiments with Sugar-beet. To the Order Chenopodiaceae, and to the species Seta Sugar-leet. wigaris, we owe many varieties of sugar-beet, and also many- varieties of feeding-beet or mangel-wurzel. Mangel-wurzel is a very important agricultural crop in some localities of our own country, whilst sugar-beet is not. Trials have, however, been made on the growth of sugar-beet for the production of sugar; and as we have experimented on the subject, we wUl in the first place illustrate the influence of various manures on the growth of the crop, and on the pro- duction of sugar in it ; and afterwards, in more detail, give somewhat similar results relating to the mangel. The experiments with both crops were made in the same field and on the same plots as those on which first Norfolk whites and afterwards Swedish turnips had been grown. The last crop of Swedish turnips was taken in 1870, and sugar-beet then followed for five years in succession, 1871-75 inclusive. Experiments with the mangel were then com- menced in 1876, and have been continued up to the present time, so that the crop of 1894 was the nineteenth in suc- cession. It has been stated that by the continuous growth of the one description of crop, the Swedish turnip, with one character and limited range of roots, the surface-soil had become close, and a somewhat impervious pan was formed below it. Therefore before growing sugar-beet the land was ploughed more deeply. During the first three of the five years of sugar-beet, the Plan qfex- arrangement of the plots and of the manures was substan- P^""'^*- tially the same as afterwards for mangels ; but during the last two years of the five, neither farmyard nor any other nitrogenous manure was applied, the object being to deter- mine the effects of the unexhausted residue of the nitrogen- ous applications during the preceding three years. Sugar-beet has a very much more deeply penetrating root character- than the turnip, and more even than the feeding-beet or i^pe growth mangel. In fact, great command of the resources of the heo!^'^' soil and subsoil is a characteristic of the cultivated plant. The root found to give the highest percentage of sugar is very characteristically fusiform ; and by careful selection of plants from which to grow seed, varieties are obtained nearly the whole of the swollen root of which forms under the surface of the soil — the percentage of sugar being much lower in the above-ground portion exposed to light. To such perfection has the art of selection, cultivation, and ac- climatisation reached, that some descriptions, when grown 32 THE KOTHAMSTED EXPEEIMENTS. Produce from dvmg alone and from, drnig and other Table 9 ex- Artificial manvres. Produce from min- eral man- ures alone and with addition of in suitable soils and localities, will yield nearly, and some- times quite, 20 per cent of sugar! For brevity, and as such heavy manuring is not adopted for the growth of beet for the manufacture of sugar, the results obtained with farmyard manure will not be given in any detail. It may, however, be observed that over the three years of the application, the average produce per acre of roots of farmyard manure alone was about 16 tons, which was raised to nearly 24 tons by the annual addition of 86 lb. of nitrogen per acre as nitrate of soda ; to about 22 tons by the same quantity of nitrogen as ammonium-salts ; to nearly 25 tons by 98 lb. of nitrogen as rape-cake ; and to more than 25 tons by 184 lb. as rape-cake and ammonium-salts together. These facts are sufficient to show how powerful a feeder and grower is the sugar-beet when liberally manured ; and that, provided other supplies are not deficient, nitrogenous man- ures very greatly increase the produce. The following Table (9, p. SS) shows the average produce of sugar-beet ; in detail roots only, and in the summary roots and leaves, over the three years, the two years, and the five years, under three conditions of mineral manuring, each alone, and each cross-dressed as indicated, by various nitro- genous manures. The table shows that when superphosphate was used either without nitrogenous manure or with nitrate of soda, the pro- duce was as great as when potash was applied in addition ; but when the nitrogen was applied as ammonium-salts, am- monium-salts and rape-cake, or rape-cake, the addition of potash to the superphosphate shows more effect. And it will be seen further on, that in the case of the mangels in subse- quent years, the effect of the potash was very much more marked — that is, when under the continuous use of super- phosphate without potash, the potash of the soil had doubt- less become more and more exhausted. That the deficiency of produce is much less marked where the superphosphate is applied with nitrate of soda than where with ammonium- salts or rape-cake, is probably due to the roots of the plant penetrating more deeply under the influence of the more soluble and more rapidly distributed nitrate with its more readily available nitrogen — thus securing a better command of the supplies of potash (and other constituents) in the lower layers of the soil and subsoil. Turning to the summary at the foot of the table, which gives the average results over the three years for plots 6 and 4 (with potash supply) both without and with nitrogenous manures, it is seen that whilst the mineral manures alone give an average of less than 6 tons of roots, the addition of EOOT-CEOPS. 33 TABLE 9. — Su&ak-Bebt. Eesults showing the effects of exhaustion and manures. Manures and produce per acre per annum. standard manures. Series 1. Standard manures only. Plot. Series 2. Sodium nitrate = 861b. nitrogen. Series 3. Ammoni- um-salts = 86 lb. nitrogen. Series 4. Ammoni- um-salts and rape- cake = 1841b. nitrogen. Series 6. Bape-oake = 98 lb. nitrogen. MEAN OP 3 TEARS, 1871-73, WITH NITROGENOUS MANURES (ROOTS ONLY). tons. cwt. tons. cwt. tons. cwt. tons. cwt. tons. cwt. 5 Superphosphate Superphosphate and po- 1 tassium sulphate j 5 18 19 11 13 9 17 15 16 5 M 5 6 17 19 11 16 22 3 17 4 ( Superphosphate, potassi- \ '{ um, and magnesium ( sulphates, and sodium f chloride ; 6 9 19 15 15 3 22 2 18 9 MEAN OF 2 YEARS, 1874 & 1875, WITHOUT NITROGENOUS MANURES (ROOTS ONLY). Superphosphate Superphosphate and po- 1 tassium sulphate ) Superphosphate, potassi- •\ um, and magnesium I sulphates, and sodium r chloride ) 5 15 8 15 7 11 10 16 5 8 8 3 7 11 10 19 5 19 9 2 7 13 11 13 8 9 8 17 9 3 MEAN OF 6 TEARS, 1871-75 (ROOTS ONLY) 5 Superphosphate Superphosphate and po- \ tassium sulphate ( Superphosphate, potassi- \ um, and magnesium ( sulphates, and sodium r chloride ) 5 17 15 4 11 2 14 19 13 3 M '{ 5 7 6 5 14 1 15 10 11 19 12 3 17 14 17 18 13 17 14 14 SUMMARY— MEAN OF PLOTS 6 & 4 (ROOTS AND LEAVES). Mean of 3 years, 1871-73 Eoots Leaves Total / Eoots Mean of 2 years, J Leaves 1874 and 1875 ) (. Total I' Eoots Mean of 5 years, J Leaves 1871-75 ^ { Total 14 3 6 17 16 6 18 5 23 19 8 13 2 2 10 15 14 3 18 13 14 3 19 10 18 9 12 10 12 2 14 15 22 7 29 19 11 6 3 6 14 12 17 6 23 16 17 17 3 13 21 10 9 11 14 3 17 8 VOL. VIL 34 THE EOTHAMSTED EXPEKIMENTS. nitrate of soda raises the produce to nearly 19 tons, that of ammonium - salts to nearly 15 tons, that of rape -cake to nearly 18 tons, and that of rape-cake and ammonium-salts together to more than 22 tons. It is also seen that during the succeeding two years, when no further nitrogenous manure was used, there was still more or less increase, due partly to the manure-residue of the previous applications, and partly to the increased amount of leaf that had been annually returned to the land as manure where nitrogenous manures had been employed. Thus the average produce over the two years by the mineral manures, including potash, but without nitrogenous manure, was 5 tons 14 cwt., raised where nitrate of soda had previously been applied to 8 tons 13 cwt., where ammonium - salts had been used to 7 tons 12 cwt., where rape-cake to 9 tons, and where rape-cake and ammonium- salts together to 11 tons 6 cwt. Produce The summary further shows that over the three years of of leaf. |.jjg application of nitrogenous manures, the produce of leaf was raised from 1 ton 7 cwt. with the mineral manures alone, to 5 tons 2 cwt. by the addition of sodium nitrate, to 3 tons 10 cwt. by ammonium-salts, to 3 tons 13 cwt. by rape-cake, and to 7 tons 16 cwt. by rape-cake and ammonium-salts to- gether. Over the next two years, without further nitro- genous manuring, but with some nitrogenous manure -residue, and increased return of leaf to the land, where nitrogenous manures had been applied, the produce of leaf was raised from 1 ton 2 cwt. by the mineral manure alone, to 2 tons 2 cwt. where in addition nitrate of soda had previously been applied, to 1 ton 10 cwt. where ammonium-salts had been used, to 2 tons 8 cwt. where rape-cake, and to 3 tons 6 cwt. where rape - cake and ammonium - salts had been applied together. TaiieW The next Table (10, p. 35) which relates to the mean pro- duce of plots 6 and 4 (with potash), over the three years during which the nitrogenous manures were annually applied, shows the proportion of leaf to 1000 of root, some particulars of the percentage composition of the root, and of the leaf, and the amounts of certain constituents per acre in the root and in the leaf. Proper- The first line of figures shows a range of from 205 to 354 te^a^cZ P*^*^ °^ ^^^^ *o 1000 of root, according to the manure, and root. the consequent degree of luxuriance and of maturity. The proportion of leaf was thus much higher than in Swedish turnips ; it is also higher than in mangel-wurzel, but much lower than in common turnips. ^Sfieaf T^® percentage of dry matter in the root is more than and root, twice as high as in common turnips, more than one and a-half EOOT-CEOPS. 35 TABLE 10.— Sugar-Beet. Mean of plots 6 and 4 ; 3 years, 1871-73. Series 1. Mineral manure alone. Series 4. Series 2, Series 3. Mineral Mineral Mineial and and and ammoni- sodium. ammoni- um-salts nitrate um-salts and =8611). =86 lb. rape-cake nitrogen. nitrogen. = 184 lb. nitrogen. Series 5. Mineral and rape-calce = 98 lb. nitrogen. LEAP TO 1000 BOOT. 230 269 232 354 205 PER CENT. Dry matter {!--* ; Nitrogen in / In root . dry \In leaf . Mineral mat- / In root . ter in dry \ In leaf . Potash in dry {!-[-* ; Phosphoric / In acid in dry \ In root leaf 18.75 14.65 16.83 11.19 18.16 12.12 17.04 10.20 0.58 2.18 0.95 2.61 0.84 2.30 1.27 2.76 4.11 23.83 5.13 22.13 4.75 23.47 5.59 22.08 1.45 5.29 1.67 4.52 1.72 4.82 1.84 4.68 0.67 0.78 0.55 0.67 0.52 0.64 0.57 0.62 17.88 11.28 0.82 2.34 4.54 22.86 1.61 5.21 0.56 0.81 PEE ACEB, LB. ("In root . Dry matter V''^'^^ ' ' 2463 435 6996 1248 6086 934 8444 1768 7096 925 vLeaf + or - root -2028 -5748 -6152 -6676 -6171 rln root . Nitrogen . J ^^ ^^^^ • " 14.3 9.5 67.0 32.8 51.2 21.5 105.5 48.8 58.4 21.6 1 (.Leaf + or - root -4.8 -34.2 -29.7 -66.7 -36.8 rln root . Mineral mat- 1 In leaf . ter 1 (.Leaf-t- or -root 101.2 103.7 364.2 276.9 288.5 217.9 469.6 390.0 322.1 210.2 -1-2.5 -87.3 -70.6 -79.6 -111.9 ("In root . Potash . -I^l^^f • • .35.6 23.0 117.1 66.4 104.4 45.0 156.1 81.0 113.9 48.2 .Leat-h or -root -12.6 -60.7 -59.4 -74.1 -65.7 / In root . Phosphoric ) In leaf . acid ) (.Leaf-l- or -root 14.1 3.4 38.8 8.3 31.5 6.0 48.3 11.0 39.4 7.6 -10.7 -30.6 -25.5 -37.3 -31.9 36 THE KOTHAMSTED EXPERIMENTS. time as high as in swedes, and considerably higher than in the feeding-beet or mangel-wurzel. It will afterwards be seen that this increased amount of solid matter in the root is chiefly sugar. As in the case of the mangel leaf, the percentage of dry matter in the sugar-beet leaf is actually lower than in the case of the turnips ; and it is very much lower than in the sugar-beet root, whilst in the turnip it was very much higher in the leaf than in the root. The percentage of nitrogen in the dry substance of the root is much lower than in the case of the turnip ; and it is in a less degree lower than in the mangel-root grown by the same manures. As in the case of the other descriptions of roots, the percentage of nitrogen in the dry matter of the sugar-beet leaf is very much higher than in that of the root. The percentage of mineral matter in the dry substance of the leaf is four or five times as high as that in the root ; in fact the mineral matter constitutes more than one-fifth of the total dry substance of the leaf. It is higher than in the case of the mangels, and about twice as high as in that of either Swedish or common turnips. To determine the amounts of potash and phosphoric acid in the root and in the leaf, respectively, of both sugar-beet and mangel-wurzel a large series of analyses of the ashes of the root and of the leaf of the experimentally grown sugar- beet and mangel-wurzel, has been made. Table 10 (p. 35) shows that the percentage of potash in the dry matter of the sugar-beet leaf is very much higher than in that of the root. Of phosphoric acid, on the other hand, the percentage in the dry matter of the leaf is but little higher than in that of the root; whilst in the dry matter of both root and leaf it is very much lower than is that of potash. Effect of The lower division of the table shows that, notwithstand- ta/Zna"" ing the comparatively large proportion of fresh leaf to root, root. the proportion of the total solid matter of the crop which is accumulated and remains in the leaf is, owing to the very high percentage of solid matter in the root and very much lower percentage in the leaf, much less than would be con- cluded from the weight of the fresh produce only. Thus, with the lowest proportion of leaf, as in Series 5 with rape- cake, there was more than 3 tons per acre of solid matter in the root, and much less than half a ton in the leaf ; whilst with the highest nitrogenous manuring, the greatest luxuri- ance, the heaviest crops, and the highest proportion of leaf to root, as in Series 4 with rape-cake and ammonium-salts together, there are more than 3f tons of solid matter per acre in the root, and little more than | ton in the leaf. It KOOT-CROPS. 37 will be seen further on how large a proportion of the solid matter of the root of this highly artificial vegetable produce is sugar. The lower division of the table further shows that, whilst there was only 14.3 lb. of nitrogen per acre in the roots with- out nitrogenous supply, the amount was raised — by nitrate of soda to 67 lb., by ammonium-salts to 51.2 lb., by rape-cake to 58.4 lb., and by rape-cake and ammonium-salts together to 105.5 lb. Then the amount of nitrogen per acre in the leaf was — with mineral but without nitrogenous manure 9.5 lb., with the addition of nitrate of soda 32.8 lb., of ammonium- salts 21.5 lb., of rape-cake 21.6 lb., and of rape-cake and ammonium-salts together 48.8 lb. A point of interest in regard to the amounts of nitrogen per acre in the crops is, however, that there was in every case very much more accumulated in the root than in the leaf, which is chiefly of value only as manure again. It is further seen that with the same mineral, but varying nitrogenous supply, the amount of total mineral matter per acre in the roots was — only 101.2 lb. without nitrogen supply, 364.2 lb. with nitrate of soda, 288.5 lb. with ammonium-salts, 322.1 lb. with rape-cake, and 469.6 lb., or more than 4 cwt., with the rape-cake and ammonium-salts to- gether. Lastly, the total amount of mineral matter per acre in the leaf was, with the very high percentage in the dry substance, very large ; but it was in each case, with nitro- genous supply, considerably less in the leaf than in the root. It is remarkable that with the same mineral supply in each case there was, without nitrogen, less than 2 cwt. of mineral matter per acre per annum in root and leaf together, whilst with the highest nitrogenous supply in addition there was nearly 7f cwt. of mineral matter in the total crop. There is here evidence both of how liberal must be the supply of available mineral constituents for the luxuriant growth of the crop, and how great will be the exhaustion of them if the crop be sold off the farm. Bearing in mind that the same amount of potash was applied Nitrogen per acre in the case of each of the five series, it is of interest "i^^f^f^ 1 11 fl I'll! '^^^ suga/r- to observe that the percentage oi potash in the dry substance production. of the root was distinctly higher in the four series with nitrogenous supply than in Series 1 without it ; and when we consider, as will be fully illustrated further on, that the amount of sugar produced depends very materially on the amount of nitrogen taken up, and that a liberal supply of available potash has also much influence on the amount of sugar produced, it is what might be expected that, with liberal nitrogen-supply and increased production of sugar, we 38 THE EOTHAMSTED EXPERIMENTS. should find an increased amount of potash taken up. In fact, the lower division of the table shows that, with the same potash supply by manure, there was, compared with the amount stored in the root without nitrogenous supply, more than three times as much where nitrate of soda was added, nearly three times as much where ammonium-salts were used, about three times as much where rape-cake was employed, and nearly four and a-half times as much where rape-cake and ammonium-salts were applied together, supply- ing an excessive amount of nitrogen. The actual amounts of potash per acre in the roots were indeed — only 35.6 lb. per acre per annum without nitrogenous supply, 117.1 lb. with nitrate of soda, 104.4 lb. with ammonium-salts, 113.9 lb. with rape-cake, and 155.1 lb. with the excessive supply of nitrogen in ammonium-salts and rape-cake together. Although, as has been seen, the percentage of potash was very much higher in the dry substance of the sugar-beet leaf than in that of the root, the figures in the lower division of the table show that under all conditions as to nitrogenous supply there was much less potash per acre in the leaf than in the root. As, however, the leaf would be returned to the land as manure, there should be no loss of the potash of the farm by the amount of it left in the leaf. And again, as the very much larger amount of potash in the roots should, when consumed on the farm, be almost wholly recovered in the manure of the animals fed upon them, there should be but little loss to the farm of the potash they contained. If, how- ever, either the roots or the leaves are removed or sold off the farm, the exhaustion of potash may be very considerable. Phosphoric Turning to the amounts of phosphoric acid, the supply of mdUay:''^ which was the same for each of the five series, it has been seen that the percentage of it in the dry substance of the roots varied comparatively little ; but the figures in the lower division of the table show that the actual quantities per acre in the roots varied very considerably, and to a great extent in proportion to the amounts of growth as influenced by the nitrogenous supply. It is further seen that the amounts of phosphoric acid remaining in the leaf are very small com- pared with those in the root. Produce It has already been shown when considering the results mZ.^nf recorded in Table 9 (p. 33) relating to the selected artificially- and, red- manured plots, that the produce over the two years after the due-action, cessation of the application of the nitrogenous manures in- dicated considerable increase over that where no nitrogen had been applied, due partly to the residue of the nitrogenous manures previously applied, and partly to the residue (leaves, &c.) of the larger crops previously grown. It will be of in- terest here to show the average produce of roots per acre per EOOT-CEOPS. 39 annum on the different divisions of the farmyard manure plot over the three years of the direct application of the manures, and over the succeeding two years of manure- and crop-residue. It was as follows : — TABLE 11. Series 1. Farmyard manure alone (3 years only). Farmyard manure, and— Series 2. Sodium nitrate =86 lb. nitrogen (3 years only). Series 8. Ammoni- um-salts = 86 lb. nitrogen (3 years only). Series 4. Ammoni- um-salts and rape- cake = 184 lb. nitrogen (3 years only). Series 5. Rape- cake = 98 lb. nitrogen (3 years only). 3 years of direct appli-\ cation / 2 years of residue of\ manure and crop J tons. cwt. 16 6 14 tons. cwt. 23 16 15 16 tons. cwt. 22 6 16 3 tons. cwt. 25 2 17 17 tons. cwt. 24 18 17 2 Difference 2 6 8 6 3 7 5 7 16 Thus there was an average of little more than 2J tons of roots per acre per annum less over the two years of unex- hausted residue of the farmyard manure than over the three years of its direct application. There was also less leaf over the two years of residue. It is seen, however, that on the divisions of the farmyard-manure plot, where artificial nitro- genous manures were used in addition, there was an average of from 7 to 8 tons of roots less over the two years of residue than previously. There was also considerable reduction in the produce of leaf. Still the greater produce over the two years of residue-action, where the nitrogenous manures had been previously used in addition than where the farmyard manure had been used alone, show considerable effect from the residue either of the artificial nitrogenous manures them- selves, or from their increased crop-residue ; and so far as there is any direct effect from the manure-residue of the pre- viously applied nitrate or ammonium-salts, it is probably chiefly due to nitrates being drawn up again from the sub- soil. Even in the case of the rape-cake, the residue-effect is also doubtless largely due to crop-residue, but to a consider- able degree to manure-residue also — a portion of the nitro- genous matter of such organic manures becoming very slowly available in the soil. To sum up on this point : In the case of the nitrate and Mamtre- ammonium-salts, the effect of residue will be in the least pro- anf^op- portion due to manure-residue, and in the greatest to crop- residue. residue. With such manures as rape-cake, the effect will be due in a large proportion to manure-residue, and also largely to 40 THE EOTHAMSTED EXPERIMENTS, Tabu 12 Leafa/nd root. crop-residue. With farmyard manure, so far as there had been larger crops, there will be much crop-residue ; but a very large proportion of the effect on future crops is to be attributed to slowly decomposing manure- residue. The next Table (12) shows for the produce of the two years without further application of nitrogenous manures, the same particulars as to composition as Table 10 for the preceding three years — namely, the amount of leaf to 1000 root, and the percentages, and the amounts per acre, of certain constituents the root and in the leaf. The results need not be con- m sidered in much detail. Excepting in the case of Series 5, the proportion of leaf to root is considerably less over the two years, with the less supply of nitrogen within the soil, and the consequent much less luxuriance. There is, nevertheless, over the two years a lower percentage of dry substance in the root, doubtless owing to the less formation of sugar with the less nitrogen available to the plant. There is also generally a somewhat lower percentage of dry or solid substance in the leaf over the two years of comparative exhaustion. Again, there is, where nitrogenous manures had previously been applied, generally a lower, and in some cases a considerably lower, percentage of nitrogen in the dry substance of the roots over the two years of only residual supply. The percentage of nitrogen in the dry substance of the roots is indeed very low over both periods, but especially in the second; and it will be seen further on that it is much lower than in either of the descrip- tions of roots cultivated for feeding purposes. In fact, so much is the sugar-forming habit of the plant developed, and so largely does the amount of the non-nitrogenous substance — sugar — contribute to the percentage of dry matter, that the percentage of the nitrogenous bodies is relatively very low, even though a large amount of nitrogen may have been taken up over a given area. As in the case of the three years with direct nitrogenous manures, so now over the two years with only residual supply of nitrogen, the percentage of nitrogen in the dry substance of the leaf is very much higher than in that of the root. It is, however, in each series somewhat higher over the two years than over the three of direct sup- ply, perhaps owing to somewhat less matured — that is less exhausted — condition of the leaves over the two years. Turning now to the percentage of total mineral matter in the dry substance over the two years, it is seen that in the root and leaf respectively it is approximately the same over the two years as over the preceding three ; and it is as was the case over the three years, four or five times as high in the dry substance of the leaf as in that of the root. KOOT-CROPS. 41 TABLE 12.— Sogar-Beet. Mean of plots 6 and 4 ; 2 years, 18V4-75. The mineral manures, every year, and— Series 1. (No nitro- genous manure). Series 2. (Previ- ously sodium- nitrate). (Previ- ously ammoni- um-salts). Series 4. (Previ- ously ammoni- um-salts aod rape-cake). Series 5. (Previ- ously rape- cake). LEAF TO 1000 ROOT. 206 248 197 294 PER CENT. Dry matter Nitrogen in dry Mineral mat- ter in dry J* In root \ In leaf / In root \In leaf. { Potash in dry -{ In root In leaf Phosphoric acid in dry In root In leaf fin root \ In leaf 17.77 11.21 15.71 10.18 16.67 11.41 16.31 10.45 0.66 2.47 0.71 2.65 0.84 2.61 0.87 2.85 4.27 . 22.05 5.15 22.64 4.94 21.30 5.37 21.01 1.56 5.37 1.91 4.99 1.86 4.31 1.81 4.46 0.54 0.81 0.49 0.71 0.55 0.75 0.61 0.76 16.01 10.24 0.80 2.74 5.41 22.14 1.79 5.08 0.58 0.77 PER ACRE, LB. fin root . Dry matter J ^"^ ^^^^ ■ • 2259 296 3026 493 2843 385 4138 790 3232 557 l.Leaf+ or -root -1963 -2533 -2458 -3348 -2675 flu root . Nitrogen . J I" ^^^^ • " 14.5 7.2 22.6 13.0 23.2 10.1 35.7 23.1 26.4 15.4 1 I.Leaf + or - root -7.3 -9.6 -13.1 T-12.6 -11.0 rln root . Mineral mat- J In leaf . 95.8 64.7 154.6 110.4 140.5 79.9 218.8 163.1 171.0 119.2 ter 1 l.Leaf+ or -root -31.1 -44.2 -60.6 -55.7 51.8 ("In root . Potash . ]I"1«^^ ■ • 35.3 15.9 57.7 24.6 52.9 16.6 75.1 35.2 57.8 28.3 l.Leaf+ or -root -19.4 -33.1 -36.3 -39.9 -29.5 rln root . Phosphoric J In leaf . 12.3 2.4 14.9 3.5 15.7 2.9 25.2 6.0 18.9 4.3 acid 1 LLeaf+ or -root -9.9 -11.4 -12.8 -19.2 -14.6 42 THE EOTHAMSTED EXPEEIMENTS. al matter in the root. of soil i Potash m the root. Eeferring to the results given in the lower division of the Table (12) relating to the amounts per acre of dry matter, nitrogen and total mineral matter, it is seen that, comparing the other series with Series 1, there is a considerable increase in the amount of dry substance per acre in the root, and some in the leaf also, due to nitrogenous residue. There is, more- over, notable increase in the amount of nitrogen stored up in both the root and the leaf over a given area, due to residue ; but much less than there was under the influence of direct supply. Comparing the average annual amounts of dry substance, of nitrogen, and of mineral matter, per acre, over the two years of the action of residue with those over the three years of direct supply, there is in each of the Series 2, 3, 4, and 5, less than half as much dry matter per acre in the roots over the two as over the three years. There is about or less than half, and even only one-third, as much nitrogen accumulated in the roots over the two years ; and there is also generally less than half as much increase of nitrogen in the leaves over the two years. Further,. though the supply was the same each year, there was less than half as much total mineral matter in the roots, and generally less than half as much in the leaves, under the influence of the re- stricted supply of nitrogen and coincident restricted growth. In reference to these points, it is to be borne in mind that the leaves were always returned to the land. Whilst there is in the above facts clear evidence of con^^ siderable effect from previously unexhausted nitrogenous manure and crop-residue, there is at the same time in the lower percentage of nitrogen in the roots, and in the much lower amounts per acre, both of dry substance and of nitro- gen in the crops growing under the influence of only residual supply, clear indication that the nitrogenous accumulations available within the soil, whether from manure- or from crop- residue, were rapidly becoming exhausted. The figures relating to the potash per cent in the dry matter of the roots, and per acre in the roots, show (with the continued annual supply of potash), as in the case of the three years, a high percentage in the dry matter with high luxuriance — that is, where there had been a large amount of nitrogenous manure- and crop-residue; and the percentages are with one exception higher over the two years, with the same supply of potash, but much less available nitrogen, and much less luxuriance and total growth, than over the three years with the direct supply of nitrogen. On the other hand, the quantities of potash per acre in the roots, although much larger with nitrogenous residue and increased growth than with the mineral manure alone, are, with the much less EOOT-CEOPS. 43 ' growth than during the three years, generally only about half as much as over the preceding period ; but, as above stated, the amount was greater in proportion to the dry sub- stance produced — the supply of potash being the same, but the available nitrogen and the consequent growth much less. Further, as over the three years, so now over the two years with only residual nitrogenous supply, and very much less growth, the percentage of potash in the dry matter of the leaf is very much higher than in that of the root ; but also as over the three years, the actual quantity of potash per acre in the leaf is very much less than that in the root. As to the phosphoric acid, its percentage in the dry sub- Phosjihoric stance of the root is fairly uniform throughout the five aci^™'^ series with the same supply of it by manure, but with great difference in the available supply of nitrogen and in the amounts of growth. The amounts of phosphoric acid per acre in the roots are, however, by no means uniform in the different series, but have a very obvious relation to the quantities of dry substance grown. The percentage of phos- phoric acid in the dry substance of the leaf is also pretty uniform throughout the different series; but the quantities per acre in the leaf, as in the root, have distinct relation to the amounts of growth. They are, however, in all cases much smaller than those in the root, and very much smaller than the amounts of potash in the leaf. The relation of the potash and phosphoric acid to the amount of substance grown wUl be further referred to presently. • The following Table (13) shows — in the upper division the Produce percentage of sugar in the sugar-beet roots under the "/^S''""'- specified different conditions of manuring; in the second division the amounts of sugar yielded per acre (in lb.) ; in the third division the increase of sugar per acre by the nitro- genous manures ; and in the bottom division the increased amount of sugar for 1 lb. of nitrogen supplied in manure. The mean results are given for the three years of the direct nitrogenous supply, for the two years of residual supply only, and for the five years, three with, and two without, the direct supply. Further, the results are given both for plot 5 with superphosphate only as the standard or mineral manure, and for the mean of plots 6 and 4, the former with superphosphate and potash, and the latter with superphos- phate, potash, soda, and magnesia, as the mineral manure. It may in the first place be observed that the percentage Effect of of sugar is about one and a-half time as high as in mangel- pg^entam' roots grown under similar conditions as to manuring. Ee- of sugar. ferring to the results for the first three years, the table shows that the percentage of sugar is the highest in Series 1 — that u THE EOTHAMSTED EXPERIMENTS. TABLE 13.— Su&ae-Bebt. Sugar per cent and per acre per annum in the roots. Averages of 3 years, 1871-73; 2 years, 1874-75; and 5 years, 1871-75. Plot. Standard manures. Series 1. Standard manures only, every year. Standard manures, every year, and— Series 2. Sodium nitrate = 80 lb. nitrogen (3 years only). Series S. Ammoni- um-salts = 86 lb. nitrogen (3 years only). Series 4. Ammoni- um-salts and rape- cake = 184 lb. nitrogen (3 years only). Series 6. Bape- cake =98 lb. nitrogen (3 years only). SUGAR PER CENT. 3 years, J 1871-731 2 years, J 1874-75) 5 years, J 1871-75] 5 4&6 5 4&6 5 4&6 Superphosphate / Superphosphate \ and potash Superphosphate f Superphosphate \ and potash Superphosphate / Superphosphate \ and potash 13.08 }l2.97 10.66 11.04 11.88 12.16 9.89 10.66 12.31 |l2.05 10.36 10.60 11.61 11.99 10.78 11.17 12.77 |l2.60 10.54 10.86 11.77 12.09 10.25 10.86 12.17 12.07 10.72 11.22 11.59 11.73 SUGAR PER ACRE, LB. e 1731 4661 3563 3886 ■1704 J 4635 4063 5279 1684 2053 1963 2591 }l531 2045 2047 2825 1672 3618 • 2923 3368 }l635 3599 3257 4297 3 years,! 1871-731 2 years, J 1874-75] 5 years, J 1871-751 5 4&6 5 4&6 5 4&6 Superphosphate ( Superphosphate \ and potash Superphosphate / Superphosphate \ and potash Superphosphate / Superphosphate \ and potash 4407 4788 2065 2262 3470 3778 INCREASE OP SUGAR PER ACRE OVER SERIES 1, LB. 3 years, J 1871-731 2 years, J 1874-751 5 years, J 1871-751 5 4&6 5 4&6 5 4&6 Superphosphate / Superph osphate \ and potash Superphosphate / Superphosphate \ and potash Superphosphate / Superphosphate \ and potash 2930 1832 2155 } - 2931 2359 3575 469 379 1007 } ■■■ 514 516 1294 1946 1251 1696 }... 1964 1622 2662 2676 3084 481 731 1798 2143 LB. INCREASE OF SUGAR FOR 1 LB. NITROGEN IN MANURE. 3 years, J 1871-731 5 years, J 1871-751 5 4&6 5 4&6 Superphosphate / Superphosphate \ and potash Superphosphate / Superphosphate \ and potash 34.1 21.3 11.7 1- 34.1 27.4 19.4 37.7 24.2 15.4 }... 38.1 31.4 24.1 27.3 31.5 30.6 36.4 EOOT-CEOPS. 45 is, without nitrogenous supply, with the least luxuriance, and the smallest and ripest roots, the mean for plots 6 and 4 amounting to 12.97 per cent. On the other hand, in Series 4, with the highest nitrogenous manure, the greatest luxuri- ance, and the least maturity, the percentage is only 10.66. Comparison of the percentages of dry matter and of sugar show that the sugar constituted about or more than two- thirds of the total dry or solid substance of the root. As a rule, where nitrogenous manure was used there was a some- what higher percentage of sugar with than without potash supply. There was also generally a somewhat higher per- centage over the three years of direct nitrogenous supply than over the succeeding two years. Eeferring to the second division of the table, which shows Manuring the amounts of sugar per acre under the different conditions "?fZ^ as to manuring, it is seen that over the three years the mean produce of plots 6 and 4 with potash was, without nitro- genous manure 1704 lb. ; with nitrate in addition 4635 lb. ; with ammonium-salts 4063 lb. ; with ammonium-salts and rape-cake 5279 lb.; and with rape-cake 4788 lb. In other words, with little more than three-fourths of a ton of sugar per acre with the mineral manure alone, there was, with nitrogenous manure in addition — when as ammonium-salts more than If ton, with nitrate more than 2 tons 1 cwt., with rape-cake nearly 2 tons 3 cwt., and with rape-cake and am- monium-salts more than 2 tons 7 cwt., of sugar produced per acre. Over the subsequent two years, without further nitro- genous supply, there was, however, generally about, or not much more than, half as much sugar yielded. The third division of the table shows that with superphos- Superphos- phate and potash as the mineral manure, there was over the ^^^^ three years an average annual increase of sugar yielded, per acre, due to the nitrogenous supply, of 2931 lb. by the nitrate, of 2359 lb. by the ammonium-salts, of 3575 lb. by the ammonium-salts and rape-cake, and of 3084 lb. by the rape-cake. Over the succeeding two years, however, the in- creased production of sugar, due to the nitrogenous residue, was, with the nitrate less than one-fifth, with the ammonium- salts rather more than one-fifth, with the ammonium-salts and rape-cake more than one-third, and with the rape-cake alone less than one-fourth, as much as over the three years with the direct supply of nitrogen. Upon the whole, therefore, it is evident that even with a Depend- full supply of mineral manure the produce of sugar was ^f^gf small, and that the increased production of that non-nitro- nitrogen. genous substance was dependent on the available supply of nitrogen within the soil. Examination of the table will 46 THE EOTHAMSTED EXPBEIMENTS. Sugar-pro- duction Carbohy- d/rates of swpply of ous man- wresfor crops poor im nitro- further show that where ammonium-salts, ammonium-salts and rape-cake, or rape-cake alone, was employed, there was considerably more sugar produced on plots 4 and 6, where potash was supplied, than on plot 5, where superphosphate was the only mineral manure. Doubtless with the continued supply of superphosphate alone as the mineral manure, and the growth forced by nitrogenous supply, the amount of pot- ash available within the range of the roots had become more or less exhausted. Where the nitrogen was applied as nitrate, however, there was no deficiency of sugar -production with superphosphate only as the mineral manure ; a result prob- ably due, as already observed, to the greater range of the roots induced under the iniluence of the soluble and more rapidly distributed nitrate, thus securing a better command of the potash of the soil and subsoil. The bottom division of the table illustrates very strikingly the interesting fact of the dependence of the amount of the non-nitrogenous substance — sugar — produced on the amount of nitrogen available within the soil. Thus, taking the results for plots 6 and 4, with full mineral supply including potash, there is over the three years — for 1 lb. of nitrogen supplied — when as nitrate 34.1 lb., as ammonium-salts 27.4 lb., as rape-cake 31.5 lb., and when applied in excessive amount in ammonium -salts and rape-cake together 19.4 lb., of sugar produced. Taking the results for the five years, three with direct supply and two with residue only, the increased pro- duction of sugar for 1 lb. of nitrogen supplied is somewhat greater — namely, with the nitrate 38.1 lb., with the ammo- nium-salts 31.4 lb., with the rape-cake 36.4 lb., and with the ammoniuin-salts and rape-cake together 24.1 lb. It will be seen, however, that when superphosphate without potash was used as the mineral manure, the produce of sugar for a given amount of nitrogen in manure was, excepting in the case of the nitrate, distinctly less. It is not only in the case of sugar-beet that the amount produced of the special carbohydrate of the plant is largely influenced by the supply of nitrogen. It is so in the case of root-crops generally, which may be fitly called sugar-crops. As we shall see further on, the result is very similar in the case of grain crops, the produce of which is greatly increased by nitrogenous manures ; and in their case it is the carbo- hydrates — starch and cellulose — that are chiefly produced. It is_ also much the same with potatoes, the increased pro- duction of starch being then the characteristic result. In fact it will be found that nitrogenous manures are chiefly used for crops poor in nitrogen, the increased produce of which is characteristically that of non-nitrogenous bodies. Without attempting to give a physiological explanation of EOOT-CEOPS. 47 the result, it may at any rate be stated as a matter of fact that nitrogenous manures greatly increase the general vege- tative activity of such plants, and consequently, if the other necessary supplies are not wanting, the activity of the formation of their natural or characteristic products is enhanced. It has been seen that the supply of potash as well as of Potash and nitrogen has much to do with the amount of root-develop- f^^/-^ ment, and the amount of sugar produced. The following table shows the amounts of sugar for 1 of potash, in the roots. The supply of potash was the same in all cases ; in Series 1 without any nitrogenous manure, but in the other series the nitrogenous manures as indicated, in each of the first three years. The results are the means of plots 6 and 4, over the three years with the direct supply of nitrogen, over the two years without further nitrogenous supply, and over the five years, three with and two without, nitrogenous manure on Series 2, 3, 4, and 5. SUGAK FOR 1 OF POTASH IN THE ROOTS. Series 1. Without nitrogenous manure. With sodium nitrate Series 3. With ammonium- Series 4. With rape- cake and ammonium- salts. Series 5, With rape-cake. 3 years, 1871-73 2 years, 1874-75 5 years, 1871-75 47.9 43.4 46.1 36.5 38.7 38.9 34.1 37-6 34.9 42.0 39.1 41.3 In the first place, it is to be observed that the amount of sugar for 1 of potash in the roots is considerably the greater where no nitrogen was supplied by manure, and where there was no luxuriance, and by far the ripest roots ; con- ditions under which the sugar produced would presumably be the maximum for the amount of nitrogen available, and probably also the maximum for the amount of potash present in the roots. On the other hand, the lowest amounts of sugar for 1 of potash are, upon the whole, in Series 4, where there was excess of nitrogen, great luxuriance, the lowest maturation, and consequently the crudest juice. Comparing period with period, the least amount of sugar for 1 of potash in the roots was generally over the two years with full supply of potash, but deficient supply of nitrogen, and de- ficient yield of sugar. In the cases of most normal growth, it would seem that there were for 1 part of potash about, or nearly, 40 parts of sugar in the roots. In reference to these results, it is to be borne in mind that the percentage of potash remaining in the dry substance of the leaf, where 48 THE EOTHAMSTED EXPERIMENTS. Nitrogen manure and re- crop. carbohydrates are so largely formed, was much higher than in that of the root ; though, as Tables 10 and 12 show, by far the greater part of the total potash of the crop was found in the root, where is the great accumulation of sugar. Before leaving the subject of the experiments with sugar- beet, it will be well to refer briefly to the amount of the nitrogen supplied in manure which is recovered in the in- crease of crop. Below are shown the amounts recovered in the increased produce of the roots only, taking the mean of plots 6 and 4, with potash as well as superphosphate as the mineral manure. The results are given for the three years of the direct supply of the nitrogenous manures, and for five years, three with and two without, the direct supply ; and the figures show the amounts of nitrogen recovered in the increased produce of roots for 100 supplied in manure: — 3 Years. 5 Year With nitrate of soda . 61.3 66.9 With ammonium-salts 42.9 49.0 With rape-cake 45.0 52.7 With rape-cake and ammonium-salts . 49.6 57.4 As the leaves are annually returned to the land as manure, it will be obvious that, taking the average over a number of years, it is only the amount in the roots that can be credited as immediate return from the manure employed. It is seen that the highest amount recovered is from nitrate of soda — namely, 61.3 per cent over the 3 years, and 66.9 per cent over the 5 years; next we have 49.6 per cent over the 3 years, and 57.4 per cent over the 5 years, with ammonium- salts and rape-cake ; then 45 per cent over the 3 years, and 52.7 per cent over the 5 years, with rape-cake ; and lastly, only 42.9 per cent over the 3 years, and only 49.0 per cent over the 5 years, with ammonium-salts. These amounts are, however, higher than those obtained with wheat or barley — a result no doubt chiefly due to the period of accumulation and growth extending much later in the season than in the case of those grain crops ; and hence also, no doubt, is to be explained the much greater accumulation of nitrogen under equal conditions of soil by maize than by either wheat or barley. We shall recur to this subject further on. Plan of ex- pe/riments with man- 4. Experiments with Mangel- Wurzel. We have now to consider the results of experiments with manzel-wurzel, a variety of beet largely used in some districts of our own country for feeding purposes. The experiments were made in the same field, and on the same plots as those EOOT-CEOPS. 49 with the turnips and sugar-beet ; and following the sugar- beet, they were commenced in 1876, and are still continued — the last crop, that of 1894, being therefore the nineteenth in succession. We propose to draw our illustrations from results obtained in the field during the 17 years, 1876-92, and in the laboratory during shorter periods. Table 14 (p. 50) gives the average produce — roots, leaves, and total — over the 17 years for six plots, each with five different conditions as to nitrogenous supply. A glance at the table shows that the produce of roots of Mamgds the mangel-wurzel is on a much higher level than that of ^/'^^ either common or Swedish turnips, and there is also much piured, more leaf. There was, however, a general similarity in amount of produce obtained under similar conditions of manuring with the mangel as with the sugar-beet. Compared with turnips, the mangel-seed is sown earlier, and the plant has a longer period of growth. It has a much more deeply penetrating tap-root, throws out a less proportion of its feed- ing-roots near the surface, and exposes a comparatively large area of leaf to the atmosphere. With its more extended root-range, it is less dependent on continuity of rain when growth is once well established; and it bears, or rather requires, for full growth a higher temperature than the turnip. These conditions determine in what localities it is most suitably grown in this country. But where the soil and climate are suitable, very much larger crops can be obtained than of turnips. The mangel requires, however, very heavy dressings of manure if it is to yield full crops. The Table (14) shows that with farmyard manure alone. Dung The question as to the condition of the nitrogen in vegetable foods, and especially in such crude and immature EOOT-CEOPS. 63 products as our feeding roots, is therefore one of great im- portance. In the early reports of the Eothamsted feeding experiments, published more than forty years ago, we called attention to the fallacy of estimating the whole of the nitrogen of our stock-foods as protein or albuminoid com- pounds, especially in the case of succulent and unripened products. Table 19 (p. 64) gives results as to the condition of the Swedes. nitrogen in Swedish turnips grown in the experimental rota- tion at Eothamsted in 1880 ; also in the mangels grown ia the experiments in 1878, 1879, and 1880. It should be explained that one portion of the rotation wuhmit land has been entirely unmanured throughout, and that the "«"»«»•«■ roots so grown are quite abnormal, none of the characters of the cultivated root being developed under these circum- stances. The results given relate to the roots grown in 1880 as the first crop of the ninth course. . It is seen that with an abnormally high percentage of total nitrogen in the roots (0.347 in the fresh, and 2.758 in the dry), there was also a high percentage of albuminoid nitrogen ; which cor- responded, however, to only 32.9 per cent of the total nitrogen. The next plot had received, for the roots, superphosphate SuperpJws- of lime alone. Under these conditions the roots of the ninth p¥-*^ course show a very low percentage oi nitrogen m their dry substance (0.984), but 59.1 per cent of it existed as albu- minoid compounds. Lastly, the third plot received for the roots of each course CompUx a complex manure, both mineral and nitrogenous. The ^a""™- percentage of total nitrogen in the dry substance of the roots (1.539), though not high, was nevertheless more than one and a-half time as high as in the case of the roots grown by superphosphate alone ; and the proportion of the nitrogen which was as albuminoids was only 42.5 per cent. Then, again, it is seen that in the cultivated roots by far M.. 0. "3 S •A" S "3 - la » ■e »oos 00 0) ,| p-ff- ffi 0.0- o« CtM(7- 0" doo 00 ■*CO CO a ea p< CM o 1 ^ rt OirHiffl OIM o w !-oo 000 000 00 e:] u fe odd GO odd odd CO e6 d q )-H p< 5 00 rH 00 4^ J- a ^hT t| K a oo^cn S OQO m H Ofr- (xl ■*»n 01 IS) M 1 Sg s § P S3 S P< i-H r-4 i-H rH rH rH rHr-J rH 1 &: ^ S J ■3 'm • r^OiH ^ rHt- rH H a cqfr- to Eg ■>ai-\ CO ■s S-p g "dividual years during the latter half of the whole period. There was, in fact, great fluctuation due to season; but there is also evidence of reduction due to exhaustion in some cases. The bottom line of the tables, which shows the percentage reduction in the amount of produce over the second twenty years compared with the first twenty, enables us to dis- criminate in some degree between the effects of exhaustion sgecu of and those of season. exhaustion It is seen that the four plots of Series 1 show a reduction over the second twenty years of from about 30 to 40 per cent, or about twice as much as in the case of either of the other series. There is here evidence- that in the case of Series 1, without nitrogenous manure, much of the reduction over the second half of the period was due to nitrogen iVUrogen, exhaustion. exhaustion. In Series 2, with ammonium-salts, there is about 21 per cent reduction on plot 1, where the ammonium-salts are used alone, nearly as much on plots 2 and 3 with defective mineral manuring, and only about 12 per cent where full Effect of mineral manures are used in addition. mineral Tit (ZTtUT'BS In Series 3, with sodium nitrate, there is a reduction of about 23 per cent where the nitrate is used without mineral manure, of 21 per cent where it is used with potash, soda, and magnesia, but without phosphate (plot 3), and of only 14 to 17 per cent where phosphates were used in addition to the nitrate. Lastly, in Series 4, with rape-cake, which contains a con- siderable amount of mineral matter, ther^is a reduction of about 18 per cent on plots 1, 3, and 4, but of only about 14 per cent on plot 2 with superphosphate only as the mineral manure. As already intimated, that there should be any reduction influence in the yield over the second half of the period where rape- o/««««o«. cake with its annual residue and accumulation is used, is ^hShm evidence that part of the reduction is due to an average of amd^hos- less favourable seasons over the later period. But that there ^hauaion. 78 THE EOTHAMSTED EXPERIMENTS, General view. should be the greatest reduction in Series 1, where no nitrogen is supplied, is. evidence of nitrogen exhaustion under those conditions ; and that, within Series 2 and 3 respectively, there should be the greatest reduction where the ammonium- salts or nitrate is used without phosphates, is evidence of phosphoric acid exhaustion in those cases. Leaving the results relating to the produce of each indi- vidual year, or of limited series of years, as given in Tables 22 and 23, a general view of the effects of the sixteen different conditions as to manuring is conveniently obtained in the summary Table 24. There is there given the aver- age produce oyer the forty years on each of the sixteen TABLE 24. — Summakt showing the average produce op Bar- ley PER ACRE PER ASNTJM, OVER FORTY YeAR^, BY DIFFERENT Manures. Plot. 200'lb. 275 lb. No nitro- ammon.- sodium genous salt3= nitrate 1 manxire. 431b. = 43 lb. nitrogen. nitrogen. 1000 lb. rape-eake2 =49 lb. jiitrogen. DEESSED GRAIN PER ACRE, BUSHELS. 1 Without mineral manure . 164 29 32i 41J 2 Superphosphate 2ia 421 45i 431 3 Potassium, sodium, and magnes- ium sulphates 18 311 334 394 4 Superphospate, and potassium, sodium, and magnesium sul- phates 22f 434 454 43i STRAW PEE ACRE, LB. 1 Without mineral manure . 1044 1793 2127 2624 2 Superphosphate 1210 2674 3018 2792 3 Potassium, sodium, and magnes- ium sulphates 1076 2011 2322 2627 4 Superphosphate, and potassium, sodium, and magnesium sul- phates 1279 2904 , 3186 2875 TOTAL PRODUCE (GRAIN AND STRAW) PEE ACRE, LB. 1 Without mineral manure . 1976 3420 3964 4953 2 Superphosphate 2422 5080 5596 6251 3 Potassium, sodium, and magnes- ium sulphates 2079 3773 4208 4876 4 Superphosphate, and potassium, sodium, and magnesium sul- phates . . . . • 2530 5365 5761 5319 ^ Ammonium-salts = 86 lb. nitrogen first 6 years, =43 lb, next 10 years; Ddium nitrate = 43 lb. nitrogen each year since. sodium nitrate = 43 lb. nitrogen each year i '^ 2000 lb. rape-cake first 6 years, 1000 lb. since. BAELEY. 79 plots. The first column gives the results for the four plots of Series 1, without nitrogenous manure ; the second column those for Series 2, with ammonium-salts equal to 43 lb. nitrogen per acre per annum ; the third those for Series 3, first with ammonium-salts and afterwards sodium nitrate ; and the fourth those for Series 4, with rape-cake. The upper division of the tahle gives, for each plot, the average produce of grain per acre in bushels; the middle division the average produce of straw in lb. ; and the lower division the average total produce (grain and straw together) in lb. Eeferring first to the results on the four plots without nitrogenous manure, as given in the first column of the table, it is seen that plot 2 with superphosphate, and plot 4 with superphosphate, and potassium, sodium, and magnesium Phosphates sulphates, give considerably more produce than plot 3 with '^"'^ "i|"^"* the potash, soda, and magnesia, without phosphate. There is nitrogen. more of straw as well as grain, and of course, therefore, of total produce, with than without the phosphate. There is, indeed, very marked effect by phosphatic manure, and very little by the alkalies. The second column, with the same four conditions as to mineral supply, but with, in each case, 43 lb. of nitrogen per wm acre per annum as ammonium-salts, shows a very great "»*™S'«»- increase. Even with the ammonium-salts alone there is a great increase ; "there is somewhat more on plot 3, where the alkalies are alao-applied, but very much more still on plot 2, where superphosphate, and on plot 4, where alkalies and superphoshate, are also used. The third column shows that, with a larger amount of Greatest nitrogen supplied in the first six years, and with sodium ^"^^^^^g. nitrate instead of ammonium-salts in the later years, there is gen and still greater increase ; and again, the increase is by far the ^^ff^^' greater where the superphosphate is used. The four plots of Series 4, with the rape-cake, show a Rape-cake much greater uniformity of result with the different mineral ""^ '''^*'" manures. Still, the two phosphate plots (2 and 4) give more produce than the two without phosphate. Eeferring to the produce of grain in illustration, it is seen that plots 1 and 3 with rape-cake ' without superphosphate, give considerably more produce than the same plots (1 and 3) in either Series 2 with the ammonium-salts, or in Series 3 with sodium nitrate. The explanation of this is that the rape-cake itself contains phosphates. On plots 2 and 4, on the other haiui, where phosphates are added, there is about as much produce in Series 2 with the ammonium-salts, and more in Series 3 with the nitrate, than in Series 4 with the rape-cake. Thus, then,, whilst there is evidence that the phosphate of the rape-cake was effective when none was otherwise supplied, 80 THE EOTHAMSTED EXPERIMENTS. Nitrogen available. Potash mwnii/res. Potash of the soil. results with artificial manmes. Superphos- phate for spring- sovm crops. when it was so applied in addition, there was more effect with the nitrate, with its more rapidly available nitrogen, than with the rape-cake with its greater actual amount of nitrogen, but in a less rapidly available condition. Comparing the produce of plot 2 with superphosphate with- out potash, with that of plot 4 with superphosphate and' potassium, sodium, and magnesium sulphates in addition, it is remarkable that, both in Series 2 with the ammonium-salts, and in Series 3 with nitrate of soda, there is, over the whole period of forty years, almost identically the same amount of barley grain without as with the potash. There is, however, rather more straw, and total produce, with than without the potash. Thus we have, with the ammonium-salts an average of 42f bushels without potash, and 43J bushels with potasjh ; and with the nitrate of soda 45f bushels without, and 45^ bushels with potash. Of straw, however, there is with the ammonium-salts an average of 2674 lb. without, and 2904 lb. with the potash ; and on the nitrate plots 3018 lb. without, and 3186 lb. with potash. , It will afterwards be seen that where nitrogen and phos- phoric acid were liberally supplied without potash, the avail- able potash of the soil itself became deficient ; though this deficiency was to the last comparatively little manifested in the produce of grain. It is obvious, however, that with gradual reduction in the amount of total plant, the yield of grain must also in time materially diminish. So much for the influence on the barley crop of different conditions of manuring, each continued for more than forty years, on the same plot, and in a field of somewhat heavy loam, with a raw clay subsoil, and chalk below giving good natural drainage. It is seen that nitrogenous manures alone had much more effect tha,n mineral manures alone. It was obvious, therefore, that the exhaustion induced by the continuous growth of the crop was characteristically that of nitrogen. Both with and without nitrogenous supply, phosphates were more effective than potash salts, showing that the avail- able store of phosphoric acid in the soil became deficient sooner than that of potash. With the shorter period of growth of barley than of wheat, and its greater proportion of surface- rooting, both nitrogenous, and mineral exhaustion are sooner developed ; and so far as mineral exhaustion is concerned, the available supply of phosphoric acid was sooner exhausted than was that of potash. Indeed, in ordinary agricultural practice, it is clearly established that superphosphate is more effective with the spring-sown than with the autumn-sown cereals. BAELEY. 81 Influence, of Season on the Amounts of Produce. It has been seen that there were, under all conditions of Variations manuring, very great variations in the amount of produce of.W'o^'^ from year to year, according to season. The extent and i Z character of the influence of season will be brought promi- ^^'^<>™- nently to view by comparing the produce of the best and the worst seasons of the forty, and comparing the characters of the seasons themselves. Tables 25 and 26 illustrate these points. Table 25 (p. 82) gives the produce of grain, the weight per bushel of the grain, the produce of straw, and the total produce (grain and straw together), of six very different conditions as to manuring in each of the best two seasons, and in the worst season of the whole series. There is also given4he deficiency of produce in the bad season compared with that in each of the two good seasons. For wheat, 1863 was the best season of the forty. For barley, 1863 was also a very good year for both grain and straw ; but it was not so good for such a variety of manures as were 1854 and 1857, which (in the table) are adopted as the best seasons. For almost all conditions of manuring, 1854 was the season Best of the highest total produce, grain and straw together ; that ^"^°^- is, it was the season of the greatest luxuriance or vegetative activity. But 1857 was, especially for the highest manur- ing, the one of the highest produce of grain, and of the high- est quality or maturity of grain, as evidenced by the weight per bushel. Thus, 1854 was the highest for luxuriance, and 1857 the highest for maturation, of the crop. For wheat, 1879 was decidedly the worst season of the forty. Worst For barley also 1879 was a very bad season; but 1887 was "" worse still, especially for high manuring, and it is therefore adopted as the worst season for barley. The plots selected for illustration are those without manure, with farmyard manure, with mixed mineral manure alone, with mixed mineral manure and ammonium-salts, with mixed mineral manure and nitrate of soda, and with mixed mineral manure and rape-cake. The figures speak for themselves, and will repay careful study ; but we can only refer to them very briefly here. The lower division of the table shows that, under each of the six very different conditions as to manuring, 1854 yielded a much higher total produce (grain and straw together) than 1857. But the upper division shows that, notwithstanding there was the less amount of plant in 1857, as shown by the less amount of straw and total produce, it gave, in most cases, nearly as VOL. VII. F seasons. 82 THE EOTHAMSTED EXPERIMENTS. much grain as 1854; and in two — those with the highest nitrogenous manuring (and both years were within the first six when the larger amounts were applied), 1857 gave more grain than 1854. The weight per bushel of the grain was also higher in 1857 on all the plots where nitrogenous man- ures were used. TABLE 25. — ^Peoducb of Bablet in th^K^o best Seasons, 1854 and 185y ; in the worst season, 1887 ; and the average over forty Years, 1852-1891. Descriptions of manures ; quantities per acre. Best seasons. Worst reason, 1887. 1887 + or - 1864. 1857. age of 40 DRESSED GRAIN PEE ACRE, BUSHELS. lo 7-2 40 4a iaa Unmanured Farmyard manure Mixed mineral manure alone .... Mix. min. man. and 200 lb. am.-salts = 43 lb. N. Do. and 275 lb. sodium nitrate=43 lb. N Do. and 1000 lb. rape -cake =49 lb. N 35 26i H -27i -18* -26 56f 51i 26 42 391 8f -33| -31 60| m ■.2% -38 ■ -34i 62| e^ 26i -37i -39§ 60i 62i 21 -39i -41i 16J 48f 46i 43i y ■WEIGHT PER BUSHEL OF DRESSED GRAIN, LB. lo Unmanured ... , . 53.6 52.0 51.0 -2.6 -1.0 62.0 7-2 Farmyard manure 53.9 54.2 65.3 -1-1.4 -1-1.1 54.3 4o Mixed mineral manure alone .... ^4.0 /64.3 63.7 61.8 -2.2 -1.9 53.0 4a Mix. min. man. and 200 lb. am.-salts— 43 lb. N. 54.8 53.3 -1.0 -1.6 54.1 4aa Do. and 275 lb. sodium nitrate = 43 lb. N 82.1 53.9 53.7 -1-1.6 -0.2 63.7 4c Do. and 1000 11). rape-cake=49 lb. N 62.8 54.1 53.4 -1-0.6 -0.7 53.9 STRAW PER ACRE, LB. lo 7-2 40 4a 4aa ic Unmanured Farmyard manure ... Mixed mineral manure alone .... Mix. min. man. and 200 lb. am.-salts=43 lb. N. 'Do. and 275 lb. sodium nitrate =43 lb. N Do. and 1000 lb. rape -cake =49 lb. N 2442 4171 2595 4530 1425 2649 1920 3120 648 1842 630 1706 -1794 -2329 -1966 -2826 777 807 -1290 -1416 6487 4057 2078 -3414 -1984 4712 3705 1740 -2972 -1965 1044 3247 1279 2904 3186 2876 lo 7-2 40 TOTAL PRODUCE (GRAIN AND STRAW) PEE ACEE, LB. Unmanured Farmyard manure Mixed mineral manure alone .... Mix. min. man. and 200 lb. am.-salts = 43 lb. N. Do. and 275 lb. sodium nitrate =43 lb. N Do. and 1000 lb. rape -cake = 49 lb. N 4405 7298 4969 7958 9026 8126 2878 6664 4111 6386 7734 7241 1043 3294 1088 2875 -3362 -40U4 -3881 -5029 -6571 -1836 -2270 -3023 -3407 -4279 -4366 1976 6016 2630 6366 5761 6319 Wote.— Plot 4aa, ammonium-salts = 86 lb. nitrogen Hrst 6 years, =43 lb. next 10 years ■ sodium nitrate=43 lb. nitrogen last 24 years. Plot 4c, 2000 lb. rape-oake first 6 years, 1000 lb. since. BAKLEY. 83 The contrast between the produce in these two very differ- ent good years, and that in the worst season, 1887, is very striking ; in fact, the difference amounted in several cases to more than the average crop of the country. For comparison with the produce of these selected years, the average on each of the six plots over the forty years is given. It will be seen how very much higher than the average is the produce in the good years, and how very much lower it is in the bad season ; indeed it is, in the bad season, gener- ally only about, or less than, half as much as the average. It will be of interest to consider, however briefly, some of the climatic characteristics of these various seasons. The next Table (26) shows, for each month, of each of the Tempera- three seasons, reckoning from October in the preceding year *^ainfaU. to September in the year of growth, the mean temperature, -^ and the rainfall, above or below the average. TABLE 26. — Chahacter of the two best Seasons, 1854 and 1857, AND OF THE WORST SEASON, 1887. TEMPERATURE AND RAIN- FALL + OR — Average. Mean temperature. Eainfall. Days of rain, 0.01 inch or more. Best two. Worst. Best two. Worst. Best two. Worst. 1863-4. 1856-7. 1886-7. 1863-4. 1866-7. 1886-7. 1863-4. 1866-7. 1886-7. October i November December January February March . April . Deg.F. + 1.3 -0.2 -5.2 + 2.4 + 0.8 + 2.7 +2.3 Deg.F. + 2.1 -1.6 + 1.0 0.0 + 0.5 +0.7 -0.4 Deg. F. +3.7 + 1.7 -2.7 -1.0 +0.2 -3.5 -2.0 Inches. + 1.43 -0.45 -1.30 -0.60 -0.29 -1.28 -1.11 Inches. -0.89 -1.15 -0.27 +0.60 -1.30 -0.77 -0.30 Inches. -1.39 + 0.62 + 1.50 -0.85 -0.97 -0.25 + 0.05 Days. +13 - 2 + 3 - 3 - 6 - 4 Days. - 4 - 3 + 1 + 7 - 8 - 2 + 7 Days. + 2 + 6 +2 -7 -2 May . June July . . -1.6 -2.3 -1.3 +1.5 +3.8 +2.9 -2.7 +2.7 +4.9 + 1.51 -0.99 -0.85 -1.67 +0.80 -1.50 -0.28 -0.67 -1.31 + 5 + 1 + 4 - 6 - 2 - 2 +7 -8 -1 August . September . 0.0 + 1.6 +4.9 +3.2 +1.6 -2.5 +0.21 -1.42 +0.10 +1.00 -0.05 -0.19 + 1 - 3 + 1 -2 +4 Averages Totals + 1.5 -5.14 -5.35 -3.79 + 9 -11 +1 It is obvious that different seasons will differ almost infin- itely at each succeeding period of their advance, and that with each variation the character of development of the plant will also vary, tending to luxuriance or to maturation — that is, to quantity, or to quality, as the case may be. Hence only a very detailed consideration of climatic statistics, taken together 84 THE EOTHAMSTED EXPERIMENTS. Character- istics of the good Character- istics of the bad with careful periodic observations in the field, can afford a really clear perception of the connection between the ever- fluctuating characters of season, and the equally fluctuating characters of growth and produce. It is, in fact, the dis- tribution of the various elements making up the season, their mutual adaptations, and their adaptation to the stage of growth of the plant, which throughout influence the tendency to produce quantity or quality. ^ 8till it will be seen that the limited summary of the meteorological conditions of the seasons in question, which can alone be given here, is not without significance. First, then, as to 1854, the season of great luxuriance and high total produce. The table shows that there was an ex- cess of temperature in January, February, March, and April, with a deficiency of rain from November (1853) to April in- clusive ; but that during May, June, and July — that is, the months of active above-ground growth-.— there were lower than the average temperatures, with a considerable excess of rain in May, and then a deficiency — conditions obviously favouring continued vegetation and slow maturation. For the crop of 1857, there was less excess of temperature, and less than the average amount of rain, to the end of April ; then from May to August inclusive there was both consider- able excess of temperature and considerable deficiency of rain — that is, there were throughout the period of active above-ground growth conditions favouring seeding tendency and maturation rather than luxuriance. Thus, then, the two good seasons were very different in their climatic characteristics, as they were in the character of their produce. Compared with these, it may be mentioned that the very bad season of 1879 was characterised by much lower than average temperatures throughout the winter, spring, and summer, with at the same time great excess of rain from January to September inclusive; the result being amounts of produce greatly below the average, and very low weight per bushel of the grain. The season of 1887, on the other hand, which gave even lower amounts of produce than 1 879, especially with high manuring, and which is adopted as the " worst " season, was in some important respects very different in character. Thus, whilst the crop of 1879 failed from low temperatures, combined with excess of rain throughout, the season of 1887 was characterised by low temperatures, especi- ally in March, April, and May, but associated with a defi- ciency of rain commencing in Janiiary. The result was very restricted spring growth. In June and July, however, the temperature was considerably in excess of the average, but BARLEY. 86 "with continued and considerable deficiency of rain, the com- bination further restricting growth, and bringing on prema- ture ripening. InflueTwe of Exhaustion, Manures, and Variations of Season, on the Composition of the Barley Crops. In the case of wheat it was found that the supplies within Cmnpod- the soil — both of nitrogen and of mineral constituents — had **°p^'|f. a very direct influence on the composition of the crop so long jiumced ty as it was only in the vegetative stage ; but that there was, exhaustion, nevertheless, very great uniformity in the composition of the and season. final product of the plant — the seed — ^provided only that it was perfectly matured. The composition of the straw, how- ever, showed a very direct connection with the supplies by the soU. The composition of the grain was, on the other hand, materially influenced by variations of season. But variations of season obviously have great influence on the condition of maturation; whilst difference in maturation implies difference in organic composition — the amount of carbohydrates (starch especially) formed. In fact, such vari- ations in composition imply deviations from perfect and- normal maturation; and such deviations are associated not only with differences in the organic composition — the relation of the nitrogenous to the non-nitrogenous constituents — but with differences in the mineral composition also. It follows that variations in the composition of the final and very definite product — the seed — should be much more clearly traceable to variations of season than to variations in the supplies within the soil — that is, than to exhaustion or manures. This was found to be very strikingly so in the case of wheat, and we have now to consider how far it is so with its near ally — harley. The results given in Table 27 (p. 86) forcibly illustrate the Taife 27 much greater influence of variations of season than of manures '^P'""^- on the composition of barley grain. Many complete analyses of the ash of the grain (and also the straw), grown by different manures, and in different seasons, have been made ; and taking for illustration the important and characteristic constituents, potash and phosphoric acid, the table shows, for three very different manurial conditions, the highest, the lowest, and the mean amounts, of potash and phosphoric acid, in 1000 parts of the dry substance of the grain, and of the straw, in different seasons. The manurial conditions selected are — 1, without manure ; 2, with farmyard manure ; 3, mixed mineral manure (including potash) and ammonium-salts. First as to the amounts of potash in 1000 parts dry sub- 86 THE EOTHAMSTED EXPEEIMENTS. Potash im the crop as akd man- ures. stance of the grain of the differently manured plots, in the different seasons. It is seen that there is much greater vari- ation in the proportion of the potash in the different seasons with the same manure, than there is with the different manures. Further, the seasons showing the highest amount of potash were of much higher maturing character than those showing the lowest amounts. TABLE 27. — Highest, Lowest, and Mean Amounts, of Potash and Phosphoric Acid, per 1000 Dry Sobstance. Per 1000 dry grain. Highest. Lowest. Mean. Per 1000 dry straw. Highest. Lowest. Mean. POTASH. lo Unmanured 1871 7.66 1853 6.00 6.54 1871 11.77 1856 5.25 8.55 7-2 Farmyard man. 1871 8.36 1856 5.89 6.81 1871 22.01 1856 6.76 13.23 4ffl Mix. min. man. and amm. -salts 1871 7.98 1852 5.62 6.61 1871 22.53 1852 5.67 14.05 PHOSPHORIC ACID. lo XJnmannred 1852 10.08 1854 8.85 9.27 1856 2.60 1863 1.20 1.74 7-2 Farmyard man. 1871 10.50 1854 9.23 9.99 1856 2.92 1863 1.48 2-19 4(8 Mix. min. man. and amm. -salts 1866 10.39 1863 8.84 9.68 1866 3.12 1863 1.06 1.94 crop as m- season and manwe. 1 Next it is seen that there is stUl greater, indeed enormous, variation in the amount of potash in the dry substance of the straw, with the same manure, in different seasons. There is also great variation according to manure ; comparatively little when there was full supply, but considerable without manure — that is, with exhaustion. Turning now to the phosphoric acid in the grain, there is here again much more variation in different seasons with the same manure than with the different manures. But whilst in the case of potash there is the higher proportion in the 'bettm' seasons, in that of phosphorig. acid there are lower amounts in the dry substance in the hetUr seasons. In fact, high amount of potash in the ash, and in the dry substance of the grain, is, as a rule, associated with high maturation — that is, with high proportion of starch ; whilst high proportion of phosphoric acid is generally associated with low maturation, and with high proportion of nitrogen. The proportion of phosphoric aciji in the straw also varies more with season than with manure, and it is the highest in the worst seasons. BAELEY. 87 The connection between maturation and composition is Matma,- further illustrated by the results in Table 28, which shows ^^ '^f_ the general characters of the produce, as indicated by the tion. ■weight per bushel of the grain, of four very different seasons so far as the maturation of the grain was concerned. The table further shows— rthe percentage of ash (pure) in the dry matter of the grain, and of the straw ; the percentage of potash and of phosphoric acid in the ash of the grain, and of the straw; also the potash and phosphoric acid per 1000 dry matter of grain, and of straw — the results being the means of six differently manured plots in each season. Lastly, the seasons are arranged in the order of highest weight per bushel of grain, this being, upon the whole, the best practical measure of high quality, or at least of high maturation. TABLE 28. Harvests. Weight per "bushel of grain. lb. Per cent ash (pure) in dry matter. Per cent in ash (pure). Potash. ^""a'o'la?™ Per 1000 dry matter. Potash. ^^"^^l^ QEAIN. 1871 55.9 2.65 29.80 35.33 7.89 9.39 1863, 55.3 2.55 26.59 35.80 6.78 9.15 1852 51.7 2.48 23.84 40.89 5.90 10.13 1856 47.4 2.44 24.21 41.35 5.89 10.09 STBAW. 1871 55.9 6.27 26.01 3.68 16.57 2.31 1863 55.3 5.48 24.91 2.29 13.99 1.26 1852 51.7 4.45 14.62 4.05 6.58 1.81 1856 47.4 4.49 13.51 6.42 6.10 2.89 It will be seen that the average weight per bushel of the Season and grain was in 1871, 55.9 lb.; in 1863, 55.3 lb.; in 1852, y^^'^to/ 51.7 lb. ; and in 1856 only 47.4 lb. ; or about 8 lb. less than ^™"" in- the two seasons of highest weight. There is here, then, very great variation in the character of these four seasons, and in the degree of maturation of the grain accordingly. ■^ N"o determinations of nitrogen are available; but it may Nitrogen be stated that the percentage of nitrogen is almost uniformly '^^^^Jf'*^ lower in the seasons of high maturation. Turning to the ^™™' particulars of composition given in the table for each of the four seasons, it is seen that, in both grain and straw, there is a higher percentage of ash in the dry s.ubstance the higher the quality of the grain. There are also higher percentages 88 THE EOTHAMSTED EXPERIMENTS. Potash, phospmric acid, and quality of Recapitu- lation. ^ :) Good seasons amd soda in crop. Silica in straw. Mineral mammres a/nd min- eral com- position of crop. of potash, but lower percentages of phosphoric acid, both in the ash and in the dry substance, the higher the quality of the grain. In wheat, however, there is lower not higher percentage of ash in the dry substance of the grain the higher its quality. But in wheat, as in barley, there is higher percentage of potash, and lower percentage of phosphoric acid, in the ash, the higher the quality. On the other hand, there is not in the case of wheat, as there is in that of barley, a much higher percentage of potash in the dry substance the higher the quality. This difference may be partly due to the larger proportion of starch to nitrogenous substance in the barley ; but it is probably in part also due to the palece (or chaff) of the barley, but not of the wheat, being adherent, and retain- ing the surplus potash brought up for grain-formation. In both descriptions of grain there is very uniformly a lower proportion of phosphoric acid in the dry matter the higher the quality of the grain. In the straw there is high percentage of ash in the dry matter, high percentage of potash, and low percentage of phosphoric acid, in the ash, and in the dry matter, the higher the quality of the grain. In the straw, however, the varia- tions show a much wider range, indicating much less definite- ness, and greater irregularity in condition. Thus, then, the higher the quality of the barley-grain — that is, the higher its proportion of starch — the higher is the proportion of potash and the lower is that of phosphoric acid. Though not shown in the table, it may be mentioned that with a higher proportion of potash there is generally a lower proportion of both lime and magnesia, and with a lower pro- portion of phosphoric acid there is a somewhat higher propor- tion of sulphuric acid. Another point of interest is, although it is true the amounts are small, that there is a tendency to a higher proportion of soda in the grain-ash, and in the dry matter of the grain, in the better seasons, even when there is no deficiency of potash. This, again, is probably due to the ash of the barley-grain containing that of the adherent palece. In relation to the composition of the straw, the most striking result is (though not shown in the table) that there is little more than two-thirds as high a percentage of silica in the ash of the produce of the better as in that of the worse seasons. The results in the next Table (29) illustrate the influence of exhaustion and oifull swp;ply, of mineral or ash constituents, on the mineral composition of the produce, both grain and straw. BAELEY. 89 i ,8 « -li ia n .2 a ■ a" 1 ^1 1 !ii ta Pi Pi i-( ^1 4 Ph 1 1 4 i ■"^ gf i & m (-1 P j! s 1 It o f, P^ -B « 11 « a & 1! Is" T-( < II « tH "i^ ■* cu -a [3 !S § 3'^ a flS w <) II. Is 11 4 ?! tw 1-° '1 is. 04 g & 1 e ■^ O ^1 « lb. 63.7 63.7 51.6 44.8 lb. 35.6 30.9 19.5 15.7 ■* s lb. 39.9 48.4 37.8 32.0 «5 OS CO 04 4 CO lO CO N '^ iH 1-1 i-l r-i OS lb. 13.1 14.6 11.5 97 oq 04 per coDt. 14.65 18.61 18.10 16.25 ^ CO per cent 8.54 6.41 4.41 3.38 g i6 . percent 6.52 6.82 6.99 6.90 CO per cent. 6.22 6.23 6.02 5.85 CD per cent 27.85 32.92 33.64 29.72 CO o r4 CO. per cent. 18.44 13.31 9.72 7.36 «• c^ per cent. 27.62 28.46 28.85 28.67 per cent. 26.79 25.97 25.68, 26.35 JoSopS irt (Or- CO 00 00 CO 00 s = = = OS 04 g OOt-t-04 CO CO 04 04 CO -cH04C0t- »o OOiOtHO M rHrH CD-* COO 00 eoeoou:iio ^ SQ cq--* 00 t-( ■* oooo o &ggs s ' ddoo o gsss O) «- «- oi r4 Oq CM OS OS to ,*eo«>oo oa in ■^sss o r-1 OOt-Tjf s oooo o mt-eOTH r-tO 00 OS in 04 TH* o = = = o •* lot--* CO m «oSS^ ss m^t-QO 00 S3SSS ° i CO CD CO CD ^ i "*■ cocoeo CO i-cqoor-i 04 CO CO oq CO CO e4rHrH^ 00 osooot- 1-HCqrHiH 03 I-l 00 04 OS t- CO OOOCDtH r-iCq rHrH 1^ CO Oil- OS loco mco in . i-lr-lrH 1-1 i-i s o ssse 5; """" " SSISS S3 m OS 00 OS OS Cft CM in «4 04 04 OS 1^00 00 S 00 00 00 oo' 00 . OS-^CJSCO OS 04 04 Cq CO -09 to in CO CO Oincoir- 00 CQCqcOCO CO COt-HCOCO mco CO in gsgg s m CO in in in CO CD 04 o SSfeS co cot- 00 OS ^ 1852 1862 1872 1882 s M ~ ~ ~ >1 o - r § Cs^eOO iH rHOqCOt^ ^-CD■*^ fe CDr-ICOCD in S!g33 g eor-ieoo oq Ol 04 to CD 00 in CO CO us in 00 CO CD !>- in in mco 5S CO CO CO lO OS c»ocot- 00 i-"incO rH o OS I-l 00 t- OS gsssg g oq OS00C3S § CO 04 04 CO r-fOlCOrH s n oqwrtOT Sj 3 EC r-iocoeo MfcDcqo §3 Tji ■*■ ■*' ■# "H ssss s •*■*■<*■* tH sags S SSSfe S feSSg S; &3SS in &S5IS o 00 en i^^ CO 00 i-H gggs g 00 ooooo s T--.HrH.-l CD ^- 00 OS OS 1852 1862 1872 1882 oq b'-'- ■" o - - r s 90 THE EOTHAMSTED EXPERIMENTS. They relate to the mineral composition of the produce grown for forty years in succession : 1. By ammonium-salts and superphosphate. 2. By ammonium -salts, superphosphate, and potassium, sodium, and magnesium, salts, in addition. There are given results obtained by complete analyses of the ash of samples mixed in proportion to the amount of the produce (grain and straw separately) each year — for the four ten-year periods, 1852-61, 1862-71, 1872-81, and 1882-91. The upper division of the table gives for the potash, the second for the soda, the third for the phosphoric acid, and the fourth for the silica — 1. The percentage in the ash (pure) of the grain, and of the straw. 2. The' amounts per 1000 dry matter of grain, and of straw. 3. The amounts per acre per annum, lb., in the grain, in the straw, and in the total produce (grain and straw together). Potash, First referring to the potash : its percentage, even in the grain -ash, is seen somewhat to diminish from period to /period where none was supplied in manure, and somewhat to // increase where there was an annual supply of it by manure. In the straw-ash, however, the percentage of potash went down from 18.44 over the first period to only 7.36, or less than half, over the fourth, where none was supplied ; but it increased from 27.85 per cent over the first, to 33.64 over the third, but to only 29.72 over the fourth period, where it was annually supplied. Thus the influence of exhaustion, or of full supply, of potash, has been comparatively small on the mineral composition of the grain, but very great on that of the straw. The point is further illustrated in the next results, which show the amounts of potash per 1000 dry matter of grain and of straw respectively. There is, again, comparatively \ little variation in the relation of the potash to the organic / matter in the case of the grain, but very great variation inJ that of the straw, accordingly as there is exhaustion or full supply. When it is borne in mind that the ash of barley- grain contains that of the adherent palece as well as that of the grain proper, the conclusion is that the variation in the proportion of potash to the fixed organic substance of the grain itself, is much less than the figures would indicate. It is probable that the variation, such as it is, is associated with a different relative proportion of the organic compounds themselves — of the fully -matured non- nitrogenous to the nitrogenous bodies. In fact, the evidence, duly considered. BAKLEY. 91 is not in favour of the view that there is variation in the proportion of the potash to the fixed and ripened non- nitrogenous constituents, with the formation of which it is probably to a great extent associated. The effects of exhaustion, or of full supply, of constituents. Amount of are more strikingly still brought out by a study of the figures f^l** showing the amounts of potash taken up and retained per per acre. acre by the above-ground growth, without and with the supply of it. Thus the average amounts of potash per acre per annum, in the entire crop (grain and straw together) were, over the four successive periods without supply of it — 35.6, 30.9, 19.5, and 15.7 lb. ; and with full supply they were, over the same periods — 53.7, 63.7, 51.5, and 44.8 lb. That is to say there was,, without supply, less than half as much potash annually stored up in the crop over the last as over the first ten years of the forty. On the other hand, with full supply, there was over the second period more than, and over the third about the same amount as, over the first period, but there was less over the fourth. Further, there was, over the first period about one and a-half time, over the second more than twice, over the third more than two and a-half, and over the fourth nearly three times, as much potash in the total crop with as without supply. Lastly, over the forty years there was, without supply of potash an average of only 25.4 lb., but with it 53.4 lb. of potash per acre per annum in the crop. Yet with these enormous differences in the amounts taken Potash ac- up and retained by the entire above-ground growth in the ^"^J^"*"^ different cases, there was proportionally very much less grain. difference in the amounts accumulated in the grain. Thus, over the first period, the amounts in the grain were, over the first period — without supply 13.1 lb., and with it 13.8 lb. ; over the second — without supply 14.5 lb., and with it 15.3 lb.; over the third — without supply 11.5 lb., and with supply, 13.7 lb. ; and over the fourth period — without supply 9.7 lb., and with supply 12.8 lb. Lastly, over the total period of forty years the amounts were — without supply 12.2 lb., and with supply 13.9 lb. It is thus seen that over each period there was rather less in the grain without than with supply, but that the deficiency was not material until the third period — that is, until after twenty years without supply in the one case, and twenty years with it in the other. In reference to these results, it will be of interest to con- Amount of sider what were the actual amounts of produce — grain, straw, P^"^^^- and total — on each of the two plots, over the successive 92 THE EOTHAMSTED EXPEKIMENTS. ten-yearly periods, and over the forty years. Table (30) gives particulars on these points :- TABLE 30. The following Potash and total produce. Potash in grain and straw. Dressed grain. Straw. Total produce. ATinmonium-salta-43 lb. nitrogen and superphosphate. Without potash. With potash. Without potash. With potash. Without potash. With potash. 2a 4a ia ia 2a 4a 10 years, 1852-61 . 10 years, 1862-71 . 10 years, 1872-81 . 10 years, 1882-91 . bushels. 361 bushels. m 40| 40f •owt. 27i 20i 19i cwt. 28J 28 234 23§ lb. 5683 5837 4584 4218 lb. 5827 5808 4969 4854 40 years, 1852-91 . 421 43J 23| 25J 5081 5364 It will be seen that there was almost identically the same amount of produce of grain per acre per annum over the forty years without as with the supply of potash — the aver- age annual deficiency being only f bushel; and the details show that the falling off was chiefly during the fourth period of ten years. There was, however, some deficiency of straw without potash-supply over each of the four periods. It was considerable over the third and fourth periods, and it amounted to an average of 2 cwt. per acre per annum over the forty years. It would appear, therefore, that the diminished amount of potash taken up by the plant where it was not supplied was sufficient for the exigencies of grain-formation for the greater part of the whole period; and that at least a large proportion of the excess taken up where it was liberally supplied was surplusage so far as the requirements of the grain were concerned. Some idea of how great was the surplusage may be formed by reference to the difference in the amounts of potash eventually remaining in the straw. Thus the average amounts of potash per acre per annum in the straw were — over the first period, without supply 22.5 lb., and with it 39.9 lb., or + 17.4 lb. ; over the second period, without supply 16.4 lb., and with it 48.4 lb., or + 32.0 lb. ; over the third period, without supply 8.0 lb., and with it 37.8 lb., or -|- 29.8 lb. ; over the fourth period, without supply 6 lb., and with it 32 lb., or -|- 26 lb. ; and over the forty years, without supply 13.2 lb., and with it 39.5 lb., or 26.3 lb. per acre per annum more with than without supply. It is not to BAELEY. 93 be supposed, however, that the whole of these plus amounts were surplusage ; for although the average yield of grain has been to such a great extent maintained, the character of the plant has obviously depreciated for a good many years, and several times in recent seasons even the yield of grain has been considerably deficient. Indeed it would seem that the plant has become more and more sensitive to adverse con- ditions of soil and season. Turning now to the soda, it is seen that, whether we look Soda in the at its percentage in the ash of the grain and of the straw, its ''™^' proportion in 1000 dry substance, or the amounts in the acreage crops, very much more was found in the crops grown without its supply, but where potash was deficient, than where soda was itself annually supplied. This is strikingly illustrated by reference to the average amounts per acre per annum in the total crops, grain and straw together. Thus the average amounts of soda in the total crop were — over the first period, without any supply of either potash, soda, or magnesia, 8.4 lb., and with the supply of all three, only 3.8 lb.; over the second period, without the supply 15.2 lb., and with it only 3.7 lb. ; over the third period, without the supply 11.8 lb., and with it only 2.7 lb. ; over the fourth period, without the supply 10.7 lb., and with it only 2.2 lb. ; and lastly, over the forty years, without supply of either potash, soda, or magnesia, 11.5 lb. of soda, and with the supply of all three, only 3.1 lb. of soda per acre per annum. Thus, then, not only was there much more soda taken up or retained by the plant where it was not supplied than where it was, but it is evident that there was the more soda taken up the less the supply of potash. The amounts of soda retained in the grain are, however, seen to be but small ; there was more, it is true, where there was a deficiency of potash, and where more soda was taken up. But looking to the amounts of soda per cent in the grain-ash, or per 1000 dry substance of the grain, it would seem probable that the larger amounts where there was a deficiency of potash, and more total soda taken up, were only due to larger amounts eliminated from the grain proper, and retained in the adher- ent palecB, or chaff. Whether, however, the soda has been of any avail in the earlier or merely vegetative stages of growth, as a carrier, or otherwise, may be a question. Next as to the phosphoric acid, of which there was the same Phosphme annual supply on both plots. It is seen that, whether we ^^'^ *^ take its percentage in the ash, its proportion to the dry sub- stance, or its average quantity per acre, the amounts are, in the comparable cases, comparatively uniform ; the differences not being greater than can be supposed to be connected with 94 THE EOTHAMSTBD EXPERIMENTS. Silica in the crop. Available mineral plant-food in the soil. Soil-analy- sis % able. Liebig's analyses of Rotham- tha differences in growth due to the differences in the supply of other constituents. Lastly, as to silica; the chief point of interest to- remark is that, as the figures show, its percentage in these barley- grain-ashes ranges from under 17 to more than 20, whereas in wheat-grain-ash it ranges only from about 0.5 to about 1.5 per cent ; or, if we take the proportion of silica to 1000 dry substance of grain, in barley it ranges from 4 to 5 parts, and in wheat only from about 0.1 to about 0.3 parts. This differ- ence is obviously due to the chaff being adherent in the case of barley and not in that of wheat; and the figures afford clear illustration of the material degree in which the composi- tion of barley-grain-ash is influenced by the inclusion in it of what is, in a sense, extraneous matter. It is indeed obvious that under such circumstances we should expect, as we find, less definiteness in the mineral composition of the grain of barley than in that of wheat. In reference to the foregoing results showing the influence of exhaustion and of supply, of certain mineral constituents within the soil on the mineral composition of the produce grown, it is obviously of interest to consider, as far as exist- ing evidence will permit, the amount, and the condition of availability, especially of the potash and the phosphoric acid, within the soil. Unfortunately, results obtained by the generally adopted methods of soil-analysis do not enable us to discriminate between the total and the immediately or approximately available constituents. The difficulty was recognised and pointed out at Eothamsted very early in the course of our investigations. From time to time the subject has also been discussed by others ; and in recent years several experimenters have approached it from various points of view, with the object of fixing upon some useful modification of method. More than twenty years ago, Hermann von Liebig having asked for samples of some of the plots of the Eothamsted experimental wheat-field, samples from five plots, to three depths of 9 inches each in each case, were -supplied to him. He determined in them, besides other constikients, the potash and the phosphoric acid, the former in a dilute acetic acid extract, and the latter in a dilute nitric acid extract. The results unmistakably showed differences in the amounts of potash and phosphoric acid in the soils, according to the manures employed. They further brought out the interesting fact, that comparatively very little of the applied potash or phosphoric acid had gone below the first 9 inches of soil, and that certainly none had gone into the third depth. In our own country, for some years past, Dr Bernard BAELET. 95 Dyer has been investigating the subject of " The ancdytical Dyer's determination of prolally available 'mineral' plant-food in Sto^""^ soils "; ^ and, at the suggestion of Professor Armstrong, one sted sails. of the Eothamsted Trust Committee, he asked whether we could supply him, for the purposes of his investigation, with samples of soUs from some of the experimental fields at Eothamsted, of which the manure and crop history was known. Accordingly, in 1889, we gave him facilities for taking samples of the surface-soil, to a depth of 9 inches, from twenty-two of the plots in the experimental barley-field ; and we also provided him with samples which had been col- lected in 1882, from a few selected plots, to the depth of three times 9 inches. In all these samples Dr Dyer has , determined the total potash, by acid, fusion, &c. ; the amount dissolved by hydro- chloric acid, and the amount taken up by a 1-per-cent citric acid solution ; also the amounts of phosphoric acid, by hydro- chloric acid, and by a 1-per-cent solution of citric acid. Dr Dyer's results, obtained on the surface-soils of the series of twenty-two plots, show at a glance comparative exhaustion or accumulation of both potash and phosphoric acid, whether hydrochloric acid, or the dilute citric acid solution, was used. There are, indeed, among these numerous results, some appa- rently inconsistent quantitative indications; but these are probably attributable to irregularities in the soils themselves. Difficulty and therefore to the difficulties of sampling, rather than to J^"™-^ those of analysis. It will be useful to refer a little more in detail to the re- sults obtained on the soils of plot 2a and plot 4a; the manure and' crop history of which has been pretty fully illustrated by the results given in Tables 29 and 30, and the discussion of them. It would appear that not more than two-thirds of the potash estimated to be accumulated where it was supplied, SoUaceum- was taken up by hydrochloric acid ; but that approximately "^Xalw? the whole of the accumulated phosphoric acid was so taken phosphoric up. Hence it may be judged that much of the residue of the '^"^■ supplied potash had gone into more fixed combinations within the soil than was the case with the phosphoric acid. Then as to the citric acid results, it may be observed that they are so far accordant that the sample of the surface-soil of the potash- exhausted plot taken in 1882 showed more potash than that taken in 1889, when the exhaustion was of course greater. Again, the citric acid determinations on the soil with potash-supply showed more so taken up from the 1889 than from the 1882 sample ; the accumulation having ' Trans. Chem. Soc, 1894, p. 115. iSee also the discussion on his paper, Froc. Chem. Soc, No. 134 (1893-94), p. 37. 96 THE EOTHAMSTED EXPERIMENTS. been the greater at the later date. It is also of interest to observe that the amounts determined in the potash-exhansted soil by the 1-per-cent citric acid solution were about from three to five times as much as the crops would annually take up, which is a fairly consistent relation. Further, with reference to these barley-soil results, as superphosphate was applied to both plots, the comparison of the amounts taken up on the two is of less interest than in the case of the potash; but comparison with the results obtained on another plot, otherwise similarly manured, but without superphosphate, shows, as already referred to, that the estimated accumulation of phosphoric acid was approxi- mately indicated by the amount taken up by hydrochloric acid. The results relating to the two plots are, however, of special interest as illustrating, in the one case actual exhaus- Acawmvia- tion, and in the other actual accumulation of potash, there **T "h being in the one a loss over the forty years of about 1018 lb. shown }yy of the potash of the soil, and in the other a gain from supply smi-cmaiy- ^f ^ijout 3180 lb. ; whilst of the latter amount the results show that hydrochloric acid extracted nearly two-thirds, and citric acid less than one-fourth. It is further of interest to note that Dr Bernard Dyer's results,' obtained on the 1882 Potash mid samples from the two plots, in each case to the depth of three ^^id'k^ times 9 inches, agree with those formerly obtained by Her- toiheswr- mann von Liebig on the wheat-field soils, in showing that face. little if any of either the potash or phosphoric acid artificially supplied had gone below the first 9 inches of depth. AnaZysisof Dr Dyer is also working on the soils of some of the plot's wheat-soil. q£ ^}jg experimental wheat-field, and these will afford some striking illustrations in regard to the condition of availability of accumulated residue of potash-supply over a long series of years. Thus there is a series of plots which have received the same amount of ammonium - salts and superphosphate each year for forty years, to 1891 inclusive ; one of which has received no potash either during those forty years, or during the eight preceding years ; two received potash during the first eight years, but none since ; and one, besides receiving potash during the first eight years, has received it each year since. The complete manure and crop history of each of the four plots is, so far as potash and phosphoric acid are con- cerned, available for each of the four ten-yearly periods of the forty years — as in the case of plots 2a and 4a in the barley-field. The amount and composition of the crops show great reduction in produce and exhaustion of potash, where none had been applied from the beginning; less reduction, and less exhaustion, where there was a residue of potash from the applications during the first eight years ; and lastly, main- BAELET. 97 tenanee of produce, and great accumulation of potash in the crops, where potash has heen annually applied. Further, the indication is, that the whole of the residue of potash sup- plied during the first eight years on the plots where none has been applied since, has been approximately exhausted during the succeeding forty years. It is obvious, therefore, that Dr Dyer will find new points of interest in the investigation of the experimental wheat-field soils ; for the results will afford illustrations, not only of mere exhaustion and accumulation, but of effective residue as well. On what does Strength of Straw Depend 1 It will be appropriate to refer here to the bearing of ex- SUica and perimental evidence on the question whether, as is frequently f^^'^"-^ stated, strength of straw is dependent on a high percentage of silica. Table 31 (p. 98) affords illustrations on this point. TaiieSi The upper division of the table gives results relating to «^^*'^^- wheat, and the lower corresponding results relating to barley. In the case of wheat five, and in that of barley three, very different conditions of manuring are selected for illustration ; and, for each condition as to manuring, results obtained in bad and in good seasons are given. The particu- lars indicating the character of the crops are — the percentage of grain in the total produce, and the weight per bushel of the dressed grain; and, side by side with these are- recorded — the percentage of ash in the dry matter of the straw, the percentage of sUica in the ash, and the percentage of silica in the dry matter. In the wheat in every case, and in the barley in every case Season and but one, there is a higher proportion of grain in the better p'^°'^^^- season ; and in every case, of both wheat and barley, there is a much higher weight per bushel of grain in the better season. These conditions are, in fact, proof of the superiority of the crops in the main characters of seed-forming tendency, and ripening. The percentage of ash in the dry matter of the straw is not /Season and a very significant character ; and it is seen that in the case '^'^ ™ of the wheat it was on the average somewhat the lower, but in that of the barley uniformly the higher, in the better seasons. The percentage of silica in the ash of the straw is more Silica in significant; and in both the wheat and the barley it is, ««*''«'^ ^, ,,1 f. p • Till ',1 a/n/ matter under all the conditions of manuring, much the lower m the of straw. : better seasons. More significant still is the percentage of silica in the dry matter of the straw ; and it is seen that with the wheat under each condition of manuring, and .with VOL. vn, a 98 THE EOTHAMSTED EXPERIMENTS. the barley under most conditions, it is considerably lower in the better seasons. It may be observed that the exceptions in the case of the barley were, where organic manure, as in rape-cake and farmyard manure, was employed. TABLE 31. Per cent grain in total produce. Weight per bushel of dressed grain. Per cent ash in ■dry matter. Per cent silica in WHEAT. BAELEY. Per cent silica in dry matter. Without manure (1856 ■,1858 36.4 40.6 54.3 60.4 5.5 4.9 71.47 65.85 3.93 3.23 Ammonium-salts alone (1856 tl858 34.8 40.3 55.5 59.6 3.9 4.0 66.23 57.47 2.58 2.30 Mixed mineral manure (1856 11858 36.7 43.6 56.4 61.5 5.7 5.6 68.74 64.67 3.92 3.62 Mineral manure and amm.-salts (1856 11858 33.6 38.2 58.0 62.2 4.9 5.0 64.63 55.60 3.17 2.78 ■Farmyard man- ure (1856 ': 1858 34.5 39.6 58.6 62.6 6.7 6.54 69.56 59.71 4.66 3.90 Rape-cake . (1852 11871 44.3 45.4 51.7 56.3 4.75 6.64 57.49 42.04 2.73 2.33 Eape-cake . (1856 11863 39.1 48.4 46.1 56.3 4.63 5.17 49.39 45.62 2.29 2.36 Mineral manure and amm.-salts (1852 11871 43.2 43.3 51.4 56.5 4.19 6.70 62.21 32.71 2.61 2.19 Mineral manure and amm.-salts (1856 11863 40.2 47.3 46.4 56.5 5.48 6.32 57.47 35.24 3.15 2.23 Farmyard man- ure (1852 11871 47.0 43.8 52.8 56.6 5.15 7.55 57.38 42.71 2.96 3.22 Farmyard man- ure (1856 ". 1868 42.8 48.3 47.1 57.2 . 4.92 6.21 57.86 43.08 2.85 2.68 Season anct^ Direct analytical results clearly show, therefore, that the itrnviT /P™portion of silica is as a rule lower, not higher, in the straw of the better grown and better ripened crops. This result is quite inconsistent with the usually accepted view that high quality and stiffness of straw depend on a high amount of silica. Pierre and Bretschneider have, indeed, concluded from their experiments that this is not the case, and at Eothamsted we have long maintained a contrary view, In fact, high proportion of silica means a relatively strength of straw not upon BAHLET. 99 low proportion of organic substance produced. Nor can there Woodi/ be any doubt that strength of straw depends on the favour- XeS*"?/ able development of the woody substance ; and the more this straw. is attained the more will the accumulated silica be, so to speak, diluted — in other words, show a lower proportion to the organic substance. It may be mentioned that in our own neighbourhood, where the straw-plait industry prevails, the complaint during seasons of bad harvests has been that an unusually large proportion of the straw was brittle and broke in the work- ing ; and considering the character of the seasons, there can be no doubt that this was associated with low development of the woody matter, and high proportion of silica. Summary and Conclusions. We have now illustrated the influence of exhaustion, of manures, and of variations of season, on the amounts of produce, and on the composition, of barley. The results have shown that on the growth of barley for Summwry more than forty years in- succession on rather heavy ordinary "f'^^^^- arable soil, the produce by mineral manures alone was higher than that without manure ; that nitrogenous manures alone gave more produce than mineral manures alone; and that mixtures of both mineral and nitrogenous manure gave much more than either used alone — indeed generally twice, or more than twice, as much as mineral manures alone. Of imineral constituents, whether used alone or in mixture with nitrogenous manures, phosphates were much more effective than mixtures of salts of potash, soda, and magnesia. The averages show that, under all conditions of manuring (except- ing with farmyard manure) the produce was less over the later than over the earlier periods of the experiments, a result partly due to the seasons. But the average produce for the Most effect- forty years of continuous growth of barley was, in all cases ^J^"" where nitrogenous and mineral manures (containing phos- bm-leif, phates) were used together, much higher than the average produce of the crop grown in ordinary rotation in the United Kingdom, and very much higher than the average in most other countries when so grown. It is seen that the requirements of barley within the soil. Barley amd and its susceptibility to the external influences of season, are '^^"*' '"™" very similar to those of its near aUy, wheat. There are, how- ' ever, distinctions of result dependent on differences in the ■habits of the two plants, and in the conditions of their culti- vation accordingly. "Wheat is with us, as a rule, sown in the autumn, on a 190 THE BOTHAMSTED EXPEKIMENTS. Barley a snrface- Root-range heavier soil, and has four or five months in which to dis* of wheat. j[,£})yxte its roots, and so gets possession of a wide range of soil and subsoU, before barley is sown. Barley is sown in a lighter surface-soil, and, with its short >p,eriod for root-development, relies in a much greater degree on the stores within the surface-soil. Accordingly, it is more susceptible to exhaustion of surface-soil as to its nitro- genous, and especially as to its mineral, supplies; and in the common practice of agriculture it is found to be more benefited by direct mineral manures, especially phosphatic manures, than is wheat when sown under equal soil conditions, The exhaustion induced by both crops is, however, char- acteristically that of available nitrogen; and when, under 'm!,dbmUy. the ordinary conditions of manuring and cropping, artificial manure is still required, nitrogenous manures are, as a rule, requisite for both crops, and for the spring-sown barley, superphosphate also, Soils for ■ ) liastly, although barley is appropriately grown on lighter w1^* ""'^ ^°^^^ *^^^ wheat, good crops, of fair quality, may be grown on the heavier soils after another grain crop, by the aid of artificial manures, provided that the land is sufficiently clean. requisite for wheat SECTION III.— EXPERIMENTS ON THE GROWTH OF VARI- OUS LEGUMINOUS CROPS FOR MANY YEARS IN SUCCESSION ON THE SAME LAND; ALSO ON THE QUESTION OF THE FIXATION OF FREE NITROGEN, Inteoduction. We now come to the third element of the ordinary four- course rotation — namely. Leguminous Crops, which, indeed, have a place in most other rotations also. Character- It is found that, within certain, limits, the requirements, istics of anji tiie results of growth, of different members of one and cfo^ps. the same family show certain characteristics in common ; whilst those of different families show more or less of dis- tinctive character. Nevertheless there are some important points of similarity, as well as of contrast, between the re- quirements of the agricultural representatives of the Grar minese, the Cruciferse, the ChenopodiaceES, and the Solan'^ae. It will be seen, however, that the agricultural representa- tives of the Legi:iminos8e, all of which are included in the sub-order PapilionaceEE, and some of which are of much im- portance in our agriculture, show very marked differences as compared with those of any of the other Orders above enumerated, , , , . ■; LEGXJMINOUS CROPS. lOl It SO happens that both the scientific interest and the Legumm- practical value of these crops; whether as elements in rota- ^^-^^^ tion, or as grown in the mixed herbage of graSs-land, depend gm. very largely on the amount of nitrogen which they contain, and on the sources of their nitrogen ; and especially on the great differences in these respects between them and the representatives of the other Orders with which they are grown, either in alternation in our rotations, or in, association in our meadows and pastures. So much is this the case, that it is essential to a proper understanding and appreciation of the characteristics of growth of these crops, and for theillustration of their value and importance as depending on those characteristics, to compare and to contrast the conditions and results of their growth with those of the crops of other Orders.. We wiU, therefore, first briefly call attention to the differ- ence in the amounts of nitrogen assimilated over a given area by different crops when each is grown for many years in succession on the same land, without any nitrogenous manure— that is to say, under conditions in which the soil is to a great extent exhausted of accumulations of nitrogen due to recent supplies by manure, and when, therefore, the plants have to rely largely on what may be called the natural resources of the soil, and on those of the atmosphere. Yield of Nitrogen ;per acre in different Crops. Table 32 (p. 102) shows the yield of nitrogen per acre per Yield of annum, with mineral, but without any nitrogenous manure — dmS^ in wheat and in barley as gramineous crops, in turnips as crops. representatives of the Cruciferae, in sugar-beet and mangel-, wurzel of the Chenopodiaceae, and in beans and clover as leguminous crops, when each is grown for many years in succession on the same land. Incidentally it is to be noticed that in the case of each of GraOmi the crops — wheat, barley, and ieans — thus grown year after 3^^l^ year on the same land for many years in succession without nitrogenous manure, there was a reduction in the yield of nitrogen per acre per annum over the second period com- pared with the first ; that is, as the previous accumulations within the soil became reduced. Disregarding this tendency to reduced yield, it is seen Yield of _ that over the same period of 24 years, with full mineral but ^^f!!lT without nitrogenous manure, the wheat yielded an average of barley. 22.1 lb., and the barley 22.4 lb. of nitrogen per acre per annum ; the two allied crops, therefore, yielding almost identical amounts in their above-ground produce without 102 THE EOTHAMSTED EXPERIMENTS. nitrogenous manure, on soil very poor in available nitrogen, so far as accumulations due to recent applications of nitro- genous manure are concerned. TABLE 32. — Nitkogen pee aoee pee anndm, in vabious Ceops GEOWN AT ROTHAMSTED, WITH MINERAL BDT -WITHOUT NlTEO- GENOUS Manure. Tidd of mtrogen in root-crops. Wheat Barley Eoot-crops Beans Glover ^Swedisli turnips I Sugar-beet I Mangels Total . Duration of experiment. 63 75 12 years, 1852-63 12 years, 1864-75 24 years, 1852-75 12 years, 1852- 12 years, 1864- 24 years, 1852-75 *15 years, 1856-70 5 years, 1871-75 10 years, 1876-85 30 years, 1856-85 12 years, 1847-58 tl2 years, 1859-70 24 years, 1847-70 t22 years, 1849-70 * 13 years, 2 years failed. + 9 years beans, 1 year -wheat, 2 years fallo-w. j 6 years clover, 1 year -wheat, 3 years barley, 12 Average nitrogen per acre per annum. lb. 27.0 17.2 22.1 26.0 18.8 22.4 18.5 14.7 14.0 16.4 61.5 29.5 45.5 39.8 years fallow. Turning no-w to the yield of nitrogen in the root-crops — turnips, sugar-beet, and mangel- -wurzel — it may be mentioned that, prior to the period referred to in the table, turnips had been gro-wn for a number of years, and had yielded 42 lb. of nitrogen per acre per annum, due to the accumulations from comparatively recent nitrogenous manuring. But it is seen that after these accumulations had been reduced, s-wedish turnips gave, over 15 years, an average of only 18„5 lb.; sugar-beet over the next 5 years, an average of only 14.7 lb. ; and mangel--wurzel over the succeeding 10, years, an : average of only 14.0 lb. of nitrogen per acre per annum. Or, reckoned LEGUMINOUS CROPS. ' 103 over the whole period of 30 years, after the recent accumu- lations had been worked out, the root-crops gave an average of only 16.4 lb. of nitrogen per acre per annum. It is remarkable how very similar is the amount of , nitrogen annually accumulated in gramineous, cruciferous, ™«™'«*»* O of ThltTOQ&Hb and chenopodiaceous crops, after the soil had been exhausted m grain of the more recent and more readily available nitrogenous "■ndnot- accumulations. Thus, over the second half of the period, the "^ wheat gave 17.2 lb., and the barley 18.8 lb., against 16.4 lb. over 30 years in the various root-crops. We now conie to the yield of nitrogen in leguminous crops. Yield of Eeferring first to the results obtained with beans, it is seen ™*™5'f» ™ that over the first half of the period of 24 years, the average crops. annual yield of nitrogen in the crop was 61.5 lb. per acre ; whilst over the second 12 years — in 3 of which the crop failed, so that there were only 9 years of beans, one of wheat, and two of fallow — the annual yield was less than half as much, or only 29.5' lb. per acre. Nevertheless, the average yield over the 24 years without any nitrogenous manure, was 45.5 lb. per acre per annum. That is to say, under very similar conditions as to soil-supply, the highly nitrogenous leguminous crop, beans, has yielded over a given area twice as much nitrogen as either wheat or barley, and more than twice as much as the root-crops; The last results in the table relate to the leguminous crop oicmer sick- — clover. It is well known that clover faUs when it is "^**' attempted to grow it too frequently on the same land ; and, in the case recorded in the table, it happened that clover was obtained in only 6 years out of the 22 for which the yield of ridd of nitrogen is given ; so that there are included, owing to the ?/o™^f^g^ failures, 1 year of wheat, 3 of barley, and 12 of fallow. '^^' Notwithstanding this, there was, with the occasional inter^ polation of the clover, an average yield over the 22 years of 39.8 lb. of nitrogen per acre with mineral, but without nitro- genous supply. The next illustrations show more strikingly still the riddsof greater yield of nitrogen in leguminous than in gramineous ^^'''''^j" , crops, when grown under equal soil conditions. They relate cmd clover to the yield of nitrogen in barley and in clpver, grown side \)y..ocmpa/red. side in the same field ; and the results are given in Table 33, The field had grown one crop of wheat, one of oats, and three of barley in succession, w;ith artificial mineral and nitro- genous manures ;. but 'vs^ithout any farmyard or other organic manure. In 1872 barley was again sown ; on one half alone, and on the other half with clover. In 1873 barley Vas again grown on the one half, but the clover on the other; 104 THE EOTHAMSTED EXPERIMENTS. The table shows that the barley yielded 37.3 lb. of nitrogen per acre, whilst the three, cuttings of clover contained 151.3 lb. In the next year, 1874, barley was grown over both portions ; and on the one where barley had yielded 37.3 lb. of nitrogen in the previous year, it now yielded 39.1 lb. ; but on the portion where the clover had yielded 151.3 lb., the barley succeeding it yielded 69.4 lb. That is to say, the barley yielded 30.3 lb. more nitrogen after the removal of 151.3 lb. in clover/than after the removal of only 37.3 lb. in barley. TABLE 33. — Nitrooen per acee in Barley and in Clover, GROWN IN Little Hoosfield, Eothamsted. Nitrogen per acre. isv3iS :::::::: ,3,jB-ieyl^rjK : : : : : lb. 37.3 151.3 39.1 69.4 V. Barley after clover more than, after barley 30.3 Glover en- The fact is, that the clover had not only yielded so much ricking soil more nitrogen in the removed crops, but it had also left the mm rogen. gm-fa^g.gQii considerably richer in nitrogen. Thus in October 1873, after the removal of the barley and the clover, samples of soil were taken from ten places on each of the two portions, and the nitrogen was determined in the samples — ;from each of four of the individual holes separately, in the mixture of the four, and in the mixture of the samples from the other six .places. The determinations in the numerous separate Samples consistently showed that, to the depth of 9 inches, the clover-land-soil, which had yielded so much more nitro- gen in the crops, was nevertheless determinably richer in nitrogen than .the barley-land-soil, which had yielded so much less. This is sufBciently illustrated by the following figures, showing the mean percentage of nitrogen in October 1873, in the fine dry. soil, of the clover-land, and of the barley-land, respectively: — Mean per cent nitrogen. In clover-land-soil .... 0. 1566 In barley- land-soil .... 0.1416 This was the case notwithstanding that all visible vegetable dihris had .first been removed from the samples. It was, LEGUMINOUS CROPS. 105 further found that the ahove- and under-ground vegetable residufe picked from the clover-land samples was inuch more in quantity, and contained much more nitrogen, than that from the barley-land samples. In 1874, and in 1875, barley only was sown over both por- Further tions. In 1876, barley was again sown over the whole of the ^^^ land, with clover as well on the portions where it had grown in 1873 ; but the plant failed in the winter, and gave no crop in 1877. In 1877, barley was again sown over the whole ; this time with clover on half of the previously clover portion, and on half of the previously only barley portion. In the autumn of 1877 soil-samples were again taken; this time from four places on each of the differently cropped portions. The determinations of nitrogen in the surface-soils consistently showed, as before, a higher percentage where clover than where only barley had grown. It is, of course, well known in agriculture, that the growth mtrogm of clover, which removes much more nitrogen than a cereal *'' legumm- crop, increases the produce of a succeeding cereal as if nitro- gramineous genous manure had been applied. But attention is specially '^P^- to be directed to the fact, that a leguminous crop accumulates a great deal more nitrogen over a given area than a gramin- eous one under equal soil-conditions. But not only is the yield of nitrogen per acre much less in the cereal crops, but the percentage of nitrogen in the dry substance of the gramineous produce is much less than in that of the leguminous produce. The corn of the leguminous crops — beans and peas, for example — contains more than twice as high a percentage of nitrogen in its dry substance as that of the gramineous grains. The dry substance of leguminous straws also contains about twice as high a percentage of nitrogen as that of cereal straws. Again, the dry substance of clover-hay contains not far short of twice as much nitrogen as that of meadow-hay. Lastly, the dry substance of roots contains about the same percentage of nitrogen as that, of the cereal grains, but only about half as much as that of the leguminous corn. The leaves of the root-crops are, however, high in nitrogen. The general result is, then, that the nom-leguminous crops, especially those of the gramineous family, are characterised, both by yielding much less nitrogen in their produce over a given area, and by containing a much lower percentage of nitrogen in their dry substance, than the leguminous crops. Bearing these facts in mind, let us now turn to the consider- ation of the effects of direct nitrogenous manures on the various crops. 106 THE EOTHAMSTED EXPERIMENTS. Effects of ovs man- v/res lyoon iiwrious craps. Table 34 Method of calcula- tion. Effects of Nitrogenous Manures in increasincf the Produce of various Crops. It is fully recognised that, under the conditions in which the crops are grown in ordinary agriculture, nitrogenous manures have very marked effects in increasing the amounts of produce of wheat, of barley, of turnips, of mangels, and of potatoes — that is, of the comparatively low-in-nitrogen wow- leguminous crops. It is to be borne in mind, too, that in the case of wheat and barley the increased produce consists characteristically of the non-nitrogenous substances starch and cellulose, in that of the root-crops of the non-nitrogenous substance sugar, and in that of potatoes of the non-nitrogen- ous substance starch. The influence of nitrogenous manures in increasing the production of the non-nitrogenous constituents of our crops is very strikingly illustrated by the results given in Table 34. ' The first column of figures shows — the estimated amounts of carbon per acre per annum, in the total produce of wheat and of barley, in the roots of sugar-beet and mangel-wurzel, in the tubers of potatoes, and in the total produce of beans, when each is grown by a complex mineral manure without nitrogen, and also with the same mineral manures with nitrogenous manure in addition. The second column shows the estimated gain of carbon — that is, the increased amount of it assimilated under the influence of the nitrogenous manures. The third column gives the estimated increased production of total car- bohydrates, under the influence of the nitrogenous manures ; and the last column the estimated gain of carbohydrates for 1 of nitrogen in manure. The calculations are based on the average produce by the different manures, of wheat over 20 years, of barley over 20 years, of sugar-beet over 3 years, of mangel-wurzel over 8 years, of potatoes over 10 years, and of beans over 8 years. The mode of calculating the amounts of carbon and of carbokydraCes is as follows : From the amount of dry sub- stance in the crops, the amounts of mineral matter and of nitrogenous substance are deducted; and the remainder rep- resents the amount of carbohydrates. The amount of carbon in the nitrogenous substance is calculated, and then that in the carbohydrates, on the assumption that, in the wheat, barley, and beans, starch and cellulose are the main products ; in the sugar-beet and mangel-wurzel, cane-sugar, pectine, and cellulose ; and in the potatoes, starch^ and cellulose. Such estimates can, obviously, be only approximations, to the truth ;' but, accepted as such, they are useful, as conveying some lEGUMINOITS CEOPS. 107 definite impression of the influence of nitrogenous manures on carbon-assimilation, and on carbohydrate-formation. TABLE 34.— Estimates op the Yield and Gain ow Carbon, and ow THE Gain of Caebohydeates, per acee per anndm, in vahiotis' Experimental Crops grown at Kothamsted. Carbon. Carbohydrates. Actual. Gain. Gain. For 1 nitrogen in manure. 1 WHEAT 20 TEARS, 1852-71. lb. lb. lb. lb. Mineral manure ■. 988 Mineral manure and 43 lb. nitrogen as ammonia . 1590 602 1240 28.8 Mineral manure and 86 lb. nitrogen as ammonia . 2222 1234 2550 29.7 Mineral manure and 86 lb. nitrogen as nitrate . 2500 • 1512 3140 36.5 BARLEY 20 YEARS 1862-71. Mineral manure . . Mineral manure and 43 lb. nitrogen as ammonia , 1138 2088 950 1992 ) 46.3 SUGAE-BEET 3 TEARS, 1871-7. . Mineral manure . . . . Mineral manure and 86 lb. nitrogen as ammopia . Mineral manure and 86 lb. nitrogen as nitrate . 1123 2600 3031 1477 1908 3188 4052 37a 47.1 MANGEL-WURZEL S YEARS, 1876-83. Mineral manure Mineral manure and 86 lb. nitrogen as ammonia . Mineral manure and 86 lb. nitrogen as nitrate . 759 1889 2129 1130 1370 2376 2771 27.6 32.2 POTATOES 10 TEAES, 1876-86. Mineral manure ....... Mineral manure and 86 lb. nitrogen as ammonia . Mineral manure and 86 lb. nitrogen as nitrate 1021 1783 762 1507 1752 731 1416 17.5 16.5 BEAKS 8 TEARS, 1862 AND 1864-70. ) Mineral manure . . . . . . Mineral manure and 86 lb. nitrogen as nitrate . 726 992 266' 474 s's ~ It is thus seen that, independently of the underground growth, the wheat was estimated to assimilate 988 lb. of 108 THE EOTHAMSTED EXPEEIMENTS. Yield of carbon per acre per annum, under the influence of a complex "and"!^^^ mineral manure alone ; and that the amount was increased to outnitro- 1590 lb. by the addition of 43 lb. of nitrogen as ammonium- mawe ' ^^^^^' *° ^^^^ ^^- ^J ^^ ^^- °^ nitrogen as ammonium-salts, and manure.. ^^ ^ggo lb. by 86 lb. of nitrogen as sodium-nitrate. Accord- ingly, as shown in the second column, the increased assimila- tion of carbon was — by 43 lb. of nitrogen as ammonium-salts '602 lb., by 86 lb. as ammonium-salts 1234 lb., and by 86 lb. as sodium-nitrate 1512 lb. Eeckoned in the same way, the increased assimilation of ' carbon in the barley was, for 43 lb. nitrogen as ammonium- salts 950 lb. per acre — that is, one and a-half time as much as by the same application in the case of wheat. In the sugar-beet, the roots only (the leaves being left on the land), the increased assimilation of carbon was 1477 lb. per acre by the application of 86 lb. nitrogen as ammonium- salts,; and 1908 lb. by 86 lb. nitrogen as sodium - nitrate. Ther^ was, therefore, considerably more increased assimilation of carbon, and accumulation of it in the roots of the sugar-beet, .. .than in the grain and straw of wheat, by the same applications of nitrogenous manure. ■ In mangel-wurzel roots (the leaves beiiig returned to the land), the increased assimilation of carbon was 1130 lb. by 86 lb. of nitrogen as ammonium-salts, and 1370 lb. by 86 lb. as nitrate — that is, less than in the removed crops (grain and straw) of wheat, and considerably less than in the removed crops (the roots) of sugar-beet. In the potatoes, reckoned on the increased production of tubers only (the tops being left on the land), the increased yield of carbon by 86 lb. of nitrogen as ammonium-salts was 762 lb. per acre, and by 86 lb. as sodium-nitrate 731 lb. — that is, there was considerably less increased production of starch in potatoes than of sugar in either sugar-beet or mangel- wurzel by the same applications of nitrogenous manure. Lastly, in the leguminous crop — beans, with its high yield — of nitrogen per acre, and the high percentage of nitrogen in its dry substance-T-the increased assimilation of carbon under the influence of nitrogenous manure was comparatively quite in- significant. Thus there was, by the application of 86 lb. of nitrogen as sodium-nitrate, an increased assimilation of carbon of only 266 lb. per acre, or little more than one-sixth as much as in wheat, and little more than one-eighth as much as in Tield of . sugar-beet, by the same application. "^^twith Turning to the figures in the third column, it is seen that mdimth- there was a very greatly increased production of the non- nitrogenous bodies, the carbohydrates, by the use of nitrogen- out mtro- manure. ous manures. LEGUMINOUS CROPS. 109 Thus, by the use of 43 lb. of nitrogen as ammonium-salts, there was an estimated increase of 1240 lb. of carbohydrates in the grain and stra'wj of -wheat, and of 1992 lb. in those of barley. By the application of 86 lb. of nitrogen as ammonium- salts, there was an increased formation of 2550 lb. of carbo^ hydrates in wheat, of 3188 lb. in sugar-beet, of 2376 lb. in mangel-wurzel, and of only 1507 lb. in potatoes ; and when 86 lb. were applied as sodium-nitrate, there was an increased production of 3140 lb. in wheat, of 4052 lb. in sugar-beet, of 2771 lb. in mangel-wurzel, and of only 1416 lb. in potatoes. Whilst, compared with these amounts, there was by the same application, an increase of only 474 lb. of carbohydrates in beans. The last column shows the estimated increased amounts of carbohydrates produced for 1 of nitrogen in manure, in the different cases. Thus, when 43 lb. of nitrogen were applied as ammonium-salts, 1 lb. of nitrogen in manure gave an in- creased production of 28.8 lb. of carbohydrates in the grain and straw of wheat, and of 46.3 lb. in those of barley ; when 86 lb. nitrogen were applied as ammonium-salts, 1 lb. gave an increase of 29.7 lb. carbohydrates in wheat, 37.1 lb. in the roots of sugar-beet, 27.6 lb. in those of mangel-wurzel, and 17.5 lb. in potatoes. Again, when 86 lb. were applied as sodium-nitrate, 1 lb. gave an increase of 36.5 lb. carbohy- drates in wheat, 47.1 lb. in sugar-beet, 32.2 lb. in mangel- wurzel, 16.5 lb. in potatoes, and only 5.5 lb. in the legumin- ous crops — ^beans. It is natural to ask. What is the explanation of the appar- Seemingly ently anomalous result, that the crops which are charac- '^^^"^ terised by containing comparatively little nitrogen, and by plained. yielding large amounts of non-nitrogenous products — starch, sugar, and cellulose — are especially benefited by the applica- tion of nitrogenous manures ; and that, under their influence, they yield greatly increased amounts of those non-nitrogenous bodies ? It is, perhaps, little more than stating the facts in another way to say, as is the case, that the luxuriance or activity of growth of all these crops is very greatly enhanced by nitro- genous manures ; and that, since their special products are these non-nitrogenous substances, the natural result of the increased luxuriance is to increase the formation of the bodies which are their essential or characteristic products. A further possible explanation of the curious result has, however, been suggested.^ Thus, on purely chemical and physiological grounds, and 1 See yines'. Lectures on We Physiology of Plants, p. 140 et seq. 110 THE EOTHAMSTED EXPERIMENTS, Vines' mews. An cmal- ogy from world. ^so far as would appear without any special reference to the fact that, in the case of our chief starch- and sugar-yielding crops, the production of those substances is greatly enhanced -by the use of nitrogenous manures, it has been suggested that the substance first formed in the chlorophyll-corpuscle from carbon dioxide and water is not starch, but a substance possibly allied to formic aldehyde {GB.fi), which goes to construct proteid, by combining with the nitrogen and sul- phur absorbed in the form of salts from the soil, or with the nitrogenous residues of previous decompositions of. proteid. It is supposed, however, that starch may nevertheless be the first visible product of the constructive metabolism ; since, unless protoplasm were being formed, no starch could be produced. This view is partly founded on the consideration of the analogy that would then be established between the forma- tion of starch and that of the carbohydrate — cellulose, which is by some experimenters supposed to be derived directly from protoplasm. It is true that such a supposition is at any rate not incon- sistent with the conditions which we have seen to be favour- able for the increased production of starch and sugar in agricultural plants. At the same time, it is admittedly at present little more than hypothesis. It would, indeed, re- quire more evidence than is at present available, to establish such a conclusion; whilst there are considerations which would lead us to hesitate to adopt the view in question with- out clear experimental proof. Thus, it seems difScult to suppose that the undoubted con- nection in some striking cases between the amount of nitro- gen taken up by the plant, and the amount of starch or sugar formed, is to be explained by an assumption which implies that a chief office of the nitrogenous bodies of plants is to serve as intermediate only, in the transformations necessary for the formation of the non-nitrogenous substances. The view does not, however, assume that nitrogen is eliminated from the plant in the process, and so lost. Then, again, plants, such as many of the Leguminosee, which are character- ised by assimilating relatively very large amounts of nitrogen over a given area of land, and by the formation of very large amounts of proteid in proportion to plant surface,' produce relatively small amounts of the carbohydrates. Nor is it irrelevant to refer to the fact that, from theo- retical considerations, it was for many years assumed, espe- cially in Germany, in opposition to the teachings of our own numerous direct experiments, that in the animal body the non-nitrogenous , substance — fat^-was mostly, if not always, LEGUMTNOUS CROPS. Ill produced by the degradation of proteid ; the nitrogenous by- products being for the most part, if not entirely, eliminated from the body as weiste matter. It is, however, now indubit- ably established, at any rate in the case of the herbivora ■n-hich produce the most fat, that that substance is derived largely, if not exclusively, from the non-nitrogenous constitu- ents of the food — the carbohydrates. In the case of the supposed transformation in plants, the same prodigal expenditure of the nitrogenous bodies in the formation of the non-nitrogenous is, however, as has been said, not involved. Effects of Nitrogenous Manures on Leguminous Crops. We have now to illustrate the influence of nitrogenous manures on various leguminous crops which, as has been pointed out, are characterised by containing a high percent- age of nitrogen in their dry substance, and by assimilating a large amount of nitrogen, from some source, over a given area of land. It will be seen that the results bring to view some very remarkable failures, but also some not less signal and significant successes. Our first illustrations relate to experiments with bearis, Effects of grown for many years in succession on the same land, with- «*™^««<«« out manure ; with a purely mineral manure (consistmg of on leans. superphosphate, and salts of potash, soda, and magnesia); also with the same mineral manure, and nitrogenous manure in addition, suppUed either as ammonium-salts or as sodium- nitrate. Table 35 (p. 112) gives a summary of the results obtained under each of the three conditions as to manuring over a period of 32 years of continued or interrupted experi- ments, from 1847 to 1878 inclusive. The upper division gives the average amount of total produce (corn and straw together) per acre per annum, over each of the four 8-yearly periods, and over the total period of 32 years. But, as there were frequent failures of crop, the lower division of the table gives the average produce per acre per annum over the years of crop only during each period. Before referring to the figures, it should be explained that Nitrates in the first 5 years the nitrogen applied to the third plot was ^^*^' in the form of ammonium-salts. The effects were, however, ammon- so small and irregular, that the application of nitrogenous ^^'"■■saUs. manure was then suspended for some years — ^indeed for 10 years ; after which, it having been observed that nitrates were more beneficial to Leguminosse than ammonium-salts, sodium -nitrate was applied instead ; in amount supplying 86 lb. nitrogen per acre per annum, or nearly twice as much 112 THE EjOTHAMSTED EXPEKIMENTS. as had been given as ammonium-salts in the earlier years. This application of the nitrate commenced in 1862, and with some breaks owing to severe or wet winters, which prevented the seed being sown or destroyed the plant, it was continued up to 1878, when the experiments were finally abandoned. TABLE 35.^Beans. . Ayerage Produce per acre per annum in lb. Total produce (corn and straw). Unmanured. Mixed mineral manure (including poliasb). Mixed mineral manure and nitrogen. Average pee acre per annum, over each 8 years, and over the 32 years, 8 years, 1847-54 8 years, 1855-62 8 years, 1863-70 8 years, 1871-78 lb. 2421 1664 606 864 lb. 32081 2466 1622 1506 lb. 3555 2629 2198 •1646 32 years, 1847-78 1389 21682 2507 Average per acre per annum, over the years of crop only, each period. 1st 8 years, 8 crops . 2nd 8 years, 7 crops . 3rd 8 years, 7 crops . 4th 8 years, 4 crops . 32 years, 1847-78, 26 crops 2421 1902 692 1729 1709 32083 2818 1854 3011 2688* 3555 3005 2513 3292 3086 crops grown at short m- teniaU. 1 7 years, excluding 1849, in which year the produce was accidentally not weighed. 2 31 years, excluding 1849. 3 7 crops, excluding 1849. 4 25 crops, excluding 1849. Faiiwre of The Occasional entire failures above referred to as mainly due to adverse seasons, were also materially dependent on the conditions induced in the land by the continuous cropping with this plant; which, as is the case with most Leguminosse, is very susceptible to parasitic attacks of various kinds when the conditions of growth are not normal and favourable. Indeed, when there was not absolute failure, there was a general tendency to decline in yield, and then to recover again more or less after a break. This was some- what marked after a year of fallow in 1860, and the growth of wheat in 1861 ; after which there was^ in 1862, fair pro- duce, especially on the third plot, where the nitrate was now applied. ; The .land was again fallow in 1863, and this was LEGUMINOUS CROPS. 113 again followed by improved growth, after which there was declining produce for a number of years to 1870 inclusive, and again recovery in 1874 after 3 years of fallow. This general view of the results is of interest, as fixing attention on the great tendency to failure of this leguminous crop, when grown year after year on the same land. Independently of the occasional entire failures, there were also considerable fluctuations from year to year according to season; and the table shows that there was, besides, upon the whole considerable decline from period to period. Turn- ing now to the effects of the different manures, it is seen that there was, over each period, a considerable increase of increased produce by the use of the mineral manure containing potash, ^'■<«^«<=«. but that there was comparatively little further increase by eral man- the addition of nitrogenous to the mineral manure. Thus, "™- as shown in the upper division of the table, the average •^'''^™- annual total produce over the 32 years (which, however, nitrogenma included 7 without any bean crop) was — without manure "inanure. 1389 lb., with the mineral manure alone 2168 lb., and with the mineral and nitrogenous manure together 2507 lb. That is, whilst the mineral manure without nitrogen gave an average annual increase of 779 lb., the addition to it of nitrogenous manure only further raised the produce by 339 lb. Or if, instead of taking the average of the 32 years, we take it only over the 26 years in which there was any bean crop, as shown in the lower division, the average total produce was — without manure 1709 lb., with purely mineral manure 2688 lb., and with the mineral and nitrogenous manure to- gether 3086 — that is, there was an annual average increase of 979 lb. by the mineral manure containing potash, and of only 398 lb. more by the addition of nitrogenous manure. It may be added that details not given in the table further show, that in two of the last 8 years the total produce was, without manure, only exceeded three or four times during the whole period — namely, during the first five years ; with mineral manure alone, it was only exceeded four or five times ; and with the mineral and nitrogenous manure together, it was only exceeded six times. Indeed the table shows that on both of the manured plots the average total produce over the last 4 years of actual crop (with 4 of fallow in the 8 years) was nearly as much as the average of the first 8 years of crop. Thus, with the purely mineral manure, the average total produce of the first 8 years was 3208 lb., and over the last 4 years of crop it was 3011 lb. ; and with the mineral and nitrogenous manure it was, over the first 8 years 3555 lb., and over the last 4 years of TOL. vn. H 114 THE ROTHAMSTED EXPERIMENTS. crop 3292 lb. It will be seen further on that the average annual yield of nitrogen was also nearly as great over the last 4 years of crop as over the first 8 years. Ammrni- It may be observed that nitrogen supplied as ammonium- tZiSie s^^*s to the highly nitrogenous leguminous crop seldom gives foriegu- any increase, and is sometimes injurious in the year of minous application ; though some benefit may afterwards result from crops. ^^^ residue after the ammonia has been converted into nitric Nitrates acid. Even nitrates, however, directly applied as manure, wncertain. ^re very uncertain in their action, and at any rate yield very much less increase of produce with the highly nitrogenous Leguminosae than with the Graminese, and crops of other Orders yielding produce of low percentage of nitrogen in their dry substance, and accumulating comparatively little nitrogen over a given area of land. Continuous It is specially to be noted, that whilst the cereal crops may vAthhea ^^ successfuUy grown for many years in succession on the afaiiure. Same land, provided only that mineral and nitrogenous manures are liberally supplied, this leguminous crop — beans — gradually fails when so grown ; and although characteristi- cally benefited by mineral manures containing potash, neither these alone, nor a mixture of mineral and nitrogenous manure, has sufficed to maintain even fair growth for a number of The reason years in succession. The result is, however, not entirely due "'*2'- to deficiency in the supply of constituents within the soil, but is also in a considerable degree dependent on the fact that, by the continuous growth of the crop, with its special habit and range of roots, the surface-soil acq[uires a close and unfavourable condition, and a somewhat impervious pan is formed below. The improved result in the later years with the intervention of fallow, further illustrates the fact that the previous failures were not wholly due to exhaustion. Amount of The next Table (36) shows the amounts of nitrogen in the imn^mm. ^^^^ crops, the produce of which we have been considering. The table is on the same plan as that relating to the produce ; the upper division giving the averages for the four 8-yearly periods, and for the total period of 32 years, and the lower division those for the years of crop only, within each period ; and, as in Table 35, the results for the total produce only (corn and straw together) are given. Eeferring to the figures in the upper division of the table, it may be observed that, notwithstanding there were 6 blank years, and one year of wheat, out of the 32, and notwithstand- ing that the produce declined much, and gave on the whole much less than the average obtained under ordinary agri- cultural conditions, yet the average yield of nitrogen in the crops grown without any supply of it was much more than LEGUMINOUS CEOPS. 115 in either of the cereals, the root-crops, or potatoes, grown under similar conditions. Thus, as the bottom line of the upper division shows, there was an average over the 32 years, of 24.8 lb. of nitrogen per acre per annum in the crops without any manure, but of 35.4 lb. with the mineral manure without nitrogen; whilst the amount was raised to only 42.4 lb. by the addition of nitrogenous manure. Over the first 8 years, however, the yield was very much higher, being for the three plots re- spectively 48.4, 60.2, and 69.0 lb. Over the second period of 8 years the average was not far from that of the whole 32 years, but over the third and fourth periods it was much less. Yields of nitrogen without manure, with min- eral man- ure, and ■with nitro- genous manure. TABLE 36.- -Beans. Yield op Nitro&en, average pee ache PEE ANNUM, LB. 8-YeAE PERIODS. Total produce (corn and straw). Unmanured. Mixed mineral manure (including potash). Mixed mineral manure and nitrogen. Avbeaqb pek acre per annum, over each 8 years, and over the 32 years. 8 years,. 1847-54 8 years, 1855-62 8 years, 1863-70 8 years, 1871-78 lb. 48.4 25.3 9.2 16.4 lb. 60.21 34.3 23.5 26.7 lb. 69.0 36.8 35.1 28.7 32 years, 1847-78 24.8 35.42 42.4 Avbeaqb per acre pee annum, over the years of crop only, each period. 1st 8 years, 8 crops 2nd 8 years, 7 crops 3rd 8 years, 7 crops 4th 8 years, 4 crops . 48.4 28.9 10.4 32.7 60.23 39.2 26.8 53.3 69.0 42.1 40.0 57.4 32 years, 1847-78, 26 crops . 30.5 43.94 52.2 1 7 years, excluding 1849, in whicli year the produce was accidentally not weighed. 2 31 years, excluding 1849. * 7 crops, excluding 1849. 4 25 crops, excluding 1849. As in the case of the total produce itself, so also in that of the nitrogen in the total produce, if we take the averages of the years of crop only, as given in the bottom division of the table, we have a much higher average yield per annum over the 4 years of crop of the last 8 years, than over the years of 116 THE EOTHAMSTED EXPEEIMENTS. Influence of fallow on beans. Failwe of clover grown at short in- tervals. crop of either the second or the third period of 8 years. In- deed, on the two manured plots there is an average annual yield of nitrogen per acre over the 4 years of crop during the last 8 years not very far short of the average of the first 8 years. Thus, with the purely mineral manure, there is an average annual yield of nitrogen over the first 8] years of 60.2 lb., and over the 4 years of crop of the last 8 of 53.3 lb.; and, with the mineral and nitrogenous manure together, over the first 8 years of 69.0 lb., and over the 4 years of crop of the last 8 years, of 57.4 lb. That is, with the intervention of fallow, we have, though not good agricultural crops, yet really large yields of nitrogen compared with those obtained in many of the preceding years ; and very large yields without any supply by manure, com- pared with those obtained under the same conditions with any of the non-\egum.mous crops. It would appear probable, therefore, that if a suitable mechanical condition of the land could have been maintained, fair crops, and large yields of nitrogen, would also have been maintained. Upon the whole, then, although the crop practically failed when it was thus attempted to grow it year after year on the same land, it nevertheless accumulated, in its above-ground produce, much more nitrogen over a given area than the crops of the other Orders, but was little benefited by an arti- ficial supply of nitrogen. We have now to record a still greater failure than that with beans — namely, when it was attempted to grow another leguminous crop year after year on ordinary arable land — this time Trifolium pratense, or Eed clover. The results are summarised in Table 37. The table is headed Red clover, sown frequently on the same land. The period of experiment was in fact 29 years — from 1849 to 1877 inclusive. But the details, not given in the table, show that although clover was sown fifteen times in the 29 years, in only 7 was any clover crop ob- tained; whilst about one-fifth of the produce of the whole series of years was yielded in the first year, 1849. It is, in- deed, fully recognised that in our own country clover will not grow under ordinary conditions more frequently than once in a certain number of years, which varies according to soil and other circumstances, but is seldom so few as four, and frequently as many as, or more than, eight years. It should be stated that when the clover failed, sometimes a cereal crop, wheat or barley, was sown ; but more frequently the land was left fallow. Further, the amounts of produce entered in the column headed Series 1 are in each case the LEGUMESrOUS CHOPS. 117 means of those on three plots, each of which occasionally re- ceived a mineral manure containing potash ; and the results given in the column Series 2 are also the means of three plots, each with the same mineral manure as Series 1, and nitrogenous manures occasionally applied in addition. TABLE 37. — Eed Clovee. Sown frequently on the same land. Total Produce per acre per annum, as Hay. Series 1. Mineral manure alone. Series. 2. Mineral and nitrogenous manures. Summary. Peoduoe. 29 years, 1849-77 . . {'^^^^^^ Years of crop only . . Average lb. 52,991 1,827 4,416 lb. 60,689 2,093 4,668 Years of clover only (7) . { ^^^^^^ 29,195 4,171 31,886 4,555 Summary. Nitkoqen (estimated). 29 years, 1849-77 . . j ^^^^^^^^ Years of crop only . . Average 929.4 32.0 77.5 l,04.S.l 36.0 80.2 Years of clover only (7) . { 1° ^^^age 700.7 100.1 765.3 109.3 It should be explained that very large crops of clover were Variations obtained in the first year, 1849; less than one-quarter as *^^*^^^^ much in the third year, 1851 ; and in the fourth about half as much as in the first. No more clover was then obtained until the seventh year, when there was very little. After this, there was more or less in the eleventh, seventeenth, twenty-third (on Series 2), and lastly, (on Series 1) in the twenty-seventh year ; but in no case, excepting in the fourth year, was the amount of produce half as much as in the first. Comparing the results without and with the nitrogenous iifects of manure, the table shows that the average annual total pro- '"'"'■ti'ogencms duce of clover-hay, and other crops, was, reckoned over the cimer. 29 years, 1827 lb. without, and 2093 lb. with, the nitrogenous manure ; and, reckoned in the same way, the average annual 118 THE EOTHAMSTED BXPEEIMENTS. Failv/re of continuous clover-crop- img on land. yield of nitrogen was, without nitrogenous manure 32 lb., and with it 36 lb. Eeckoned, however, over the years of crop only, the yield of nitrogen in the clover and other crops was 77.5 lb. per acre per annum without, and 80.2 lb. with, the nitrogenous manuring. Or, reckoning the nitrogen in the clover alone, and only over the years when it gave any crops, the average annual yield of it over those 7 years was, without nitrogenous manure 100.1, and with it 109.3 lb. There was, therefore, comparatively little increase, either in the pro- duce, or in the yield of nitrogen, by the use of nitrogenous manures. To conclude in regard to these experiments : The attempt to grow clover year after year on this ordinary arable land, by means of such mineral manures as increase the luxuriance of growth when there is a fair plant, or even by the addition to these of nitrogenous manures, has entirely failed. In view of this failure to grow the crop continuously on ordinary arable land, the next results to which we have to call atten- tion are of much interest and significance. Growth of Red Clover, year after year, on rich Garden Soil. Success of In 1854, after it seemed clear that the plant would not 'dover^ov- co^tiii^® to grow on the arable land, clover was sown in a pmgon garden only a few hundred yards distant from the experi- gardmsml. mental field, on soil which had been under ordinary kitchen- garden cultivation for probably two or three centuries. It is remarkable that, under these conditions, the crop has grown luxuriantly almost every year since — 1893 being the fortieth season of the continuous growth. Further particulars will be given on the point presently, but it may here be premised that, at the commencement, the percentage of nitrogen in the surface-soil of the garden was four or five times as high as in that of the arable soil in the field ; and it would doubtless be richer in all other manurial constituents also. Indeed, after the growth of clover for 25 years in succession, even the second 9 inches of depth was found to be still very much richer in nitrogen than the first 9 inches in the field. Table 38 gives the results for each of the 40 years of experiment with clover on the rich garden-soil. The first column after the dates shows the number of cuttings each year, the second the amounts of produce per acre, reckoned in the condition of dryness as hay, the third the amount of dry substance, the fourth that of the mineral matter, and the last the estimated amounts of nitrogen per acre in the crops. At the bottom of the table are given the average annual results, over periods of 10, 10, 10, 10, and 40 years. It Condition of the ga/r- Table 38 LEGUMINOUS CROPS. 119 TABLE 38. — Red Clover. Grown year after year on rich Garden Soil. 40 years, 1854-93. Hay, Dry Matter, Mineral Matter, and Nitrogen, per acre per annum. Number of cuttings. As hay. Dry matter. Mineral matter. Nitro- gen (esti- mated). Seed sown. lb. lb. lb. lb. 1854 2 5,191 4,326 435 125 1854, March. 1855 3 18,113 15,094 1560 435 1856 2 11,027 9,190 1116 265 1857 3 14,855 12,379 1384 357 1858 2 7,608 6,340 792 183 1859 2 6,227 5,189 687 149 ... 1860 1 8,679 7,233 806 208 1860, May. 1861 2 13,353 11,128 1285 321 ... 1862 2 10,042 8,368 991 241 1863 2 11,798 9,832 971 283 1864 2 5,500 4,583 446 132 1865 1 2,044 1,704 190 49 1865, AprU. 1866 2 10,466 8,713 908 251 1867 2 6,748 5,624 573 162 1868 1 991 826 106 24 1868, April. 1869 2 4,183 3,486 387 100 1870 1 1,741 1,451 148 42 , 1871 1 4,513 3.761 458 108 1871, April. 1872 2 10,142 8,452 899 243 1873 2 9,287 7,740 772 223 1874 3 5,899 4,916 540 142 1874, May and July. 1875 1 2,731 2,276 230 66 1875, July and September. 1876 2 3,517 2,931 279 84 1876, September. 1877 1 3,533 2,944 326 85 1877, May. 1878 3 13,416 11,180 1336 322 1879 1 2,788 2,282 428 66 1879, May. 1880 2 5,742 4,785 643 138 1880, April. 1881 2 4,262 3,552 330 102 1881, April (mended). 1882 3 6,433 5,361 641 154 1882, April mended). 1883 1 2,716 2,264 315 65 1883, May. 1884 3 9,990 8,325 863 240 1885 3 6,511 5,426 615 156 1886 1 2,702 2,252 313 65 1886, April. 1887 2 3,287 2,739 264 79 1887, April (mended). 1888 1 1,841 1,535 211 44 1888, April (mended June). 1889 2 8,664 7,221 754 208 1889, April (mended). 1890 1 2,817 2,348 367 68 1890, April. 1891 2 6,696 5,580 574 161 1891, May (mended). 1892 1 3,568 2,973 355 86 1892, May 7 (May 27, mended). 1893 2 5,941 4,951 500 143 1893, April (mended). AVERAGE PEE ACRE PER ANNUM. 10 years, 1854-63 „ 1864-73 „ „ 1874-83 „ „ 1884-93 10,689 5,561 5,099 5,202 8,908 4,634 4,249 4,335 1003 489 507 482 257 133 122 125 40 „ 1854-93 6,638 5,532 620 169 120 THE EOTHAMSTED BXPEEIMENTS. should be stated that, as the garden clover plot is only a few yards square, calculations of produce per acre can only give approximations to the truth; but it is believed that they can be thoroughly relied upon so far as their general indications are concerned. It may be added that five times during the whole period, gypsum has been applied to one-third, and a mineral manure containing potash, but no nitrogen, to another third of this plot. Produce of We shall confine attention to the amounts of produce i^«OTwZ« reckoned as hay, and to the estimated amounts of nitrogen in ^own the produce. Casting the eye down the column of produce cimer. ^^ j^^^y^ jt jg ggen at a glance that, excepting a few occasional years of very high produce during the later periods, the amount of crop is very much greater during the first than during either of the subsequent periods of 10 years. In fact, as is seen at the foot of the table, there was an average annual produce equal to 10,689 lb. of hay over the first period of 10 years, but of only 5561 lb. over the second, 5099 lb. over the third, and 5202 lb. over the last 10 years. Now, even these latter amounts correspond to what would be considered fair though not large crops, when clover is grown in an ordinary course of rotation, once only in 4, or in 8 years, or more ; so that the produce in the earlier yeate on this rich garden-soil was very unusually heavy. Indeed the average annual produce over the whole period of 40 years — namely, 6638 lb., or nearly 3 tons of hay — would be a very good yield for the crop grown only occasionally in the ordinary course of agriculture. Amount of But it is when we look at the figures in the last column of TOiroffOT. in ^jjg table, which show the estimated amounts of nitrogen in musiy the crops, that the importance and significance of these grovm results obtained on rich garden-soil are fully recognised; and this is especially the case when they are compared with those obtained on ordinary arable land. Thus the amount of nitrogen in fair crops of wheat, barley, or oats, will be from 40 to 50 lb. per acre ; of beans about 100 lb. ; of meadow-hay about 50 lb. ; and of clover grown occasionally in rotation from 100 to 150 lb. ; but here, on this rich garden-soil, the produce of clover has in one year contained more than 400 lb. of nitrogen, in three years more than 300 lb., in several more than 200 lb., and in only thirteen years of the 40 less than 100 lb. In fact, as the figures at the bottom of the table show, the estimated average annual yield of nitrogen in the above- ground growth was — over the first 10 years 257 lb., over the second 10 years l33 lb., over the third 10 years 122 lb., over the last 10 years 125 lb., and over the whole period of 40 LEGUMINOUS CEOPS. 121 years 159 lb ; whilst, as the details show, the yield of nitrogen in the thirty-first year (1884) was about 240 lb., in the thirty-second year 156 lb., in the thirty-sixth year 208 lb., in the thirty-eighth year 161 lb., and in the fortieth 143 lb. Further, the averages over the second, third, and fourth, 10 years of the continuous growth (133, 122, and 125 lb.) were about as much as in a fair but not large crop grown occa- sionally under the ordinary conditions of agriculture ; whilst the average of the 40 years, 159 lb., is as much as in a really good crop grown occasionally in rotation. There would seem, then, to be clearly indicated, a soil- GondiUon source of failure on the arable land, and a soil-sotcree o^ fjingfn- success on the garden-soil. jiumee. The results given in Table 39 will throw some further light on this point. It shows the percentage of nitrogen in the first 9 inches of depth of the garden-soil, in 1857 and in 1879, between which periods the growth of 21 years had been removed. It also shows the estimated amounts of nitrogen per acre in the surface-soil at the two periods, and the reduction in the amount during the 21 years. TABLE 39. — Red Clover, grown on rich Garden-Soil. Nitrogen per cent, and per acre, in the fine soil, dried at 100° 0. (First 9 inches of depth.) 1857. 1879. Difference. per cent. 0.5095. per cent. 0.3634 per cent. 0.1461 Per acre, .... Total Per acre per annum (21 years) lb. 9528 lb. 6796 lb. 2732 130 It may be mentioned that the percentage of nitrogen given for the sample collected in October 1857, is the mean of duplicate or more determinations, made in 1857, in 1866, and again in 1880 ; and it is almost identical with the results obtained at the latest of these dates. The first point to notice is that the first 9 inches of depth Richness of of this rich garden-soil contained more than half a per cent the garden of nitrogen — that is, nearly four times as much as the average nitrogen. of the Eothamsted arable soils, and nearly five times as much as the exhausted arable clover-land-soil where the crop failed. It is, of course, true that the garden-soil would be corre- spondingly rich in all other constituents ; but some portions of the arable soil where the clover failed, had received much more of mineral constituents by manure than had been re- moved in the crops. The result given for 1879 is the mean of determinations 122 THE EOTHAMSTED EXPERIMENTS. Redaction made on three separate samples, for whicji the determinations in nitrogen agreed Very well. The results can leave no doubt that there soil under had been a great reduction in the stock oi nitrogen in tne clover. surface-soil s-ince 1857. The reduction amounts to nearly 29 per cent of the whole in the 21 years ; and, reckoned per acre, it corresponded, as shown in the table, to a loss of 2732 lb. during the 21 years ; and although, as has been seen, fairly average, and even good crops, were still grown, it is obvious that coincidentally with this great reduction in the stock of nitrogen in the surface-soil, there has been a very marked re- duction in the clover-growing capability of the soil. Reduced On this point it may be mentioned that, whilst fresh seed persistence ^a,s only sown iive times during the first 20 of the 40 years, Ouced'pro- it has been fully or partially sown twenty-one times during duceofthe the last 20 years. It is obvious, therefore, that the plant was able to stand very much longer in the earlier than in the later condition of the soil. Indeed, both the reduced persistence of the plant, and the reduced produce, have been coincident with a considerable reduction in the stock of nitrogen in the soil. The question arises. What relation does the amount of nitrogen lost by the soil bear to the amount taken off in the crops? Amounts of It is admittedly necessary to accept with some reservation nitrogen results of calculations of produce per acre from amounts ob- the crop tained on a few square yards, but the general indications may and lost ly doubtless be trusted. Such estimates show more than 160 lb, ^ **" ■ of nitrogen to have been removed per acre per annum in the crops over the 21 years ; whilst the estimated loss of the sur- face-soil corresponds to about 130 lb. per acre per annum. That is to say, the loss by the surface-soil is sufficient to account for a large proportion of the nitrogen removed in the crops. There is, however, evidence leading to the conclusion that, when excessive amounts of farmyard manure have been applied, as had been the case with this garden-soil, there may be some loss by the evolution of free nitrogen ; and obviously, so far as this may have occurred, there will be the less of the ascertained loss to be credited to assimilation by the growing clover. Clover On the other hand, it is known that when growing on warnmb- O'^cli^^iT arable soil, the clover plant throws out a large aM. amount of feeding roots in the lower layers ; and although in the case of so rich a surface-soil the plant may derive a larger proportion of its nutriment from that source, we must at the same time suppose that it has also availed itself of the resources of the subsoil. Unfortunately, in 1857 samples LEGUMINOUS CEOPS. 123 were only taken to a depth of 9 inches, so that no compari- son can be made of the condition of the subsoil at the two periods. In 1879, however, the second 9 inches of the gar- den-soil was found to contain a much higher percentage of nitrogen than the first 9 inches of the clover-exhausted arable field, and about three times as high a percentage as the sub- soil of the arable field at the same depth. It cannot be doubted, therefore, that the subsoil of the garden plot has contributed nitrogen to the clover crops. Here, then, notwithstanding the very little effect of direct Smi-som-ce nitrogenous manures on either the beans or the clover grow- /»*™?«»^ . . ^ TOT CtOV&iT, ing on the ordinary arable land, there would seem to be very clear evidence of a soil-source of, at any rate much of the enormous amounts of nitrogen assimilated over a given area by the clover growing on the rich garden-soil. It may here be observed that, in experiments on the mixed herbage of permanent grass-land, in which the growth of leguminous herbage was much increased by the application of mineral manure containing potash, it was found at the end of 20 years that the amount of nitrogen in the surface-soil had been considerably reduced, compared with that of a plot which had been unmanured, and had yielded very much less leguminous herbage. The conclusion was that, as in the case of the clover growing on the rich garden-soil, the nitrogen of the surface-soil had been a source of, at any rate much of the nitrogen of the increased produce of Leguininosse in the mixed herbage of the grass-land. Bed Clover grown after the Beans. After the cessation of the experiment with leans in 1878, the land was left fallow for between four and five years, to 1882 inclusive, when grass-seeds were sown, but failed. On this land, on which the attempt to grow the leguminous crop, beans, year after year had failed, and been abandoned, barley and clover were sown in the spring of 188.3. In April 1883, however, before the barley and clover were Exhaustion sown, the surface-soil (free of stones, and reckoned dry) of the o/™*™?*™ plot, which had been entirely unmanured during the 32 years of the experiments with the beans, was found to contain 0.0993 per cent of nitrogen, that of the mineral-manured plot 0.1087 per cent, and that of the plot which had received both the mineral and nitrogenous manure 0.1163 per cent — amounts which show considerable nitrogen exhaustion of the surface- soil. Also in 1883, the nitrogen as nitric acid was determined in samples,' each of 9 inches of depth, down to a total depth of 124 THE BOTHAMSTED EXPERIMENTS. Table 40 72 inches. In the case of several plots the results show, cal- culated per acre, that the total amount of nitrogen as nitric acid to the depth of eight times 9 inches, or 72 inches in all, was 27.95 lb. in the unmanured plot, 20.72 lb. in that with purely mineral manure, and 25.38 lb. in that of the plot which had received both mineral and nitrogenous manure. In the soil of the farmyard manure plot, on the other hand, the amount was about twice as much — namely, 50.46 lb. Excluding this last result, it may be said that the amounts of nitrogen already existing as nitric acid, to the depth deter- mined, were very small. These, then, were the conditions of the soil when the barley and clover were sown in the spring of 1883. The clover grew very luxuriantly from the first, so much so as to considerably interfere with the growth of the barley. Table 40 shows the amounts of nitrogen per acre in the bar- ley and clover in 1883, and in the clover in 1884 and 1885. TABLE 40. — Barlet and Clover, grown aeter Beans, Geescropt Field. Nitrogen removed per acre in. tie crops. Previous condition of manuring. 1883. Barley and clover. 1884. Clover. 1885. Clover. Total. Without manure . Mineral manure and some \ nitrogen * ) Mineral manure only . lb. 45.0 57.2 59.3 lb. 183.2 193.1 206.4 lb. 52.7 79.9 81.6 lb. 280.9 330.2 347.3 It should be stated that the plots, the yield of -nitrogen of which is here given, do not exactly correspond with those for which the yield of nitrogen in the beans was given ; some of the barley and clover crops having been taken together where no difference in the produce was observable. Thus, half the plot represented as without manure had been unmanured from the commencement — that is, for nearly 40 years, but the other half received some nitrogen to 1878 inclusive, but had since been entirely unmanured. Again, the results given in the second line relate to the produce of a plot part of which received purely mineral manure, but the other part ammo- nium-salts or nitrate up to 1878, but none since. The results given in the third line relate, however, to a plot which has not received any nitrogenous manure from the commencement of the experiments with the beans, but which was not brought under experiment until 5 years later than the other plots. Thus, on a plot where a purely mineral manure containing potash, but no nitrogen, had been applied for 27 years, to LEGUMINOUS CKOPS. 125 1878 inclusive, and no manure since, 347.3 lb. of nitrogen Yield of were gathered per acre, almost wholly by the leguminous "■*"?«"• crop — clover. On a plot on part of which the mineral manure only, and on part the same mineral manure and am- monium-salts or nitrate had been applied up to 1878, but nothing since, 330.2 lb. of nitrogen were removed in the crops. Lastly, where to half of the plot no manure whatever had been applied for nearly 40 years, but to the other half ammonium-salts or nitrate up to 1878, the yield of nitrogen in the barley and clover was 280.9 lb. Here, then, in a field where beans had been grown for i^arge many years in succession, and had yielded much less than 'f°^soif^ average crops, and the land had then been left fallow for poor in several years ; where the surface-soil had become very poor "'■^"5'««- in total nitrogen ; where both surface and subsoil were very poor in ready -formed nitric acid; and where there was a minimum amount of crop-residue near the surface for decom- position and nitrification ; there were grown very large crops of clover, containing very large amounts of nitrogen. Not only was so much nitrogen removed in the crops, but ThesoU the surface-soils became determinably richer in nitrogen as Sooere hy the results in Table 41 show. There are there given the thedover. percentages of nitrogen in the sifted dry surface-soil of the three plots for which the produce and the nitrogen in the beans have been given. The results relate to samples taken in April 1883, before the sowing of the barley and clover, and in November 1885, after the removal of the crops. Tlie first two columns show the percentages of nitrogen, and the other columns the calculated amounts of it per acre, in the surface-soils, 9 inches deep, at the different dates ; also the estimated gain of nitrogen under the influence of the growth of the clover. TABLE 41. — Nitrogen, pee cent, and per acre, in the sitrfacb- SOILS, BEFORE AND AFTER THE GROWTH OP THE BaRLEY AND OlOVER. Nitrogen in sifted dry soil. Per cent. Per acre. 1883. 1885. 1883. 1885. 1885 + or - 1883. 1. Without manure . 2. With mineral manure "1 containing potash J 3. With mineral manure "1 and nitrogen J per cent. 0.0993 0.1087 0.1163 per cent. 0.1083 0.1149 0.1225 lb. 2441 2672 2859 lb. 2662 2824 3011 lb. + 221 + 152 + 152 126 THE EOTHAMSTED EXPEKIMENTS. Large ac- "Without assuming that the figures represent accurately o/SroS ^'^^ amounts of nitrogen accumulated per acre, it cannot be —where doubted that the surface-soils had become considerably richer. yramV""" If, for the sake of illustration, we assume that 300 lb. of nitrogen were removed per acre in the crop, and that 150 lb. were accumulated in the surface-soil, we have 450 lb. of nitrogen to account for, as gathered by the crops within a period of little more than two years. It is clear that we have in the experimental results them- selves no conclusive evidence as to the source of so large an amount of nitrogen. As the surface-soil became determinably richer, it is obvious that it must have been derived either from above or below it — from the atmosphere or from the subsoU ; and, if from the subsoil, the question arises, whether it was taken up as nitric acid, as ammonia, or as organic nitrogen ? Eesults relating to these points will be referred to presently ; but it must be admitted that there is nothing in the experimental results themselves to show that so large an amount of nitrogen could have been available as nitric acid. There remains the question whether the free nitrogen of the atmosphere has in any way been brought into combi- nation, either within the soil or within the plants ? Evidence on these points will be adduced further on. Various Leguminous Plants grown after Bed Clover. We have now to adduce another and even much more striking instance of successful growth, and of great accumu- lation of nitrogen, by plants of the leguminous Order, on soil where another plant of the same order had failed, and where the surface-soil had become very poor in nitrogen. The experiments were made on the plots where it had been attempted to grow red clover year after year on ordinary arable land; where, in fact, clover had been sown twelve times in 30 years, and where, in eight out of the last ten trials, the plant had died off in the winter and spring succeeding the sowing of the seed — in four cases without any crop at all, and in the other four yielding very small cuttings. In 1878, the land was devoted to experiments with various leguminous plants, differently manured, having regard, how- ever, to the previous manurial history of the plots. Object of The object was to ascertain whether, among a selection of plants all belonging to the leguminous Order, but of different habits of growth, and especially of different character and range of roots, some could be grown successfully for a longer time, and would yield more produce, containing more nitrogen, as well as other constituents, than others ; all being supplied LEGUMINOUS CEOPS. 127 with the same descriptions and quantities of manuring sub- stances, applied to the surface-soil. Further, whether the success in some cases, and the failure in others, would afford additional evidence as to the source of the nitrogen of the Leguminosse generally, and as to the causes of the failure of red clover when grown too frequently on the same land. Accordingly, fourteen different Leguminosse were selected, Crops and sown in 1878. These included eight species or varieties ^f^^l^-^" of Trifolium, two species of Medicago, Melilotus leucantha, Lotus corniculatus, Vicia sativa, and Lathyrus pratensis. Of these, six of the eight Trifoliums have already failed, and heen replaced by other plants ; as also have the Medicago lupilina, the Lotus corniculatus, and the Lathyrus pratensis, the last being replaced in the second year by Onohryehis sativa. The plants which have maintained fair, but very varying, character of growth, are the Trifolium repens, Vicia sativa, Melilotxis leucantha, and Medicago sativa ; and we pro- pose to give some account of the growth of these plants on the clover-exhausted soil. That the surface-soil had become very poor in nitrogen is Soil poor evident from the fact that the mean percentage of it in the »"«*<'£'««• sifted dry surface-soil of five of the clover plots was, in March 1881, only 0.1058, which is considerably lower than was found in the same field many years before ; and lower than has been found in any of the fields at Eothamsted, excepting those where crops have been grown for many years on the same land without nitrogenous manure. It is a point of interest, however, that the percentage in the surface-soil was not so low as in immediately adjoining land, which had been under alternate wheat and fallow for nearly 30 years without manure. The real interest of the results depends on the amounts, Thepomts and on the difference in the amounts, of nitrogen which the »/»«<«»■««'■ various plants have assimilated over a given area, all growing side by side on the same red clover-exhausted land, and with the same mineral manures, without any supply of nitrogen. Accordingly, the upper part of Table 42 (p. 128) shows the Table 42 estimated average amounts of nitrogen in the gramineous ^p^"'^"^- crop — wheat, grown in alternation with fallow, over 27 years to 1877 inclusive, and in the red clover (together with other crops when it failed) over 29 years, also to 1877 inclusive. Then, in the body of the table are given the amounts of nitro- gen in the wheat alternated with fallow, a,nd in the produce of five different leguminous plants during the subsequent years, commencing with 1878, and extending in some cases to 1891. Thus, over the preliminary period, the wheat gave an 128 THE EOTHAMSTED EXPERIMENTS. average annual yield of nitrogen per acre of 15 lb., and the clover gave, over much the same period, an average of 32 lb. of nitrogen. TABLE 42. — Estimated yield of Nitrogen per acre, in lb., in Wheat ALTERNATED WITH FALLOW, AND IN VARIOUS LeGUMINOTIS CeOPS, WITHOUT Nitrogenous Manure. Unman- ured. Mineral manures only. Fallow wheat. Trifolium pratense. Trifolium repens. Vida satvaa. Melilotus leuccmtha. Medicago saliva. PRELIMINARY PERIOD— Wheat and Fallow, 27 years, 1851-77 ; Red Clovee, &o., 29 years, 1849-77. Average per acre per annum lb. 32 EXPERIMENTAL PERIOD. lb. lb. lb. lb. lb. lb. 1878 14 51 53 1879 5 50 82 46 130 1880 12 8 58 36 28 1881 9 21 8 65 60 28 1882 9 18 74 146 145 HI 1883 13 101 27 143 1884 15 113 56 337 1885 16 7 15 97 16 90 52 58 270 1886 Lupins 167 1887 13 6 64 82 247 1888 9 ^ r 60 32 161 1889 9 14 Medi- - cago J I saliva \ 65 61 ) , 79 / 23 163 1890 Fallow Trifolium ( 124 t 147 1891 18 J I Faba vulg. praiensQ Total, 14years, 1878-91 163 1121 2832 1051 7022 1916 Average, 14 years, 1878-91 12 141 242 75 582 137 Average for years of crop 12 22 47 75 64 160 ■ 8 years only, 18YS-85. 2 12 years only, 1878-8! Yield of the various crops. Against these amounts the various crops yielded, over the subsequent years, averages per acre per annum as follows : The fallow-wheat, over 14 years 12 lb. ; the red clover (Tri- folium pratense), over 8 years 14 lb. ; the white clover (Tri- folivmi repens), over 12 years 24 lb. ; the vetch (Vicia sativa), over 14 years 75 lb. ; the Bokhara clover (Melilotus leucantha), over 12 years 58 lb. ; and the lucerne (Medicago sativa), over 12 years 137 lb. Or if we take the average amounts over the years of actual crop only, they were— in the wheat 12 lb., in the red LEGUMINOUS CROPS. 1-29 clover 22 lb., in the white clover 47 lb., in the vetch 75 lb., in the Bokhara clover 64 lb., and in the lucerne the enormous amount of 160 lb., of nitrogen per acre per annum. Again, if we take the total yields of nitrogen over the experimental periods, we have — in the wheat 163 lb., in the red clover 112 lb., in the white clover 283 lb., in the vetch 1051 lb., in the Bokhara clover 702 lb., and in the lucerne 1916 lb. ; that is, in the lucerne about twelve times as much as in the wheat, nearly twice as much as in the vetch, and very much more than in either of the other Leguminosse. Indeed, this very deeply and very powerfully rooting-plant yielded, in its above-ground produce alone, 337 lb. of nitrogen in 1884, 270 lb. in 1885, 167 lb. in 1886, 247 lb. in 1887, and an average of 146 lb. over the next four years. Not only have these large amounts of nitrogen been Smien- removed in the above-ground produce, but determinations of ™^'* *™ nitrogen in the soils of the vetch plot in 1883, and of the white clover, the Bokhara clover, and the lucerne plots, ia 1885, have shown, as in the case of the clover after the beans, that the surface -soil had gained rather than lost nitrogen, due to the accumulation of nitrogenous crop- residue. Here again, then, it is obvious that the original Nitrogen source of the nitrogen of the crops has not been the sui"face- ■^'■?" ?^^ soU itself. It must have been derived either from the theatmo- atmosphere or from the subsoil. sphere. The next results will throw some light on this point. Thus, having made initiative experiments of the same kind some years previously, in July 1883 samples of soil were taken to the depth of twelve times 9 inches, or 108 inches in all, on the wheat-fallow plot, on the white clover plot, and on two of the vetch plots, for the determination of the amount of nitrogen existing as nitric acid at each depth. Table 43 (p. 130) summarises the results. The iirst point to notice is that at each depth, from the first to the twelfth, the Trifolium repens soil contained much more nitrogen as nitric acid than the wheat-fallow soil; and as the figures at the bottom of the table show, whilst to the total depth of 108 inches, or 9 feet, the wheat-fallow soil was estimated to contain only 52.4 lb. of nitrogen as nitric acid per acre, the TrifoUwm repens soil — that is, the legu- minous plant soil — contained to the same depth 145.7 lb. Now, independently of the fact that the leguminous plant Nitrogen plots had received mineral manures and the wheat-land had *'^ ^^i^ter not, the characteristic difference in the history of the two crops. plots was, that the one had from time to time grown a legu- minous crop, and the other had not ; and the one which had grown leguminous crops contained, to the depth of 9 feet, VOL. VIL I 130 THE KOTHAMSTED EXPEEIMENTS. nearly three times as much nitrogen as nitric acid as the gramineous crop soil. TABLE 43.— NiTBOGEN as Niteio Acid per acre, lb., in Soils OF SOME Experimental Plots, without Nithogenods Manure FOE MORE THAN 30 Ybars ; HoosFiELD, EoTHAMSTED, Samples collected July 17-26, 1883. Depths. Wheat- fallow land un- manured. Trifolvvm repensy Series 1, Plot 4. Vicia saii/va, Series 1, Plot 4. Vicia sati/va, Series 1, Plot 6. THfolium repens, -f or - Wheat- land, -I- or - Trlfoliwm repens. Vicia sativa. Plot 4. Vicia sativa, Plot 6. Inches. 1-9 10-18 19-27 28-36 37-45 46-54 ■55-63 64-72 73-81 82-90 91-99 100-108 lb. 19.85 8.05 2.47 2.70 1.62 3.57 3.84 2.28 1.48 1.76 2.94 1.84 lb. 30.90 27.73 8.44 7.64 9.07 8.77 7.92 8.34 8.27 9.95 9.16 9.51 lb. 12.16 4.11 1.37 1.67 4.58 6.37 7.16 5.95 4.54 5.32 5.66 5.32 lb. 10.22 2.72 1.08 1.52 2.51 4.42 4.52 4.92 4.81 5.14 6.40 6.46 lb. -t-11.05 -fl9.68 + 5.97 + 4.94 + 7.45 4- 5.20 -1- 4.08 -1- 6.06 + 6.79 + 8.19 -1- 6.22 + 7.67 lb. -18.74 -23.62 - 7.07 - 5,97 - 4.49 - 2.40 - 0.76 - 2.39 - 3.73 - 4.63 - 3.50 - 4.19 lb. -20.68 -25.01 - 7.36 - 6.12 - 6.56 - 4.35 - 3.40 - 3.42 - 3,46 - 4.81 - 2.76 - 3.05 SUMMARY. 1-27 28-54 55-81 82-108 30.37 7.89 7.-60 6.54 67.07 25.48 24.53 28.62 17.64 12.62 17.65 16,30 14,02 8,45 1425 18,00 -h 36.70 -1-17.59 -1-16.93 -1-22.08 -49.43 -12.86 - 6.88 -12.32 -53.05 -17.03 -10.28 -10.62 1-54 55-108 38.26 14.14 92.55 53.15 30,26 33.95 22,47 32,25 -f54.29 -1-39.01 -62,29 -19.20 -70.08 -20.90 1-108 52.40 145.70 64.21 54.72 -1-93.30 -81.49 -90.98 Nitrifica- tion cmd soil nitro- gen. The difference is the greatest near the surface, but it is very considerable down to the lowest depths. In the first three depths there was more than twice as much nitrogen as nitric acid in the Trifolium repens, as in the wheat-fallow soil ; in the second and third three depths, there was more than three times ; and in the fourth, three more than four times as much. Hence it is obvious, that any loss by drain- age would be much the greater from the Trifolium plot, so that the difference between the two plots was probably greater than the figures show. In the case of both plots, the actual amount of nitrogen as nitric acid was the greatest near the surface, indicating more active nitrification ; and the greater amount in the TTifoliiim. LEGUMINOUS CROPS. 131 soil is doubtless due to more nitrogenous crop-residue from the leguminous than from the gramineous crop. Indeed, about 74 lb. per acre of nitrogen had been removed in the Trifoliwm repens crops, and only 18 lb. in the wheat (reckoned on the half-acre in crop) in 1882, and none from either in 1883, the year of soil-sampling ; and the crop-residue of the Trifolium repens would contain much more nitrogen than that of the wheat. But it is not probable that the excess of nitric acid in the Trifolium soil, together with the larger amount lost by drainage, could be entirely due to the nitri- fication of recent crop -residue. Some found in the lower layers was, however, doubtless due to washing down from the surface. But, as notwithstanding much more nitrogen had been removed in the crops from the leguminous than from the gramineous crop-land during the preceding 30 years, the surface - soil of the leguminous plot remained Aoam, slightly richer in nitrogen, it is obvious that the whole of J? -/"^ the nitrogen of the nitric acid could not have had its origin gen ame in the surface-soil. If, therefore, it did not come from the •^'''^ ^ atmosphere, it has been derived from the subsoil. The indication is, that nitrification is more active under mtrifica- the influence of leguminous than of gramineous growth and ^afterUm- crop-residue. There would not only be more nitrogenous mmoua matter for nitrification, but it would seem that the develop- 9'^°^^' ment of the nitrifying organisms is the more favoured. Part of the result may, therefore, be due to the passage down- wards of the organisms, and the nitrification of the organic nitrogen of the subsoil. An alternative is, that the soil and the subsoil may still Analter- be the source of the nitrogen, but that the plants may take "°'*™^- Tip, at any rate part, as ammonia or as organic nitrogen. To this point we shall recur presently. Comparing the amounts of nitrogen as nitric acid in the Remits Vicia sativa soils with those in the Trifolium repens soil, it is ^j^'^g^ to be" observed that, whilst from the Trifolium repens soil only 164 lb. of nitrogen had been removed per acre in the crops of the five years to 1882 inclusive, 366 lb. had been removed in the Vicia crops to the same date. Then, whilst none was removed in crops from the Trifolium plot in 1883, 101 lb. were removed in the Vicia crops just before soil- sampling. Under these circumstances one of the Vicia soils contained 81.5 lb., and the other 91 lb., less nitrogen as nitric acid per acre than the Trifolium repens soil. Of course we cannot know exactly how much was at the disposal of the plants at the commencement of growth ; but if there had only been as much as in the case of the Tri- folium plot, it is seen that the deficiency in the Vicia soils 132 . THE EOTHAMSTED EXPEEIMENTS. nearly corresponds with the amount removed in the crop, which was 101 lb. It may at any rate safely be concluded that most, if not the whole, of the nitrogen of the Vicia. crops, had been taken up as nitric acid. But, as the Vicia crops had removed much more in the preceding years than the trifolium crops, so also would their crop -residue be greater; and in fact much more nitrogen must have been taken up by the plants each year than the figures show — and the larger the crop-residue, the larger would be the amount of nitric acid for each succeeding crop. But the crop of 1883 was also large, and it would leave a correspondingly large nitrogenous crop - residue ; leaving, therefore, a large amount of the nitrogen assimilated to be otherwise accounted for than by previous crop-residue. Lastly in reference to these experiments, it is seen that at each of the twelve depths, the Vicia soils with growth, con- tained much less nitric acid than the Trifoliv/m, soil without growth ; and the difference is much the greatest in the upper four or five depths, within which the Vicia throws out by far the larger proportion of its feeding roots ; but the deficiency is quite distinct below this depth. The supposition is that, under the influence of the growth, water had been brought up from below, and with it nitric acid. In fact, determina- tions showed that, down to the depth of 108 inches, the Vicia soils contained less water than the Trifolium soil, in amount corresponding to between 6 and 7 inches of rain, or to be- tween 600 and 700 tons of water per acre. Fwrtherex- Experiments of the same kind were again made in 1885. ""'"'' Trifolium repens was again selected as the weak and super- ficially rooting plant, Melilotus leucantha as a deeper and stronger rooting one, and the Medicago sativa as a still deeper and still stronger rooting plant. Samples of soil were taken at the end of July and the beginning of August, from two places on each plot, and in each case as before, to twelve depths of 9 inches each, or to a total depth of 108 inches, or 9 feet. It will suffice to quote the results for the Trifolium repens and the Medicago sativa plots. They are given in Table 44. It is seen that there was much less nitrogen as nitric acid in the Trifolivmv repens soil in 1885, after the removal of 97 lb. in the crops, than in 1883 (see Table 43, p. 130), when there had been no crop. The deficiency is the greatest in the two upper layers ; but it extends to the fifth depth, repre- senting the range of the direct and indirect action of the superficial roots. Below this point there is, however, even more than in 1883 ; due, doubtless, in part to percolation from above during the two preceding seasons without growth^ LEGUMINOUS CEOPS. 133 and possibly in part to percolation of the nitrifying organ- isms, and the nitrification of the nitrogen of the suh-soil. Let us now compare the results relating to the Medicago sativa with those relating to the Tri/olium repens soils. TABLE 44. — Nitbogen as Niteic Acid per acee, lb., in the Soil and Subsoils of some Expeeimbntal Plots, without Nitrogenous Manure fob moee than 30 Years ; Hoosfield, RoTHAMSTED. Samples collected July 29 to August 14, 1885. Series 1. Mineral manures. Depths. Trifoliurfh repens, Medicago sativa. Medicare sativa, -l-or- Trifolium repens. Plot 6. Plot 6. Inches. lb. lb. lb. 1-9 11.50 8.88 - 2.62 10-18 1.38 1.11 - 0.27 19-27 0.90 0.78 - 0.12 28-36 1.86 0.81 - 1.05 37-45 7.08 0.99 - 6.09 46-54 11.31 0.93 - 10.38 55-63 13.14 0.57 -12.57 64-72 12.63 0.81 -11.82 73-81 11.19 0.70 - 10.49 82-90 10.70 0.61 -10.09 91-99 11.08 0.44 - 10.64 100-108 9.96 0.41 - 9.55 Total . 102.73 17.04 -85.69 SUMMARY AND CONTROL. 1-9 10-18 Mixture of 1 19-108 inclies J Total . 11.50 1.38 88.02 100.90 1.11 6.97 16.96 - 2.62 - 0.27 -81.05 -83.94 The table of the estimated nitrogen in the produce per acre (p. 128) shows that, from the commencement to 1885 inclu- sive, the Trifolium repens yielded only 261 lb. of nitrogen in crops, but that the Medicago gave 917 lb. Again, in 1885, the year of soil-sampling, the Trifolium gave only 97 lb., but the Medicago gave 270 lb. It is further to be observed that, quite accordantly with the usual character of growth of lucerne in agriculture, with the increasing root-range, and consequently increased command of the stores of the soil and subsoil, the yield of nitrogen increased from 28 lb. in the first and second years, to 337 lb. in the fifth year of growth, declining, however, somewhat afterwards. Under these circumstances of very large yields of nitrogen in the crops, there is at every one of the twelve depths less, 134 THE EDTHAMSTED EXPERIMENTS. and at most very much less, nitrogen as nitric acid remaining in the soil than where so much less had been removed in the Trifolivmo repens crops. The difference is distinct even in the upper layers, but it is very striking in the lower depths. Thus there is, on the average, not one-twelfth as much nitric- nitrogen in the lower ten depths of the soil of the deep-root- ing and high nitrogen-yielding Medicago sativa, as in those of the shallow-rooting and comparatively low nitrogen-yielding TrifoUum repens. Indeed, the nitric acid is nearly exhausted in the deep-rooting Medicago sativa plot ; there remaining, to the total depth of 9 feet, only about 17 lb. of nitric-nitrogen against more than 100 lb. to the same depth in the Trifolium repens soil. The total deficiency of nitric-nitrogen in the- Medicago as compared with the Trifolium repens soil, is seen to be 85.69 lb. according to one set of determinations, and 83.94 lb. according to the other. As already said, we cannot know what was the stock of nitric-nitrogen in the soil at the commencement of the growth of the season, or the amount formed during the growing period. But, with so much more Medicago growth for several previous years, it seems reasonable to assume that there would be much more nitrogenous crop-residue for nitrifica- tion than in the case of the Trifolium repens plot. Increasing But, even supposing for the sake of illustration, that each armwnts of year's growth would leave crop-residue yielding an amount of llcuxZnt- nitrogen as nitric acid for the next crop, or succeeding crops, ed for. approximately equal to the amount which had been removed in the crop, the increasing amounts of nitrogen yielded in the crops from year to year could not be so accounted for, and . there would remain the amount of nitrogen in the crop- residue itself still to be provided in addition. In fact, as- suming the proportion of nitrogen in the crop-residue to that in the removed crop to be as supposed in the above illustra- tion, nearly 700 lb. of nitrogen would have been required for the Medicago crop and crop-residue of 1884. Or, if we as- sume the nitrogen in the residue to be only half that in the crop, about 500 lb. would have been required. Doubtless, however, some of the nitrogenous crop-residue would accumu^ late from year to year. mtrie acid The results can leave no doubt that the Trifolium repens, "mu^m ^^^ *^® Medicago sativa, have each taken up much nitrogen of nitrogen from nitric acid within the soil, and that, in fact, nitric acid f^^T is an important source of the nitrogen of the Leguminosee. orops. Indeed, existing direct experimental evidence relating to nitric acid, carries us quantitatively further than any other line of explanation. But, it is obviously quite inadequate to account for the facts of growth, either in the case of the LEGUMINOUS CKOPS. 135 ■ Medicago sativa after the clover, or iu that of the clover after the beans. It is obvious that if nitric acid were the source of the Another whole, there must have been a great deal formed by the ^°^l'^„°{ nitrification of the nitrogen of the subsoil. A difEiculty in the way of the assumption that nitric acid is the exclusive, or even the main source of the nitrogen of the Legumiuosae is, that the direct application of nitrates as manure has com- paratively little effect on the growth of such plants. In the case of the direct application of nitrates, however, the hitrie acid will percolate chiefly as sodium- or calcium-nitrate, un- accompanied by the other necessary mineral constituents in an available form ; whereas in the case of nitric acid being formed by direct action on the subsoil, it is probable that it will be associated with other constituents, liberated, and so rendered available, at the same time. Numerous direct experiments have been made at Eoth- MtHfica- amsted to determine whether the nitrogen existing in a cS«!"'^'*'" comparatively insoluble condition in raw clay subsoil was susceptible of nitrification; and the methods and results have been described in various papers. It was established that the nitrogenous matters of raw clay subsoils, which con- stitute an enormous store of already combined nitrogen, are susceptible of nitrification if the organisms, with the other necessary conditions, including a sufficient supply of oxygen, are present. It was further indicated, not only that the action was more marked under the influence of leguminous than of gramineous growth and crop-residue, but that the organisms become distributed to a considerable depth, even in raw clay subsoils, especially where deep-rooted and free- growing Leguminosse have developed. But the data at command do not justify the conclusion that. the essential conditions would be ' adequately available in such cases as those of the very large accumulations of nitrogen by the red clover grown after the beans, and of the increasing and very large accumulations by the Medicago sativa for a number of years in succession. The alternatives are — either that the plant may take up mtrogm, nitrogen from the subsoil in some other way, as ammonia or ■^™™ ?^ as organic nitrogen ; or that the free nitrogen of the atmos- the air. phere is in some way brought under contribution. In reference to the first of these alternatives, the question The power suggested itself, whether roots, by virtue of their acid sap, ^™''*'y'' may not either directly take up, or at any rate attack and gmfrom liberate for further change, the otherwise insoluble organic subsdil. nitrogen of the subsoil? 136 THE KOTHAMSTED EXPERIMENTS, Accordingly, the root-sap of many plants was examined, and it was found to be more or less acid — that of the deep, strong, fleshy root of the Medicago sativa being very strongly so, The degree of acidity of the juice was determined ; and attempts were made so to free the extract from nitrogenous bodies as to render it available for determining whether or not it would attack and take up the nitrogen of the raw clay subsoil. These attempts were, however, unsuccessful. Experiments were next made to determine the action on soils and subsoils of various organic acids, in solutions of a degree of acidity either approximately the same as that of the Medicago sativa root-juice, or having a known relation to it. These experiments and their results have been fully detailed elsewhere. It is only necessary to say here that the results did not justify any very definite conclusions as to the probability that the action of roots in the soil, by virtue of their acid sap, is quantitatively an important source of the nitrogen of plants having an extended development of roots, of which the sap is strongly acid. Siibsoii not Indeed, although significant indications have been obtained, the maim ^joth as to the importance of nitric acid as a source of the nitrogen nitrogen of the Leguminosse, and as to the action of organic Tt^Ti^ acids in rendering soluble the otherwise insoluble nitrogenous Ugtmmous compounds of soils and subsoils, yet on neither of these crops. points is the evidence at present available adequate to account satisfactorily for the facts of growth. Soil and Lastly, in regard to the sources of already combined marmre nitrogen available to our crops, the evidence points to the sowces of conclusion that, independently of the small amount of com- mtrogen bined nitrogen annually coming from the atmosphere in rain, otiwr'crops. ^'^^ ^'^ minor aqueous deposits, the source of the nitrogen, at any rate of most of our crops, is the stores already existing within the soil and subsoil, or those provided by manure. It has further been seen that the combined nitrogen is largely taken up as nitric acid, or rather as nitrates. But, it is nevertheless obvious, that we have yet to seek for an ex- planation of the source of the whole of the nitrogen of the Leguminosse, We are brought to inquire, therefore, what is the evidence relating to the question of the fixation of free nitrogen, by the plant, by the soil, or otherwise ? fixation of fkee nitrogen. 137 Evidence as to Fixation of Feee Nitrogen. It can hardly be said that there remains an unsolved prob- lem in the matter of the sources of the nitrogen of our non-leguminous crops — of wheat, of barley, and of grasses, as representatives of the great Natural Order of the Graminese ; of turnips, representing the Cruciferae ; of some varieties of beet, representing the Chenopodiacese ; and of potatoes of the Solaneae. It must be admitted to be quite otherwise so far as our leguminous crops are concerned. It is nearly a century since the question whether plants Marly ex- took up, or evolved, free nitrogen became a matter of experi- ^^^^g ment and of discussion ; and it is more than half a century thMpUmts since Boussiagault commenced experiments to determine «^o »"**■««' IT 1 ° • -1 o • nitrogen whether plants assimilate free nitrogen. from the From his results he concluded that they did not ; and '**'■• those obtained at Eothamsted more than thirty years ago confirmed the conclusions of Boussingault. In fact, we con- cluded that under the conditions of those experiments, which were those of sterilisation and enclosure, in which, therefore, the action both of electricity and of microbes was excluded, the results were conclusive against the supposition that, under such conditions, the higher chlorophyllous plants can directly fix free nitrogen, either by their leaves or otherwise. It may, in fact, be concluded that, at any rate in the case of our gramineous, our cruciferous, our chenopodiaceous, and our solaneous crops, free nitrogen is not the source. Never- theless, we have long admitted that existing evidence was insufficient to explain the source of the whole of the nitrogen of the Leguminosse ; that there was, in fact, a missing link ! A missing Limiting the discussion here mainly to the question of the ^™*" sources of the nitrogen of the Leguminosae, it is generally admitted that all the evidence that has been acquired on lines of inquiry until recently followed, has failed to solve the problem. During the last few years, however, the discussion has assumed a somewhat different aspect. The question still is, whether free nitrogen is an important The new source of the nitrogen of vegetation generally, but especially '^''*™^- of the Leguminosse? But whilst few now assume that the higher chlorophyllous plants directly assimilate free nitrogen, it is nevertheless supposed to be brought under contribution in various ways ; but especially by being brought into com- bination under the influence of micro-organisms, or of other low forms, either within the soil itself, or in symbiotic growth with a higher plant. Professor Atwater made numerous experiments, both on ^^^f * the germination and on the growth of peas. In eleven out of ments. 138 THE EOTHAMSTED EXPERIMENTS. thirteen experiments on germination, more or less loss of ni- trogen was observed. In all but one out of fifteen experiments on vegetation, there was a gain of nitrogen, which was very variable in amount, and sometimes very large. As a general conclusion, he states that in some of the experiments half or more of the total nitrogen of the plants was acquired from the air. He considered that germination without loss of nitrogen was the normal process ; that loss, whether during germination or growth, was due to, decay, and therefore only accessory. He, however, goes into calculations of some of his own results, showing by the side of the actual gains, the greater gains supposing there had been a loss of 15 per cent of nitrogen, and still greater gains if there had been a loss of 45 per cent, as in an experiment by Boussingault under special conditions. Further, he says that whilst actually observed gains are proof of the acquisition of nitrogen, the failure to show gain only proves non-fixation if it be proved that there was no libera- tion. He suggests that the negative results obtained by Boussingault and at Kothamsted may be accounted for by liberation; though he recognises that the conditions of the experiments excluded the action of either electricity or microbes. It may be remarked that, in the experiments both of Boussingault and at Eothamsted, any cases of decay were carefully observed, and the losses found explained accordingly. It may, in fact, be taken as certain that the conclusions drawn were not vitiated by any such loss. Atwater concluded that his results did not settle whether the nitrogen gained was acquired as free or combined nitrogen, by the foliage, or by the soil. He considered, however, that in his experiments, the conditions were not favourable for the action either of electricity or of micro-organisms ; and he favoured the assumption that the plants themselves were the agents. Lastly, he considered the fact of the acquisition of free nitrogen in some way to be well established ; and that thus facts of vegetable production were explained which otherwise would remain unexplained. To this, and other points involved, we shall refer again presently. mihiegeVs Of all the recent results bearing upon the subject, those of results. Hellriegel and Wilfarth with certain leguminous plants seemed to be by far the most definite and significant, point- ing to the conclusion that, although the higher chlorophyllous plants may not directly utilise free nitrogen, some of them at any rate may acquire nitrogen brought into combination under the influence of lower organisms ; the development of which is apparently in some cases a coincident of the growth of the higher plant, whose nutrition they are to serve. FIXATION OF FEEE NITROGEN. 139 It was in the Agricultural Chemistry Section of- the " Naturforscher Versammlung," held in Berlin in 1886, when one of us happened to be presiding, that Professor Hellriegel first announced his new results. Quite consistently, not only with common experience in agriculture, but also with the direct experimental results of ourselves and others, Hell- riegel found that plants of the Gramineous, the Cheno- podiaceous, the Polygonaceous, and the Cruciferous Orders, depended on combined nitrogen supplied within the soil. On the other hand, he found that leguminous plants did not depend entirely on such supplies. His results were, indeed, not only very definite, but it is seen that they had a special bearing on the admittedly unsolved problem of the source of the whole of the nitrogen of leguminous crops. In the case of these plants — that of peas, for example- — it was observed that, in a series of pots to which no nitrogen was added, most of the plants were apparently limited in their growth by the amount of nitrogen which the seed supplied. Here and there, however, a plant growing under ostensibly the same conditions grew very luxuriantly; and on examination it was found that whilst no nodules were Root-nod- developed on the roots of the plants of limited growth, "^• they were abundant on those of the luxuriantly grown plants. In view of this result Hellriegel, with his colleague Dr Wilfarth, instituted experiments to determine whether, by the infection of the soil with appropriate organisms, the formation of the root-nodules, and luxuriant growth, could be induced ; and whether, by the exclusion of such infection, the result could be prevented. To this end, they added to some of a series of experimental pots 25 or 50 cubic centi- metres of the turbid watery extract of a fertile soil, made by shaking a given quantity of it with five times its weight of distilled water, and then allowing the solid matter to subside.. In some cases, however, the extract was steriMsed. In those in which it was not sterilised, there was almost always, luxuriant growth, and abundant formation of root-nodules; but with sterilisation there was no such result. Consistent: results were obtained with peas, vetches, and some other. Leguminosse ; but the same soU-extract had little or no effect in the case of lupins, serradella, and some other plants of the family which are known to grow more naturally on sandy than on loamy or rich humus soils. Accordingly, they made a similar extract from a diluvial sandy soil, where lupins were growing well, in which, therefore, it might be supposed that the qrganism peculiar to such a soil would be present;,, and, on the application of this to a nitrogen-free soil, lupins 140 THE ROTHAMSTED EXPERIMENTS. results eon- all-import- ant. Recent trials at JRotham- sted. grew in it luxuriantly, and nodules were abundantly de- veloped on their roots. Further particulars of the experiments of Hellriegel and Wilfarth, and also of the results and conclusions of Berthelot, Dehdrain, Joulie, Deitzell, Frank, Emil von Wolff, and At- water, as well as some of the later experiments of Boussin- gault which have a bearing on the present aspect of the question, will be found in our paper in the Philosophical Transactions, vol. 180 (1889), B. A short account is also given of the experiments of Br^al in our paper in the Proceedings of the Royal Society, vol. 47, 1890. It may be added that A. Petermann found gain with lupins, but doubted whether it was entirely due to root-nodule action, or whether it was from the combined or the free nitrogen of the air. (Bull. Stat. Agron. Gembloux Belg., Marck 1890.) Thus, then, not only did Hellriegel and Wilfarth get nega- tive results with plants of other families than the Legu- minosse, as all experience would lead us to expect, but they obtained positive results with the Leguminosse, in regard to the source of the whole of the nitrogen of which experience showed that there was a " missing link." Such results were obviously of fundamental and of far-reaching'' importance ; and it seemed desirable that the subject should be further investigated with a view to their confirmation or otherwise. Accordingly, it was decided to take it up at Eothamsted, and it was hoped to commence experiments in 1887, but it was not possible to do so until 1888. In that year a preliminary series was undertaken, and the investigation has been con- tinued each year since, and is, in fact, not yet completed (1894). It is proposed to give a brief account of the conditions, and of the results, of these recent experiments made at Eothamsted, which do show a fixation of free nitrogen. But, before doing so, it will be well to call attention to those of the earlier experiments, which did not indicate any fixation ; as the well-defined difference in the conditions under which such different results were obtained will bring clearly to view what are the conditions under which fixation does, and what are those under which it does not, take place. Earlier Experiments which did not show Fixation of Free Nitrogen. Experiments on the subject were commenced at Eotham- sted in 1857; they were continued for several years, and the late Dr Pugh took a prominent part in the inquiry. The soils used were ignited, washed, and re-ignited pumice FIXATION OF FEEE NITROGEN. 141 or soil. The specially-made pots were ignited before use, and Piam of cooled over sulphuric acid under cover. Each pot with its p*^^^ plants was enclosed under a glass-shade, which rested in the sted triaXs. groove of a specially-made, hard-baked, glazed stoneware lute- vessel, mercury being the luting material. Under the shade, through the mercury, passed one tube for the admission of air, another for its exit, and another for the supply of water or solutions to the soil ; and there was an outlet at the bot- tom of the lute-vessel for the escape of the condensed water into a bottle affixed for that purpose, from which it could be removed and returned to the soil at pleasure. A stream of water being allowed to flow from a tank into a large stoneware Woulff's bottle of more than 20 gallons capacity, the air passed from it by a tube through two small glass Woulff's bottles containing sulphuric acid, and then through a long tube filled with fragments of pumice saturated with sulphuric acid, and lastly through a Woulff's bottle containing a saturated solution of ignited carbonate of soda ; and, after being so washed, the air entered the glass-shade, from which it passed by the exit tube through an eight- bulbed apparatus containing sulphuric acid, by which com- munication with the unwashed external air was prevented. Carbonic acid was supplied as required, by adding a measured quantity of hydrochloric acid to a bottle containing fragments of marble, the evolved gas passing through one of the bottles of sulphuric acid, through the long tube, and through the carbonate of soda solution, before entering the shade. In 1857 twelve sets of such apparatus were employed ; in 1858 a larger number, some with larger lute-vessels and shades ; in 1859 six, and in 1860 also six. Each year the whole were arranged side by side on stands of brickwork in the open air. The numerical results obtained in the experiments of 1857 and 1858 are summarised in Table 45 (p. 142). The upper part of the table shows the results obtained, in No assimU- 1857 and 1858, in the experiments in which no combined 5y^|^o. nitrogen was supplied beyond that contained in the seed gen. sown. The growth was extremely restricted under these conditions; and the figures show that neither with the Graminese, the Leguminosae, nor the Polygonaceae (buck- wheat), was there in any case a gain of three milligrams of nitrogen. In most cases there was much less gain than this, or a slight loss. There was, in fact, nothing in the results to lead to the conclusion that either of these different descrip- tions of plant had assimilated free nitrogen. The lower part of the table shows the results obtained in the experiments in which the plants were supplied with 142 THE EOTHAMSTED EXPEEIMENTS. TABLE 45. — Summary of the Results of Expeeiments made at eothamsted in 1857 and 1858, to detbeminb whether Plants assimilate Free Nitrogen. In seed and manure, if any. Nitrogen. In plants, pot, and soil. Gain or loss. WITH NO COMBINED NITROGEN SUPPLIED BEYOND THAT IN THE SEED SOWN. Oraminese fWieat '1857 \ Barley (. Barley f Wheat < Barley L Oats . 1858 [l858»i{™ {1857 Beans ,„„ /Beans 18^8 i Peas . ■Other plants 1858 Buckwheat 0.0000 -0.0001 -0.0006 +0.0007 -0.0021 -0.0018 WITH COMBINED NITROGEN SUPPLIED BEYOND THAT IN THE SEED SOWN. •Graminese ( Wheat J Wheat j Barley V. Barley f Wheat ■I Barley tOats f Wheat U858aM Barley LOats. ('1857 1858 Xieguminosje.^ (■1858 t Clover ^ 1858a 1 Beans Other plants 1858 Buckwheat 0.0329 0.0329 0.0326 0.0268 0.0548 0.0496 0.0312 0.0268 0.0257 0.0260 0.0227 0.0712 0.0711 0.0308 0.0383 0.0331 0.0328 0.0337 0.0536 0.0464 0.0216 0.0274 0.0242 0.0198 0.0211 0.0665 0.0655 0.0292 + 0.0054 + 0.0002 + 0.0002 + 0.0069 -0.0012 -0.0032 -0.0096 +0.0006 -0.0015 -0.0062 -0.0016 -0.0047 -0.0056 -0.0016 1 These experiments were conducted in the apparatus of M. G. Ville. FIXATION OF FREE NITROGEN. 143 known quantities of combined nitrogen, in the form of a solu- tion of ammonium-sulphate, applied to the soil. The effect of this direct supply of combined nitrogen was to increase the growth in a very marked degree, especially in the case of the G-ramineae. The figures show that the actual gains or losses of nitrogen ranged a little higher in these experiments in which larger CLuantities were involved ; but they were always represented by units of milligrams only, and the losses were higher than the gains. Further, the gains, such as they were, were all in the experiments with the Graminese, whilst there was in each case a loss with the Leguminosse, aud also with the buckwheat. The losses, where beyond the limits that might be expected from experimental error properly so-called, were doubtless due to decay of organic matter, fallen leaves, &c. It should be stated that the growth was far more healthy with the Graminese than with the Leguminosse, which are, even in the open field, very susceptible to vicissitudes of heat and moisture, and were found to be extremely so under the condi- tions of enclosure under glass shades. It might be objected, therefore, that the negative results with the Leguminosae are not so conclusive as those with the Graminese. Nevertheless we concluded, and still conclude, from the results of our own experiments, as Boussingault did from his, that neither the Graminese nor the Leguminosse directly assimilate the free nitrogen of the air. That, under the conditions described, the Leguminosse as well as the Graminese can take up and assimilate already combined nitrogen supplied to them, is clearly illustrated in the experiments made in 1860 with Leguminosse alone. The series comprised — three experiments with white haricot beans — No. 1 without any other supply of combined nitrogen than that in the seed, No. 2 with a fixed quantity of nitrogen applied as ammonium-sulphate, and No. 3 with a fixed quan- tity supplied as nitrate; also three experiments with white lupins — No. 1, as with the haricots, without artificial supply of combined nitrogen, No. 2 with supply as ammonium-sul- phate, and No. 3 was nitrate. Each of these two descriptions of leguminous plant showed considerably increased growth under the influence both of ammonium -sulphate and of nitrate ; indeed the growth was much more satisfactory than in the earlier experiments. Still, owing to the atmospheric conditions within the shades, the plants lost both leaves and flowers, and were, therefore, taken up earlier than they other- wise would have been. The analytical results here again in- dicated no gain from free nitrogen, either in the experiments without, or in those with, an artificial supply of combined nitrogen — in fact, the losses were greater than the gains. 144 THE EOTHAMSTED EXPERIMENTS. results. Siioh, then, were the negative results obtained when plants were grown under conditions of sterilisation and of enclosure. There was, under such conditions, no gain from free nitrogen, in the growth of either Graminece, Zeguminosce, or other plants. Berthelot's views. Recent Rotham- sted trials. Roof-nod- ules cmd gain of Becent JEoaperiments, ivhich do show Fixation of Free Nitrogen! It was about the year 1876, that M. Berthelot called in question the legitimacy of the conclusion that plants do not assimilate the free nitrogen of the air when drawn from the results of experiments in which the plants are so enclosed as to exclude the possibility of electrical action ; and later he objected to experiments so conducted with sterilised mate- rials, on the ground that, under such conditions, the presence, development, and action, of micro-organisms are excluded. So far, however, there is nothing in the recent results, either of M. Berthelot himself or of others, which can be held to invalidate the conclusion which had been drawn from the results of Boussingault, and. from those obtained at Eotham- sted — that the higher chlorophyllous plants do not directly assimilate free nitrogen. Let us now consider what are the results obtained when the conditions of growth involve neither sterilisation nor enclosure. A preliminary series of experiments was commenced in 1888, and a more systematic one in 1889. The plants were grown in specially made pots, and arranged in a glass-house. ■ In 1888 peas, blue lupins, and yellow lupins, were grown, and there were four pots of each : 1. with washed sand, and the ash of the plant added, but no supply of combined nitro- gen beyond a small determined amount in the washed sand, and that in the seed sown ; 2. with similarly prepared sand (and ash), but microbe-seeded with the turbid watery extract from a rich garden-soil ; 3. duplicate of No. 2 ; 4. with the rich garden-soil itself. There was, under the influence of soil-extract microbe seeding, considerable formation of nodules on the roots, and considerable gain of nitrogen. In 1889, as already said, a more extended series was com- menced. It included experiments with four annuals — namely, peas, beans, vetches, and yellow lupins ;' also with four plants of longer life — white clover, red clover, sainfoin, and lucerne. And, as will be seen further on, experiments were commenced in 1890 with the same four annuals, and the same four plants of longer life, on somewhat different lines from those above referred to. Eeferring to the experiments in the glass-house, it may be stated that in 1889 and subsequently a purer white sand was FIXATION OF FEEE NITROGEN. 145 used, which was washed and sterilised by heat. The ash of the plant and a small quantity of calcium-carbonate were added. There were four pots of each description of plant, excepting in the case of the white clover, of which there were five. For the peas, vetches, beans, white clover, red clover, sainfoin, and lucerne — No. 1 was with the prepared quartz sand without eoU-extract ; Nos. 2 and 3 were with the quartz sand and garden-soil extract added ; and No. 4 was with the garden- soil itself; the fifth pot of white clover receiving calcium- nitrate instead of soil-extract. Of the lupins (both blue and yellow) — No. 1 was with the prepared quartz sand without soil-extract ; Nos. 2 and 3 were with lupin-soil extract added ; and No. 4 was with the lupin sandy soil itself, to which 0.01 per cent of the plant ash was added. The analytical details relating to the experiments com- menced in 1889, and subsequently, though now completed, have not yet been published, so that numerical results cannot be given here. The following general statement of their bear- ing wiU, however, convey a clear idea of their significance and their importance. First as to the peas. There was limited growth in pot 1, Fig. 3 ex- with sand without soil - extract, and there was an entire i'^*"^- absence of nodule - formation on the roots. The increased growth in pots 2 and 3, with soil-extract, was coincident with a very great development of nodules. In pot 4, with garden- soil, itself supplying abundance of combined nitrogen, and •doubtless micro-organisms as well, there was also a consider- able development of nodules, but distinctly less than in either pot 2 or pot 3 with sand and soil -extract only. Lastly, without soil-extract, and without nodules, there was no gain of nitrogen ; but with soil-extract, and with nodule-formation, there was much gain of nitrogen ; there being many times as much in the products of growth as in the seed sown. For illustrations of the above-ground growth, see fig. 3. With the vetches, as with the peas, there was very restricted Fig. i ex- above-ground growth in pot 1 without soil-extract seeding, P^'^"^- and this was associated with very limited root-development, and with the entire absence of nodule-formation. On the other hand, the greatly extended vegetative growth in pots 2 and 3 with soil -extract was associated with an immense development of root and root-fibre, and with the formation of numerous nodules. Again, in the garden-soil, with its liberal supply of combined nitrogen as well as micro - organisms, there was much less development of roots, and less also of nodules, than in the pots with sand and soil-extract only. F'urther, without microbe-seeding, and with no nodules, there VOL. VII. K 146 THE EOTHAMSTED EXPBKIMENTS. was no gain of nitrogen ; whilst with microbe-seeding, and with numerous nodules, there was considerable gain of nitro- gen ; there being, with much less nitrogen in the seed, and about the sanie amount in the products, as in the correspond- -^.r^/S 22 -OC-roBlK Ifvsg ■JO, :^' r i^p r ,. 4 / /a SA.Nr) ^V.- ASll, SIT.RILI'/.KI). Pot 1. Pot 2. Pot 3. Fig. 3.— peas. Pot 4. \ ing experiments with peas, very many times as much nitrogen in the vegetable matter produced as in the seed sown. See fig. 4. The experiments with yellow lupins gave very striking FIXATION OF FREE NITROGEN. 147 SAM) V \SH SAMl \ ASH, >in r.ii r/(ji. MKUOK.i -SI ( i)i:ii Pot 1. Pot 2. Pot 3. Fig. 4— vetches. Pot 4. 148 THE EOTHAMSTED EXPERIMENTS. Mg. 5 ex- results. As with the other plants, sterilised sand with ash plairied. ^^^^ ^gg^j j^ p^j-g ^^ 2, and 3, but pot 4 was filled with sandy- soil from a field where lupins were growing. Pot 1 was left without microbe-seeding, but pots 2 and 3" were microbe- seeded by a watery extract of the lupin-soil instead of garden- soil as in the other cases. The results with the yellow lupins were as follows : In the sterilised quartz sand, without mi- crobe-seeding, the growth was extremely limited, both above and under ground. Under the influence of the lupin-soil extract seeding, the above-ground growth was not only very luxuriant, but the plants developed considerable maturing tendency, flowering and seeding freely. The development of the roots generally, and that of swellings or nodules on them, were also very marked. In pot 4, with the lupin-sand itself, which would supply a not immaterial amount of combined nitrogen, although the growth was fairly normal, it was, both above ground and within the soil, much less than in the pots with sand and the soil-extract only ; and the development of nodules was also less. It was concluded that the less growth in the lupin-sand itself than in the quartz sand with the lupin- soil extract was largely due to the much less porosity of the lupin-soil, especially when watered. Again, as with the peas and vetches, so with the lupins, without microbe-seeding there was very limited growth, no formation of nodules, and no gain of nitrogen ; but with microbe-seeding there was luxuriant growth, abundant nodule- formation, and, coincidently, great gain of nitrogen. There was, in fact, very many times as much nitrogen in the products of growth as in the seed sown. See fig. 5. In the experiments with the fourth annual, the beans, the plants suffered much from aphis ; the growth was con- sequently very limited, and the gain of nitrogen but small. Results The results with peas, vetches, and yellow lupins are, how- defimteand ever. Very defiaite and very striking. They are abundantly ^"' ^' illustrative of the fact that, under the influence of suitable microbe-seeding of the soil, there is nodule-formation oa the roots, and, coincidently, increased growth, and gain of nitrogen beyond that supplied in the soil and in the seed as combined nitrogen ; presumably due to the fixation, in some way, of free nitrogen.^ As already said, experiments were also made with four plants of longer life — white clover, red clover, sainfoin, and lucerne. ^ M. M. Sohloesing fils and Laurent have shown, by growing Leguminossa in closed vessels, and by the analysis of the air before and after growth, that free nitrogen disappeared, in quantity closely corresponding to that gained in growth ; thus establishing the fact that the source of the gain was free nitrogen (Compt. Rend. cxi. 750). FIXATION OF FREE NITROGEN. 149 The white clover was sown in July 1890. Pot 1 was with sand and ash without microbe-seeding; pots 2 and 3 the same with microbe-seeding ; pot 4 with garden-soil ; and pot 5 with sand and ash, sterilised, but with calcium-nitrate added. Pot 1 gave no cutting, but pots 2, 3, 4, and 5, each gave many cuttings ; and the plants were not taken up until December 1892. On the roots of the plants in pot 1 with- out microbe-seeding there were no nodules, and there was IS M» ^O vA.Mi'*; \Sil, v^ II CiN .. vri;KiiJ/.r:i) . MK. KOIH-M 1 lll.U 'NAMr\-s(ll 1 . POTl. Pot 2. Pot 3. Pot i. Fig. 5.— YELLOW LUPINS. extremely limited growth ; on those in pots 2 and 3 with microbe-seeding there were many nodules, and in each case the produce contained several hundred times as much nitro- gen as that in pot 1. There was obviously, therefore, great Great gain gain. The plants grown by the nitrate also contained several ^^'^(■rogen. hundred times as much nitrogen as those in pot 1, but there were no nodules on the roots. The red clover was sown in July 1889, yielded many 150 THE ROTHAMSTED EXPERIMENTS. Microbe- infection and nod- ulefornw,- tion. Ndbbe's cuttings, and was not taken up until the winter of 1890-91. Pot 1, without soil-extract seeding, obviously became acci- dentally microbe-seeded ; the growth was considerable, there were nodules on the roots, and there was considerable gain. There was also much nodule-formation, and there was great gain of nitrogen, under the influence of the soil- extract seeding, but less than in the case of the white clover. The sainfoin was sown in June 1890, and the growth was very limited — supposed to be accounted for by imperfect microbe-infection of the roots— and the gain was accordingly but small. The lucerne grew much better than the sainfoin ; the roots were much more infected by the microbe-seeding, and there was accordingly considerable gain of nitrogen. In reference to the failure of growth in the cases where it was apparently due to failure to obtain suitable microbe- infection, it has already been said that Hellriegel at first found great difficulty in ensuring a good result with lupins, serradella, and some other plants, among which was red clover; and the failure to obtain good results at Eothamsted with both blue and yellow lupins in 1888, and with blue lupins in 1889, was doubtless partly due to the same cause. As bearing upon this curious and interesting point, it will be well briefly to refer here to the experiments and results of Professor Nobbe on this subject.^ He undertook an in- vestigation to determine whether leguminous trees, as well as our agricultural leguminous plants, were susceptible to microbe-infection and nodule-formation ; and also to ascer- tain whether there is one nodule - forming bacterium, or whether many bacteria have the property — each description of plant, or perhaps each group, having its special bacterium. The plants he experimented upon were peas, yellow lupins, and beans ; also as trees Rdbinia pseudacacia (locust-tree), Cytisus laburnum (laburnum), and Gleditschia triaeantha (honey locust). To each of these he applied microbe-seeding from various sources ; in some cases only soil-extracts, and in others pure cultivations, either from soil-extracts or from the root-nodules of different plants. When soil-extracts only were used, the results were somewhat irregular. But when pure cultivations were employed, the general result was that more effect was produced on any particular description of plant by the bacteria obtained from the same description than by those derived from other descriptions. Nobbe con- cluded that the results can leave no doubt that the pea and the Bobinia bacteria have different physiological actions ; ^ Versuche iiber die Stiekstoff-Assimilation der Zeguminosen. F. No'b'be, E. Schmid, L. Hiltner, E. Hotter, Versuchs-Stationen, xxxix. 327. FIXATION OF FREE NITROGEN. 151 which indicate, if not different species or varieties, at any rate different race or nutrition modifications. Beyerinck also concluded that the various papilionaceous bacteria differ more than he had formerly supposed. Of the three descriptions of leguminous trees upon which Nobbe experimented, the Rohinia and the Cytisus, which are both of the papilionaceous subdivision of the leguminous Order, were susceptible to microbe-infection and nodule- formation on their roots, and showed coincidently gain of nitrogen ; but the Gleditschia, which is not papilionaceous, but of the sub-order Csesalpinieae, was quite indifferent to such infection, although both soil-extracts and pure cultivations from various sources were tried. On the other hand, it was found that the application of calcium-nitrate and ammonium- sulphate gave considerably increased growth. Nobbe observes that the roots of Gleditschia have a very thick covering, which it would be at any rate difficult for the bacteria to penetrate ; but whether the members of this group generally behave differently from the Papilionaceae in this respect remains for future investigation to determine. It is at any rate of inter- est to note, that the only leguminous plant outside the papil- ionaceous sub-order which has yet been experimented upon has not been found susceptible to infection, or to have nodules on its roots. In 1891, F, Nobbe, E. Schmid, L. Hiltner, and E. Hotter, Phydoiogi- commenced various experiments to ascertain the physiologi- i^l^^jj. cal meaning of the root-nodules of various wo?i-leguminous nodules. plants {Eleagnus, Hippofhae, and Alnus). Meagnus sprouts were planted in two pots containing sterilised nitrogen-free sand ; a week afterwards one pot was infected with an extract of Meagnus soil. The infection had no visible effect during the whole summer, but in the autumn one of the plants began to acquire a somewhat fresher green colour than the others, and in the spring of the following year this plant was unmistakably more vigorous than the others ; it was strong, and had side shoots. All the plants (of both pots) were iso- lated in nitrogen-free sand, when it was seen that only the plant which was benefited by the inoculation had nodules. The non-infected plants were scanty and without side shoots. Only one of the infected plants began to get greener in July 1892; it had three small oblong nodules when taken up. There was no doubt that Meagnus was enabled by the possession of nodules to utilise free atmospheric nitrogen. The organisms which produced these nodules were obtained in pure cultivations, and were totally different from Bacterium radicicola. Here, then, we have experimental evidence of gain of 152 THE EOTHAMSTED EXPERIMENTS. Gain of by a non- legummous pumt. Various nodule- bacteria. Form of root-nod- ules. Fuller evidence nitrogen by a non-leguminous plant, but only with the coin- cidence of nodule-development on the roots. The conclusion drawn from the experiments of Nobbe — that there are various nodule-forming bacteria — is at any rate consistent with the descriptions which have been pub- lished as to the difference in the external appearance, and the distribution, of the root-nodules in the case of the peas, the vetches, and the lupins, grown at Eothamsted. Again, the nodules on the roots of lucerne growing in the field were observed at different periods of the season in 1887, and again more recently on plants taken from the field for that purpose ; and they are quite different in general external character from those on any other plants that have been ex- amined at Eothamsted. Among the Leguminosae growing in the mixed herbage of grass-land, in 1868 nodules were observed on the root-fibres of Lathyrus pratensis, especially near the surface of the soil ; on the ultimate root-fibres of Trifolium pratense ; and on the smaller rootlets of Trifolium repens. In the case of red clover growing in rotation on arable fand, an abundance of nodules has been found, both near the surface and at a considerable depth. They are generally more or less globular or oval. Some found on the main roots were more like " swellings " than attached tubercles, not, however, encasing the root, but only on one side. The greater number are, however, small and chiefly distributed on the root-fibres. Again, on the plot of rich garden-soil on which red clover has now been grown at Eothamsted for forty years in succession, very numerous nodules, chiefly globular and small, have been found on the roots ; for the most part within the first few inches of soil, but some to the depth of a foot or more, dimin- ishing, however, very much both in number and in size as the clayey subsoil was reached. Obviously much more evidence than is at present at com- mand is needed in regard to any difference in character, or relative prevalence, at different periods in the life and growth of the plant, and under different conditions of soil, both so far as mechanical state and porosity, and richness or otherwise in available supplies of combined nitrogen, are concerned, be- fore any clear conception can be attained of the connection between nodule-formation, luxuriance of growth, and gain of nitrogen. The subject in various aspects is being further investigated at Eothamsted, and some of the results so far obtained will be briefly referred to presently. FIXATION OF FREE NITEOGEN. 153 How is the Fixation of Nitrogen to he explaiTied ? Eeviewing the whole of the results which have been brought Assimiia- forward, there can be no doubt that the fact of the fixation of ''^rogen free nitrogen in the growth of Leguminosse under the influence from the of suitable microbe-infection of the soil, and of the resulting ^'^]^|^. nodule -formation on the roots, may be considered as fuUy established. How, then, is it to be explained ? Unfortu- Bow is u nately there is much yet to learn before a satisfactory answer l°a|^'; can be given. Obviously we must know more of the nature and mode of life of the organisms which, in symbiosis with the leguminous plant, bring about the fixation of free nitrogen, before the nature of the action can be understood. As to the mode of life of these bodies, we owe much to the investigations of Marshall Ward, Prazmowski, Beyerinck, and others ; and some of their results have been discussed in our papers. But the facts which they have established so far are insufficient to afford an adequate explanation of the phenomena involved. Nobbe, also, has recently published results on the subject. It has, indeed, been assumed that the activity of the process One as- depends on the quantity of the nitrogenous compounds at the ^''"'pOon. disposal of the roots — a supposition which implies that the source of nitrogen of the bacteria is the combined nitrogen in the soil. The experimental results which have been described clearly show, however, that the nodules may develop very plentifully in a nitrogen-free soil, and that there may, under such conditions, be great gain of nitrogen if only the soil be suitably infected ; nor would there be any such actual gain of nitrogen in nitrogen-free soils as there undoubtedly is, if the source of the nitrogen, either of the parasite or of the host, were essentially the supplies of combined nitrogen within the soil. Further, one assumption is, that the organisms become dis- otha- tributed in the soU, both during the life of the host and after- *^^o™^- wards, and that the fixation takes place under their agency within the soil itself rather than in the course of the develop- ment of the organisms in symbiosis with the higher plant. Another is, that the fixation takes place in the soil itself under the influence of microbes existing within it, and tjiat the higher plant assimilates the resulting combined nitrogen. As bearing upon these points, it may be observed that in the experiments with peas in 1888 there was practically no gain of nitrogen within the soil itself, which it may be supposed there would have been if the fixation had taken place within it, and the host had acquired its gain from the compounds there produced. Indeed, the evidence at present at command certainly does not point to the conclusion that the gain of 154 THE EOTHAMSTED EXPERIMENTS. Boussin- gault's results. Berthelot's results. Other re- sults. ■aitrogen by Leguminosse under the influence of microbe- infection of the soil, and nodule-formation, is due to fixation by organisms within the soil itself independently of the sym- biosis. It is obvious, too, that so far as free nitrogen, may be fixed by microbes within the soil, independently of connection with a higher plant, the resulting nitrogenous compounds should, directly or indirectly, be available to plants generally whether leguminous or non-leguminous. On this point it may be remarked that, from the results of vegetation experiments made by Boussingault in 1858 and 1859, in mixtures of rich soil and sand, he concluded that free nitrogen had been fixed within the soil by the agency of mycodermic vegetation; and that the nitrogenous products which remained within it were largely in the form of organic detritus. Subsequently, however, he considered that there was not satisfactory evidence that free nitrogen is fixed within the soil under the influence of the development of the lower or- ganisms. It is, nevertheless, of interest to observe that those of his results in 1858 and 1859 which showed any material gain of nitrogen, either in the vegetable matter grown or in the soil, were obtained with Leguminosse; and that, in the case in which there was the greatest gain in the plants them- selves, he records that there were numerous tubercles on their roots. In one other case in which, however, only sand was used as soil, and the gain in the plant was but small, he also observed tubercles on the roots. It is at any rate very sig- nificant, when viewed in the light of recently acquired know- ledge, that in all the cases of gain the plants grown were of the leguminous family, and that in some of them nodules were observed on the roots. Again, Berthelot's experiments showed fixation of free nitrogen by the agency of microbes within the soil, both in the absence of higher vegetation, and also coincidently with the growth of non-leguminous plants. He further considered that such fixation takes place to an extent which would be an important source of nitrogen to our crops. As referred to above, Boussingault's experiments of 1858 and 1859 showed fixation within the soil which he then attributed to the agency of mycodermic vegetation. The fact of such fixation within the soil, under the infiuence of lower plants, has also been confirmed by the recent results of some other experi- menters. _ Thus, M.M. Schloesing/Zsand Laurent have shown fixation in bare soil, and in soils growing various non-legu- minous plants, when certain Lichens and Algse were developed, but not when their occurrence was prevented. Hellriegel has also found fixation coincidently with the growth of certain Algae. Nevertheless, it may be observed that neither expe- FIXATION OF FREE NITROGEN. 155 rienee in practical agriculture, nor the nitrogen statistics of utaegain soils and crops, points to the conclusion that there is gain of ^™fS nitrogen to any material extent by the fixation of free nitro- leguminous gen under the agency of microbes within the soil indepen- ff'^"^^^- dently of leguminous growth. It was our intention to com- mence experiments on this subject at Eothamsted in 1891, but we. have not yet been able to do so. In 1888, however, Berthelot made numerous experiments with Leguminosffi, and in many of them he found very large gains of nitrogen — indeed a much higher range of gain than in his other experiments. That there should be large gain under such conditions is quite consistent with the results which have been recorded of the experiments made at Rothamsted with Leguminosse, and with those previously obtained by HeUriegel and Wilfarth. Further, these results of Berthelot, like those obtained at Eothamsted and by others with leguminous plants, are consistent with well-established facts of agricultural production, and with the nitrogen statis- tics of soils and crops, and serve, with them, to aid the solu- tion of long - recognised problems in connection with' the growth of leguminous crops. But whether or not it may eventually be established that Lower nitrogen is fixed to any material extent by microbes within organisms the soil, independently of leguminous growth, there is evi- food/or dence that in soils and subsoils containing organic nitrogen, ^jsi^'^ lower organisms may serve the higher plants by taking up or attacking and bringing into a more readily available condi- tion combined nitrogen not otherwise, or only very slowly, available for the higher plants. For example, it is probable that fungi generally derive nitrogen from organic nitrogen ; and in the case of those of fairy rings there can be little doubt that they take up from the soil organic nitrogen which is not available to the meadow plants; and that on their decay their nitrogen becomes available to the associated herbage. Then in the case of the fungus-mantle observed by Frank on the roots of certain trees, it may be supposed that the fungus takes up organic nitrogen, and so becomes the medium of the supply of the soil-nitrogen to the plant. More pertinent still is the action of the nitrifying organisms in rendering the organic nitrogen of the soil and subsoil available to the higher plants. It may well be supposed, therefore, that there may be other cases in which lower organisms may serve the higher, bringing into a more avail- able condition the combined nitrogen already existing, but in a comparatively inert state, in soils and subsoils. It may, then, be considered as fully established, that vari- Points ous Leguminosae acquire a considerable amount of nitrogen by ^t'^i'^^^- 156 THE EOTHAMSTED EXPERIMENTS. A point still un- settled. Peter- mann's trials. Barley not able tojkc free nitro- Lower amd thefix- mtf of free the fixation of free nitrogen under the influence of the sym- biotic growth of their root-nodule-microbes and the higher plant ; that there is also fixation to some extent, but quanti- tatively of much less importance, by microbes within the soil ; and that there is fixation to some, but to a comparatively immaterial amount, by lower vegetation — such as Fungi, Lichens, and some Algee. Further, it is established that there is gain from free nitrogen in the case of some mow- leguminous higher chlorophyllous plants — Meagnus, for example — but as in the case of the Leguminosse, with the coincidence of root-nodule-microbe development. There still- remains the question — ^Whether there is any fixation by the higher chlorophyllous plants themselves, independently of the associated growth of lower organisms ? Frank maintains that there is such fixation by various now- leguminous plants. In 1892, A. Petermann published the results of experiments with barley in which he found gain of nitrogen, which he attributed to fixation by the plant. He at the same time observed that the surface of the soil was partially covered with Algse. In 1893, he published the results of further experiments, in which he grew barley both with and without sterilisation. He found no gain with sterilisation, and attri- buted .that shown without it to the lower vegetation with which the surface of the sand was more or less covered. He concluded that barley is not able to fix free nitrogen; but that soils covered with lower vegetation become richer in nitrogen. He considered that the gain in his earlier experi- ments was not due, as he then supposed, to fixation by the barley itself, but was brought about by the Algse growing on the surface of the sand. His conclusion was that free nitrogen is not fixed either by the higher plants, or by soil free from lower vegetation. Liebscher, from the results of an elaborate series of experiments with various plants, includ- ing white and black mustard, concluded that these cruciferous plants have the power of fixing the free nitrogen of the air, but whether with or without the co-operation of soil-organ- isms, he considered was not proved. Lotsy, on the other hand, from the results of experiments 'with the same plants, con- cludes that there is no such fixation with sterilisation, and that it is uncertain whether it takes place under unsterilised conditions. The question is one of practical as well as scien- tific interest, as these plants are among those grown for green manuring. Liebscher's experiments certainly appear to have been conducted with very great care under the conditions selected. Nevertheless, it is difficult to accept so important a con- clusion from the results of experiments in which from about FIXATION OF FREE NITE06EN. 157 11 to 17 kilograms of soil were employed ; iu which seldom less than 10, aud frequently nearer 25 grams of combined nitrogen were involved ; in which, with these quantities, the soils and plants were exposed to free air and rain ; and in which, under such conditions, there was, with the same de- scription of plant, sometimes loss and sometimes consider- able gain of nitrogen indicated. In the case of Papilionacese growing in sand, without or with only comparatively small additions of combined nitrogen, but with due microbe-infec- tion, inducing root-nodule-formation, the gains are propor- tionally so great as to render immaterial the usual sources of error incident to experiments in the open air, and to leave no doubt whatever whether there had been fixation or not. At Fixation present, therefore, it must be considered that the fixation of "//''«« free nitrogen by the higher chlorophyllous plants themselves theUgher still requires confirmation. It may be added, that what is ^^'-orophyi- known of the nitrogen statistics of the growth in agriculture ■^ufres^ " of other cruciferous plants is adverse to the supposition that cmfirma- they avail themselves of the free nitrogen of the atmosphere. But to return to the question of the explanation of the undoubted fixation of free nitrogen in the growth of legu- minous crops under the influence of suitable microbe-in- fection, and of the development of nodules on the roots of the plants. As in the exact quantitative series of experiments made at Eothamsted in 1888 and since, some of the results of which have been briefly described, the plants were not taken up until they were nearly ripe, it is obvious that the roots and their nodules could not be examined during growth, but only at the conclusion ; when, if the gain of nitrogen be connected with their development, it would be supposed that they would be to a great extent exhausted of their nitrogenous contents. Another series was therefore commenced in 1890, a recent and is still in progress, in which the same four annuals — experiment. peas, beans, vetches, and yellow lupins, and the same four plants of longer life — white clover, red clover, sainfoin, and lucerne — were grown in specially made pits, so arranged that some of the plants of each description could be taken up, and their roots and nodules studied, at successive periods of growth : the annuals at three periods — namely, first when active vegetation was well established ; secondly when it was supposed that the point of maximum accumulation had been approximately reached ; and thirdly when nearly ripe : and the plants of longer life at four periods — namely, at the end of the first year, and in the second year when active vegeta- tion was re-established, when the point of maximum accumu- lation had been reached, and lastly when the seed was nearly 158 THE EOTHAMSTED EXPERIMENTS. Growth of root- ripe. Each of the eight descriptions of plant was grown in sand (with the plant ash), watered with the extract from a rich soil ; also in a mixture of two parts rich garden-soil and one part of sand. The pits, with their plants, were exposed to the open air, but protected from heavy rain. In the sand the infection was comparatively local and limited, but some of the nodules developed to a great size on the roots of the weak plants so grown. In the rich soil the infection was much more general over the whole area of the roots, the nodules were much more numerous, but generally very much smaller. Eventually the nodules were picked off the roots, counted, weighed, and the dry substance and the nitrogen in them determined. Among the annuals the peas, and among the plants of longer life the sainfoin, showed perhaps the most normal growth ; and the results given in Table 46 afford interesting illustrations. TABLE 46. — Experiments at Eothamsted on the Fixation of Free Nitrogen. Plants grown in pits, and taken up at successive periods, 1890-91. 1. In sand (with ash), microbe-seeded ; 2. In a mixture of rich soil and sand. Date of taking up. Number of plants. Approxi- mate number. Nodules. Weight, dried at 100° C. Nitrogen. In dry. Actual. PEAS, 1890. rlst period In sand-! 2nd n tSrd „ In soil "{ 1st period 2nd „ 3rd „ Aug. 4 Sept. 24 Nov. 29 Aug. 5 Sept. 26 Deo. 2 grams. per cent. 3 (253) 0.229 6.630 3 (335) (328) 0.616 3.592 3 0.162 2.104 3 (324) (1353 (1512) 0.743 5.022 3 1.497 3.167 3 1.600 2.797 grams. 0.0152 0,0185 0.0034 0.0373 0.0474 0.0447 SAINFOIN, 1890-91. rlst period i-H^rd ;; fist period Uth „ Dec. 10, '90 May 15, '91 June 12, '91 Sept. 11, '91 Deo. 13, '90 May 15, '91 June 12, '91 Sept. 14, '91 3 82 0.153 7.346 3 148 0.229 5.792 3 (360. 1.043 6.151 3 (2891) 4.403 4.735 3 (226) 0.040 , 6.259 3 (2018) 1.492 6.286 2 (1125) 0.649 6.363 3 (2412) 3.299 7.066 0.0112 0.0133 0.0641 0.2085 0.0025 0.0937 0.0412 0.2331 FIXATION OF FREE NITEOGEN. 159 It is seen that, stated very briefly, the general result was mtrogen that at the third period of growth of the peas in sand, the ™J°f " amount of dry matter of the nodules was very much dimin- ished, the percentage of nitrogen in the dry matter was very much reduced, and the actual quantity of nitrogen remaining in the total nodules was also very much reduced ; in fact, the nitrogen of the nodules was almost exhausted. The peas grown in rich soil, however, maintained much more vegeta- tive activity at the conclusion, and showed a very great increase in the number of nodules from the first to the third period ; and with this there was also much more dry sub- stance, and even a greater actual quantity of nitrogen in the tota,l nodules at the conclusion. Still, as in the peas grown in sand, the percentage of nitrogen in the dry substance of the nodules was very much reduced at the conclusion. In the case of the plant of longer life — the sainfoin — there was, both in sand and in soil, very great increase in the number of nodules, and in the actual amount of dry substance and of nitrogen in them, as the growth progressed. The per- centage of nitrogen in the dry substance of the nodules also showed, even in the sand, comparatively little reduction, and in the soil even an increase. In fact, separate analyses of nodules of different character, or in different conditions, showed that whilst some were more or less exhausted and contained a less percentage of nitrogen, others contained a high percentage, and were doubtless new and active. Thus the results pointed to the interesting conclusion that An inter- in the case of the annual, when the seed is formed, and the ^^^^ "'"'- plant more or less exhausted, both the actual amount of nitrogen in the nodules, and its percentage in their dry substance, are greatly reduced; but that with the plant of longer life, although the earlier-formed nodules become ex- hausted, others are constantly produced, thus providing for future growth. The results of this new series of experiments, Hoot-nod- taken togetlier with those of the quantitative series, also serve ^^^^' '^^'^ . further to show that there is intimate connection between the nitrogmT' gain of nitrogen by Leguminosae, and the development of nodules on their roots. The alternative explanations of the fixation of free nitrogen Aitematm in the growth of Leguminosse seem to be — tiols'of'the 1. That under the conditions of the symbiosis the T^la,nt fixation 0/ is enabled to fix the free nitrogen of the atmosphere by its /"« "»^™- g&n. leaves. 2. That the nodule- organisms become distributed within the soil, and there fix free nitrogen ; the resulting nitrogenous compounds becoming available as a source of nitrogen to the roots of the higher plant. 160 THE EOTHAMSTED EXPERIMENTS. 3. That free nitrogen is fixed in the course of the develop- ment of the organisms within the nodules, and that the resulting nitrogenous compounds are absorbed and utilised by the host. tjm most Certainly the balance of the evidence at present at command is much in favour of the third mode of explanation. Indeed there seems nothing in the facts to lead to the conclusion that under the influence of the symbiosis the higher plant itself is enabled to fix the free nitrogen of the air by its leaves. Nov does the evidence point to the conclusion that the nodule -organisms become distributed through the soil, and there fix free nitrogen, the compounds of nitrogen so pro- duced being taken up by the higher plant. It seems much more consistent, both with the experimental results and with general views, to suppose that the nodule-organisms fix free nitrogen, and that the nitrogenous compounds produced are absorbed and utilised by the plant. In other words, there does not seem to be any evidence that the higher chlorophyllous plant itself fixes free nitrogen, or that the fixation takes place within the soil ; but it is much more probable that the lower organisms fix the free nitrogen. If this should eventually be established, we have to recognise a new power of living organisms — that of as- similating an elementary substance. But this would only be Lower or- an extension of the fact that lower organisms are capable of seniZ^ihe Performing assimilation-work which the higher cannot ac- higher. complish ; whilst it would be a further instance of lower organisms serving the higher. Loew's Lastly, it may be observed that Loew has suggested that the vegetable cell, with its active protoplasm, if in an alkaline condition, may fix free nitrogen with the formation of ammonium-nitrate. Without passing any judgment on this point, it may be stated that it has frequently been found at Eothamsted that the contents of the nodules have a weak alkaline reaction when in apparently an active condition — that is, while still flesh-red and glistening. It will be seen that the experimental results which have been brought forward constitute only a small proportion of those obtained at Eothamsted ; and it is hoped that when the investigations and the study of them are completed, more definite answers will be forthcoming to some of the admittedly still open questions in connection with this interesting and important subject. FIXATION OF FKEE NITROGEN. 161 Of what Importance to Agriculture is the newly -recognised source of Nitrogen to Leguminous Crops ? The question yet remains, "What is the practical importance Thepracti- of the newly-recognised source of nitrogen to the Leguminosse, '^ncT^''the considered in its bearing on the known facts of agricultural new doc- production, and especially on the question of the sources of *"^- the nitrogen, not only of leguminous crops themselves, but of crops generally? Unfortunately, as in the matter of the explanation of the action by which the nitrogen is fixed, there is much yet to learn before an adequate answer can be given. Still it is desirable to report progress. It has been stated that the characteristic nodules have been found on the roots of various leguminous plants growing among the mixed herbage of grass-land, and also on those of others growing on arable land, in the ordinary course of agriculture. There can be little doubt that when such plants are growing in soil and subsoil containing an abun- dance of combined nitrogen, they will obtain some of their nitrogen from nitrates, or other ready-formed compounds of nitrogen. An apparent difficulty in the way of the assump- tion that much of the greater assimilation of nitrogen by the leguminosae than by other plants is due to a supply of nitric acid by the nitrification of the combined nitrogen of the subsoil is, that the direct application of nitrates as manure has comparatively little effect on the growth of such plants. In the case of the direct application of nitrates, however, the nitric acid will percolate chiefly as sodium- or calcium-nitrate, unaccompanied by the other necessary mineral constituents in an available condition ; whereas in the case of nitric acid being formed as a result of action on the organic nitrogen of the subsoil, it is probable that it will be associated with other constituents liberated, and so rendered available, at the same time. But, so far as the plants do obtain nitrogen derived from the fixation of free nitrogen, the question arises. Under what conditions will this supply come the more or less into play 1 In the later series of experiments made at Eothamsted, Thefonm- those conducted in pits in the open air, to which brief *^^^^°°*' reference has been made, the general, though not the in- andfixa- variable, result was, that there was a much greater number ^^^°{£l'^ of nodules formed on the roots of the plants growing in rich soil than on those grown in sand. But whilst as a rule the individual, but much fewer, nodules on the roots grown in sand, developed to a much greater size, the much larger number in the soil were very much smaller. As to the smaller number of nodules formed in sand thaii VOL. VII. L 162 THE EOTHAMSTED EXPERIMENTS. in rich soil, the explanation may simply be that, as in the sand the infection was dependent on the additions of rich- soil-extract only, the diffusion of the microbes would be only limited, and the infection of the roots therefore only local or accidental; whilst the much greater size of the individual nodules may be due to the want of power in the more weakly plant growing in nitrogen-free soil to resist the free develop- ment of the parasite. On the other hand, in the mixture of rich soil and sand, the microbes would probably be distributed throughout it, and the roots accordingly exposed to infection along their whole range. The much less development of the individual but more numerous nodules in the rich soil may be due to one of two very different causes. It may be that although the more vigorous plants grown in the rich soil could not resist the original infection, they were able to resist the further development of the parasite. Or, it may be that with the vigorous growth, the nodules were more rapidly ex- hausted of their contents to feed the host. It will be obvious that on the former supposition, some of the nitrogen of the restrictedly developed individual nodules may have been ob- tained from the nitrogenous matters of the plant itself, derived from soil-nitrogen ; in which case the gain from fixation would be less than would otherwise be indicated by the great num- ber of the nodules produced ; and in favour of this supposi- tion, which implies that in the early stages of the infection the bacteria derive nitrogenous nutriment from the stores of the higher plant itself, and only later from the fixation of free nitrogen, is the fact of the observed " nitrogen hunger stage " so characteristic of plants for some time after infection' when growing in nitrogen-free soil ; probably indicating that during that period the limited stores of the plant are being drawn upon. On the second supposition, on the other hand — namely, that the smallness of the nodules was due to their rapid ex- haustion by the host — it might be that more of the nitrogen of the nodules would be due to fixation, and that hence a larger proportion of the total nitrogen of the plant would be gain attributable to that source. Obviously more evidence is needed before a decisive opinion can be formed as to how far fixation of free nitrogen is an essential coincident of nodule-development at all its stages of accumulation, and how far, therefore, the amount of nodule- formation may be taken as a fair measure of the fixation. It is to be supposed that when nodules develop abundantly on the roots of leguminous plants growing in soil rich in readily available combined nitrogen, the nitrogen assimilated will be partly due to soil-supplies of combined nitrogen, and partly to fixation. That there is gain when red clover, for FIXATION OF FREE NITROGEN. 163 example, grows luxuriantly on ordinary arable soil, coninion experience can leave but little doubt. The evidence of fixa- tion is, however, undoubtedly much the clearer in the case of soils poor in nitrogen. Thus, in the cases of the experiments with peas, vetches, and yellow lupins, growing in nitrogen- free but duly infected sand, there being no other supply of combined nitrogen excepting that in the seed sown, the pro- portion of the total assimilation due to fixation was undoubt- edly very large. It may safely be concluded, indeed, that when luxuriant leguminous crops are obtained on soils characteristically poor in available combined nitrogen, a large proportion of the total nitrogen assimilated will be due to fixation. It is, on the other hand, by no means so clear that Abundant when such plants are grown in soil rich in available combined ^"^'f' "-^ , nitrogen, an abundant development of nodules is to be taken always in- as indicating that a correspondingly great proportion of the <^icaij« of total nitrogen assimilated is due to fixation. of nitrogen. There can, however, be little doubt that in the growth in practical agriculture of leguminous crops, such as clover, vetches, peas, beans, sainfoin, lucerne, &c., at any rate some, and in some cases a considerable proportion, of the large amount of nitrogen which they contain, and of the large amount which they frequently leave as nitrogenous residue in the soil for future crops, is due to the fixation of free nitrogen, brought into combination by the agency of lower organisms. Evidence is, however, obviously still wanting, to enable us to judge decisively under what conditions a greater or less proportion of the total nitrogen of the crop will be derived — on the one hand from nitrogen-compounds within the soil, and on the other from fixation. Incidentally the question suggests itself. How far the Causes of failure of red clover, or of other leguminous crops, may be «^<"'«'"-s»<'*- due to the exhaustion of the organisms necessary for nodule- development, and for the coincident fixation of free nitrogen ; how far to the exhaustion of combined nitrogen, or of the necessary mineral constituents, in an available condition, within the range of the roots ; or, as is sometimes the case, to insect ravages due to the condition of the soil indepen- dently of an otherwise failing condition of the plant ? Assuming it then to be established that a greater or less. Sources of and sometimes a considerable proportion, of the nitrogen of ^^'J°^-^"^ our leguminous crops will be dae to fixation under the con- ditions supposed, it is obvious that such a fact not only serves to explain the source of the hitherto unaccounted for amount of the nitrogen of those crops themselves, but that it also affords an explanation of the source of the increased amount of ni- trogen which other crops acquire when they are grown either 164 THE EOTHAMSTED EXPERIMENTS. Practical aspects of tM subject. Enriching poor soils. An Oxford- shire ex- Alternat- ing nitro- gen-accuTH- crops and in association, or in alternation, with Leguminosge. Lastly, the fact that at any rate many leguminous plants, including papilionaceous shrubs and trees, as shown by Nobbe, are sus- ceptible to the symbiosis, and under its influence may gain much nitrogen, serves to explain the source of some at least of the large amount of combined nitrogen accumulated through ages in our soils and subsoils, and also the comparatively slow exhaustion of their stores of it by cropping, drainage, and in other ways. We will, in conclusion, refer to some of the more directly practical aspects of the subject. It may be observed that in Germany, Schultz, of Lupitz, has for some years devoted a considerable area of poor, gravelly, and sandy soil, to the growth of leguminous crops — various clovers, lupins, serradella {Ormthopus saiivus), &c., by means of kainit and phosphatie manures, and he has found that the land was thereby very much enriched for future cereal and other crops. He finds, however, that it is necessary to vary the description of legu- minous crop grown. In other parts of Germany, too, the system is gradually extending of growing lupins, serradella, or other leguminous crops, especially on poor sandy soils, with a view to their enrichment in nitrogen. And, on a large estate in Hungary, visited by one of us in 1891, it was found that the results of the recent investigations indicating the fixation of free nitrogen in the course of the development of leguminous crops were being carefully studied with a view to practical application. In our own country, Mr Mason, of Eynsham Hall, Oxford- shire, after first making some experiments with various Legu- minosae on small plots, and then a considerable series in specially built tanks or pits, devoted about 200 acres to the practical application of the recently acquired knowledge in regard to nitrogen fixation. Stated in a few words, his idea is to reduce his area under roots, and to grow instead mixed crops of Leguminosse — beans, various clovers, &c. — ^liberally manured with basic slag and kainit, and to convert the pro- duce in the first year into silage, and in the second into hay. The land is thus occupied for two years ; and the assumption is that in this way highly nitrogenous crops will be obtained with mineral, but without any nitrogenous manure, and that the land will be left in high condition so far as nitrogen is concerned, for the growth of saleable crops, such as grain and potatoes, which require nitrogenous manuring. In other- words, his plan is, as he puts it, first to grow nitrogen- accumulating crops for home consumption, and afterwards nitrogen-consuming crops for sale. The experiment has been FIXATION OF FKEE NITROGEN. 165 in progress too short a time to judge how far it will be nitwgen- successful in a series of years, or of rotations. ccmsmning There is, of course, nothing new in the fact that after the growth of a leguminous crop, such as red clover, for example, the soil is left in a higher condition for the subsequent growth of a grain crop ; and that, in fact, the growth of such a legu- minous crop is to a great extent equivalent to the application of a nitrogenous manure for the cereal. Indeed, history tells The Ro- us that more than two thousand years ago it was recognised 7"^ !*"^ by the Eomans that the occasional growth of plants of the day. leguminous Order had the effect of increasing the growth of the gramineous crops with which they were alternated ; and it was stated that the effect was equivalent to that of apply- ing manure. Thus Varro says that " certain things are to be sown, not with the hope of any immediate profit being derived from them, but with a view to the following year, because being ploughed in and then left in the ground, they render the soU afterwards more fruitful;" and the plants used for this purpose were lupins, beans, vetches, and other legumes. Now, however, that the character of the action is more clearly understood — and it is certain that there is actual gain of nitrogen from sources external to the soil itself — it seems desirable that at any rate tentative trials should be made on different descriptions of soil, with a view of ascertaining whether more advantage cannot be taken of this source of nitrogen than our established practices of rotation at present secure. To sum up — the experimental results which have been Summary brought forward clearly establish that there is great gain of "f'^^i^i'ts. nitrogen under some conditions. It has also been clearly shown that due infection of the soil, and of the plant, is an essential to success. The evidence at the same time points to the conclusion that the soil may be duly infected for the growth of one description or some descriptions of leguminous plant, but not for some other descriptions. The field experi- ments on such plants at Eothamsted have further shown that land which is, so to speak, quite exhausted so far as the growth of one leguminous crop is concerned, may still grow very luxuriant crops of another description of the same Order, but of different habits of growth, and especially of different character and range of roots. This result, though undoubtedly more or less due to other causes also, is, never- theless, in some cases doubtless dependent on the existence, the distribution, and the condition, of the appropriate microbes for the due infection of the different descriptions of plant. In fact, it is pretty certain that success in any system involv- ing a more extended growth of leguminous crops in our 166 THE EOTHAMSTED EXPEKIMENTS. rotations, will not be attained without having recourse to a considerable variation in the description grown. Other essential conditions of success will generally be the liberal application of potash and phosphatic manures, and some- times chalking or liming, for the leguminous crop. Then the questions would arise, How long the leguminous crop should occupy the land; to what extent it should be con- sumed on the land, or the manure from its consumption be returned ; or under what conditions the whole, or part, of it should be ploughed in ? Lastly, it is probable that more benefit would accrue to the lighter and poorer than to the heavier or richer soils by any such extended growth of leguminous crops. SECTION IV. — EXPERIMENTS ON TEE GROWTH OF WEEAT FOR MORE THAN FIFTY YEARS IN SUC- CESSION ON TEE SAME LAND; BROADBALK FIELD, ROTHAMSTED. Inteoduction. WhAiat and It has been already pointed out, that although wheat and larUy com- barlev are closely allied botanically, and they have in some ^ ■ respects very similar requirements, yet that there are dis- tinctions as well as similarities which have to be borne in mind. Thus, whilst in our country and climate barley is generally sown in the ^ring, wheat is almost always sown in the autumn, and thus has four or five months for root-devel- opment, and for gaining possession of range of soil, before barley is sown. In the United States, on the other hand, wheat is to a great extent both a spring and an autumn sown crop ; whilst in some other exporting countries it is in some cases a spring and in others an autumn sown crop. At any rate, it is so important a crop in many countries of the world that results relating to its growth, even under widely different conditions, can hardly fail to be of interest to foreign as well as to home growers. The Field Experiments on Wheat. Plan of the The experiments on the continuous growth of wheat at !f^w!f/o Eothamsted were commenced in the autumn of 1843, the first experimental crop being harvested in 1844 ; so that the crop of 1894 was the fifty-first grown in succession on the same land — 1. Without manure. 2. With farmyard manure. 3. With a great variety of chemical manures. ■WHEAT. 167 Table 47 (p. 168) gives the number of bushels of dressed grain per acre, without manure, and with farmyard manure, in each of the 51 years, 1844 to 1894 inclusive ; also on some of the artificially manured plots, mainly selected to illustrate the effects of exhaustion and of manure-residue. In most cases in this table, and in all in the subsequent tables, the results obtained on the artificially manured plots are only given for the last 43 of the 51 years; as, during the first 8 years, various mineral and nitrogenous manures were applied, but not as a rule the same from year to year on the same plot, as they were subsequently. Without Manure every year. After a five - course rotation since manuring (turnips, barley, peas, wheat, oats), the first experimental wheat crop was harvested in 1844. The highest yield of the whole series of years without manure was 23;^ bushels in 1845, and the lowest 4f bushels in 1879. Other yields have been 21i bushels in 1854, 20 in 1857, only 5| in 1853, and only 8-9 bushels in 1867, 1875, 1876, and 1877. The upper part of the table (47) shows that the average Produce of produce without manure over the first 8 years, 1844-51, ^,^^_"" was 17f bushels, which was higher than over either of the subsequent 8 -yearly periods, due doubtless to a greater amount of comparatively recent accumulations from the previous treatment. In the bottom division of the table is given the average produce for each of the subsequent 8- yearly periods, and for the 40 years, 1852 to 1891 inclusive ; also for the whole period of 51 years, 1844-94. It is seen that, without manure, the average annual produce over these 8-yearly periods was — 16|, 13|, 12J, 10^, and 12| bushels; over the 40 years (1852-91) 13, and over the 51 years (1844-94) 13| bushels. There can be no doubt that the produce of the unmanured Sail ex- plot has gradually declined ; and, independently of the evi- ^^i^"^- dence of diminishing produce, analyses of the soil at different periods show that there has been a gradual diminution in the amount of nitrogen in it. But owing to the great fluctuations in the amount of produce from year to year dependent on season, it is by no means easy to estimate the decline due to exhaustion of the soil, as distinguisl^ed from variations due to the seasons. In the first place, it is difficult to say what figure should Former be adopted as the standard produce of the plot by which to If^^fald. compare the yield from year to year. The whole field was manured with farmyard dung in 1839, and then grew tur- TABLE 47. — Wheat grown for ^^ x^.„ -.- 1_ :: Results showing the effects of exhaustion, and of manure-residue, per acre, f reduce : Dressed Grain in bushels. Quantities Mixed min. and amm.- Balt9=172 lb. N. IS years, Mineral manure 14 tons farmyard manure every year. Without manure every year. Mixed mineral manure alone— mm. Mixed mineral manure alone— &iMe; ammo- nlum-salta alone= 86 lb. nitrogen— yellow; alternately. 1852-64. Unmanured 19 years, 1865-83. Mixed min, and sod. nit. =86 lb. K 11 years, 188i-94. aXoue—blue; ammo- nium-salts alone=86 lb. min. and anun.-salts — green; ummmured -KiMfe. Plot Noa. 2. 3. 5. 17. 18. ill. lOn. 106. Bushels. Bushels. Bushels. Biishols. Bushels, 1844 20^ 15 15* J 1845 32 23^ ;i} 1846 ■ 27i 18 m 17-§- 1847 16 25$ 25f 1848 25§ ■14 19} 25^ ^ 1849 31 19 32f 32} 1850 28^ 15j 27 18 1851 29§ 16: 28} 28-J 8 years, 1844-61 28 171 29 30J 28J 30} 26 24} 1862- 27 13 16^ 24| 14i 28 21} 22} 1853 - 19 — ' 6- 10^ 8| 19} 25} 10 15i 1854 41 21 24; 44i 23| 49i 34} 39} 1855 34 17 18: 18 33i 82i 20 28} 1856 36 14* 19* 31 173 40J 37 24i 271 1857 41 20 23J 26} 49 29} 34J 1858 -38 18 IS^ 33^ 21J 41 22} 27} 1869 36 181 201 m 32{- 34 32| 19 26} 1860 32 iS 162 26} 16J 15} 18} 1861 34 15^ 18J 32| 37 12} 16 1862 38 16 17J 27f 18} 36; 23} 24| 1863 44 17i l4 194 21} 46} 65; 39} 43| 1864 40 161 36J 17S 51} 321 1865 37 13f 12* l*i 17 Sli 32 25} 30} 1866 82 ISi 26} 12S 17h 26} 28} 1867 27 8 9| m 23t 14| 18J- 19t 1868 41 16 m 3Y§ 18S 22-^ 24J 27} 1869 38 14 16i 16J 22| 16} 20} 19} 1870 36 15 ISi 345 19 18} 21| 23} 1871 39 9 llf 16 285 13} 10} 10 1872 32 10 12| 26} 13 13} 18 18} 1873 26 11 12| llj 20| 12} 19} 201 ■ 1874 39 11 13 33} 14 11* 26} 27 1875 28 8 9i llf 26-1 10} 12} 14} 1876 23' 8i 10^ 26} loa 11 12} 14- 1877 24 8 IH 10 126 fl} 17} 18 1878 28 12 14| 29 16} 13| 27} 29} 1879 16 4 6| 3} 20| 4* 4 *% 1880 38 f 11* 17i 32$ 16 14-1 10 13} 30 13 12i 13} 32 13} 18} 19 ' 1882 32 11 12i 31 16S lOJ 23! 26 1883 35 13i 16J 161 38i 15} 17 18J 1884 32J 13 Hi 33| 13} 35 26 27 1885 1886 40^ 36J 't' 16 11} 12J 87| 33 13} 37| 44f 24} 13i 24J 12} 34 J^ 14* m 30J 39S 20| 23 1889 38 10 12i 12 16J 32 18} 23-1 m 29 18} 111 10} 12| 1890 1891 43 48i 14 13f 14i 111 36i 14g 20 311 87} 42} 18} 20} 20} 22} 33f 9| 10| 29 12} 31| 11 12 34i 9$ 14i 12} 20-3 19} 47 8 8} 1894 45i 18 22| 37} 27} 28} 3l| AVERAGES. 8 years, 1862-59 8 years, 1860-67 8 years, 1868-76 8 years, 1876-83 8 years, 1884-91 20 years, 1862-71 20 years, 1872-91 40 years, 1862-91 51 years, 1844-94 28f 33i 16} 13} 12} 10* 12} 14} u* 13f 19 16} 14 12| 13| 17 12} 18} 16} 16 12} 13} 17} 15} 32| 31} 28} 27J 82} 37} 42il 16} 2 11} 37} 32|3 .22 4 27} 22J 24 19 16} 22i 17} 20} 1 Average of 6 years, 1860-64 inclusive. 27} 27} 20} 18} 19} 26} 19 22} I Average 20 yekrs-Srst 13 years withrnixed mineral and mTn'it.p^^^ ' Average 20 years— first 12 years iimmaiiuyei SbsI- S Soo^s .^iirei alrjcsl ,--.fi ."»'"-•?---- -?-?- WHEAT. 169 nips (fed on the land), barley, peas, wheat, and oats, before the commencement of the experiments in 1843-44. The plot then grew eight crops of wheat without manure, to 1850-51, before the commencement of the period of 40 years to which the averages which have been quoted refer. Although at the conclusion of the five-course rotation since manuring above described, the land would doubtless be, in an agri- cultural sense, so far exhausted as to require re-manuring, there can be no doubt that there would nevertheless be some accumulation due to comparatively recent manuring and cropping. It would be supposed, however, that the growth of wheat for 8 years in succession without manure would remove most, if not all, accumulation which could be attributed to comparatively recent treatment. Indeed there can be little doubt that the land would suffer more or less exhaustion during these 8 years ; but, as serving to counteract the tendency to decline in yield from exhaustion during that period, it happened that, taken together, those eight seasons were of more than average productiveness. The question of the rate of decline due to exhaustion, as Fall m distinguished from fluctuation due to season, has been made ^™*«« the subject of elaborate calculation and discussion, which haustion. cannot be gone into here; but the general result may be stated as follows: — Assuming, for reasons which were fully considered, the standard produce of the unmanured plot to have been 16 bushels per acre iadependently of material exhaustion, there was an average decline from year to year of little more than one-sixth of a bushel over the 40 years 1852-91. It remains to be seen what will be the result in the future ; and whether a point has already been, or will in time be reached, at which the produce will remain constant, excepting so far as it is influenced by the fluctuations of the seasons. It is estimated that over the period of 30 years, 1851-52 Yield of to 1880-81, the unmanured plot yielded an average of 18.6 «»':™S'«»j lb. or nitrogen per acre per annum m the crop, and lost a loss of minimum of 10.3 lb. in drainage, in all 28.9 lb.; whilst, on X^S/* the mixed mineral manure plot (5), it is estimated that the crop removed an average of 20.3 lb. of nitrogen, and that at least 12 lb. were lost by drainage, or in total 32.3 lb. Further, it is estimated that the soils lost to the depth of 27 inches about two-thirds of these amounts ; leaving, say, 10 lb., more or less, to be otherwise accounted for. Of this, the rain, &c., would supply 5 lb., or perhaps rather more, and the seed about 2 lb., so that there is but little to be provided from all other sources. Further, as at the com- mencement the soil was, agriculturally speaking, exhausted, VOL. VII. L 2 2» 170 THE EOTHAMSTED EXPERIMENTS. the nitrogen supplied by it would be largely due to old accumulations. Yield with- Lastly in regard to the produce of wheat grown eraec^T"" ^° ^^''^J years in succession without manure, it may be Ainerican observed that the average yield over 40 years, 1852-91, yield. ^-^8 13 bushels per acre per annum, which is more than the average of the whole of the United States, including their rich prairie lands ; indeed it is more than the average yield per acre of the wheat lands of the whole world! That the result is not due to richness of soil will be Nitrogen obvious from the fact that the percentage of nitrogen in m the soil, ^.j^g ^^.y. gjf^^g(j gQJj^ exclusivB of stones, from samples taken in 1893, of every 9 inches of depth, down to 12 times 9, or to a total depth of 9 feet, was, for the respective depths from the first to the twelfth, as follows: 0.1110, 0.0720, 0.0609, 0.0482, 0.0445, 0.04-36, 0.0335, 0.0284, 0.0264, 0.0214, 0.0219, and 0.0251.1 xhus, the percentage of nitrogen in the surface-soil is considerably lower than in the average of wheat-lands in Great Britain ; it is considerably less than half as high as in the case of average permanent meadow- land; and it is only about one-third as high as published analyses show in some Illinois prairie soils. Th%subsoils are also very poor in nitrogen. It is further to be observed that a full mineral manure, annually applied, gave less than f bushel per acre per annum more than the unmanured plot. Hence, it may be concluded that it was not owing to any deficiency of mineral supply, but of nitrogen, that the limita- Mffect of tion of the produce was due. On the other hand, that with dowT^ a soil so poor in nitrogen the yield was nevertheless higher than the average of the United States, or of the world at large, is to be explained by the fact that great care is taken to keep down weeds, which would otherwise appropriate a large share of such fertility as the soil possessed. Farmyard Manure every year. In the application of farmyard manure every constituent is supplied in excess. The highest yields of the series of years were— 48^ bushels in 1891, 45| in 1894, 44 in 1863,. 43 in 1890, 4l| in 1868, 41i in 1857, 41^ in 1854, 40^ in 1889, 40^ in 1885, and 40 bushels in 1864. The lowest yields were— 16 bushels in 1879, 19^ in 1853, 20^ in 1844, 23| in 1876, and 24J in 1877. The average produce per acre per annum over the first ' It should te explained that these samples were not taken in our usual series for analysis, but only from one place, specially to provide iUustrative specimens of the soil and subsoil to send to the Chicago Exhibition. WHEAT. 171 8 years was 28 bushels ; and the average over each of the Produce five subsequent 8-yearly periods was— 34f, 35|, 35f, ^H,f^-^^: and 39^ bushels. Excluding the first 8 years, the average ««. produce over the 40 years, 1852-91, was 34f bushels ; and the average for the whole period of 51 years, 1844-94, was 33f bushels per acre per annum. On the farmyard manure plot, the first depth of 9 inches Great ac- shows a great accumulation. It is about twice as rich in Ifnullgm. nitrogen as any other plot in the field ; yet this richness is not proof against bad seasons, nor are the highest amounts of produce in the field obtained on this plot. It has been seen that the unmanured plot has declined in Dung in- yield and fertility; but there can be no doubt that the ^^^f^^^*^. farmyard manure plot has, on the other hand, increased in ing in soil fertility. Analyses of the surface-soil at different periods ^^nv^iy have shown that it has become about twice as rich in the crop. nitrogen as that of the unmanured plot. It has indeed been shown, that a large amount of the constituents of farmyard manure accumulates within the soil, and that they are very slowly taken up by crops. In fact, notwithstanding this great accumulation within the soil, the wheat crops on the dunged plot seldom, if ever, show over-luxuriance ; and in unfavourable seasons the produce has been comparatively small, largely owing to the encouragement of weeds, and especially of grass, which in wet seasons it has been im- possible effectually to eradicate, and what has been done has not been accomplished without injury to the crop. Let us now endeavour to estimate the average annual increased increase of produce on the farmyard manure plot, due to ^'^"'iM accumulation, independently of fluctuations due to season, creased fer- as we did the annual decline in yield on the unmanured ^''.»^2/»« plot due to gradual exhaustion. As in the case of the un- manured plot, so in that of the farmyard manure plot, we have founded an estimate of its standard produce, irrespec- tively of material accumulation, on the yield of the first 8 years ; deducting, however, the produce of the first year of all, 1844, as although the yield of the crop of the country at large in that year was high, that of the farmyard manure plot was only 20 bushels. Taking the average of the re- maining 7 years of the 8, we get 29.3 bushels, whilst 3 of the 7 yielded more than 30, and 2 others 29 bushels or more. Adopting then 29.3 bushels as the standard yield, irrespec- tively of material accumulation, the result would be an average annual increase, due to accumulation, of 5| bushels over the 40 years ; whilst the average increase from year to year, if uniform throughout the period, would be a little over \ bushel over the 40 years. 172 THE EOTHAMSTED EXPBElMfeJfTS, prodiice. In conclusion, it is seen that the average produce of the 40 years by farmyard manure was nearly 35 bushels ; which is about 7 bushels more than the average of the United Kingdom under ordinary cultivation ; and it is not far short of 3 times as much as the average of the United States, or of the whole world ! Mineral manure alone. Addition gen. titrate v. ammonia. Loss of Increase propor- tionate to available Influence of ous man- ures on non-nitro- genous con- stituent of crops. Various Artificial Manures. The next question is, Which constituents of farmyard manure are the most effective for wheat in this agricul- turally exhausted rather heavy soil, with a raw clay subsoil ? The first illustrations on this point will be drawn from Table 48. The average of the 40 years by mineral manure alone shows an increase of only 2 bushels over that of the un- manured plot, though during the preceding 8 years (1844-51) it had received mineral and nitrogenous manures, whilst the unmanured plot had, during the same period, grown eight unmanured wheat crops. The addition to the mineral manure of the first 43 lb. of nitrogen (plot 6) gives an average annual increase of 9^ bushels; the second 43 lb. (plot 7) an increase of 9, and the third 43 lb. (plot 8) only 3| bushels increase. This result affords an illustration of the inapplicability of conclusions from manure experiments when the condition of the land is too high already, or when an excess of manure is applied. A given quantity of nitrogen in the form of nitrate, yielded more produce than an equal quantity in the form of ammonia. The nitrate, being always applied in the spring, was not subject to winter drainage. It is, however, very soluble, and becomes rapidly distributed and available; but it is at the same time very subject to drainage after sowing, if heavy rains follow. Prior to 1878, the ammonium-salts were applied in the autumn, and a great loss of nitrogen by winter drainage, chiefly as nitrates, was proved. To the loss of nitrogen by drainage reference will be made further on. Thus, minerals not being deficient, the increase was in proportion to the available nitrogen, when it was not applied in excess. It will be of interest here to refer to the influence of nitrogenous manures in increasing the production of the non-nitrogenous constituents of our crops, as illustrated in Table 34 (p. 107). It shows the estimated amounts of carbon per acre per annum in various crops grown by mineral manure without nitrogen, and by the same mineral manure WHEAT. 173 TABLE 48. — Wheat geown for more than 50 Years in sitccbssion on the SAME Land, commencing 1843-4. Results showing the effects of different Manures for 43 years, 1852-94 inclusive. Quantities per acre. Produce — Dressed Grain in bushels. Superphosphate, and Sulphates Potash, Soda, and Magnesia. Alone. And am. -salts =48 lb. nitrogen. And am.-salts = 86 lb. nitrogen. And am.-salts = 129 lb. nitrogen. And sodium nitrate =86 Ib.i nitrogen. Sodium nitrate alone = 86 lb.!! nitrogen. Plots. 6. 7. 9a. Harvests. 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1886 1886 1887 1888 1889 1890 1891 1892 1893 1894 Bushels. 164 lOi 24j ISi 19J 231 18| Bushels. 20J 18i 34J 28 271 351 Bushels. 45i 44J .Ji9i 341 Bushels. 27* 23* 48S 31* 39i 481 4ll 15f 16* 17i 19t IH 14i 13i 9i 31i 25 20* 16S 31i 36* 39* 661 49| 43S 32* 30* 17 20* 16J 25| 161 39J 28| 40* 22i 25| ibi 271 368 27* 40* 30 10* llf 14J m 12J 12* 15f 151 141 22| 10* 27 21| 23* 271 16* \l\ 14* 12 15* 14* Hi 23* 23* 31* 36| 29* S5| BOi 43* 361 42| 34* 35i 36* 22 19| 38 20i 481 38* 21} AVERAGES. Bushels. 25* Hi 37f 30 Bushels. 43* 65S 61* 44* 47J 39 46* 34* 40| 35| 40* 37* 22 34* 35* 31f 43| 40* 31* 28* 26* 31J 35 26i 17| 43| 8 years, 1862-69 . 8 years, 1860-67 . 8 years, 1868-76 . 8 years, 1876-83 . 8 years, 1884-91 . 19 16* 14 ill m 26* 22 20| 241 31 28 341 36f 391 36 38* 31* 40* 39 34i 32 26* 27 22 18 20 20 years, 1852-71 . 20 years, 1872-91 . 17 12| 21| 35* 31 381 ,34| S6| 34 26 191 40 years, 1862-91 . 15 24* 33* 36* 361 22J Excess of average crop over 1 Plot 5 in bushels j 9* 18* 21* 20| 71 1 ta. nitrate of soda, e^ual 74 lb. nitrogen in 1852 ; equal 43 lb. nitrogen in 1863 and 1864 i equal 86 lb. nitrogen lnlM4, and each year to 1884 inclusive ; and eqaal 43 lb. nitrogen in 1886, and eacb year since. Ho mineral manures applied in 1862, 1853, or ISai 2 m ' Nitrate of soda, equal 74 lb. nitrogen lnl862 ; equal 86 lb. nitrogen In 1863, and each year to 1884 inclusive j and equal 43 lb. nitrogen in 1885 and eacb year to 1893 inclusive. In 1894 manured exactly as Plot 9a. 174 THE ROTHAMSTBD EXPERIMENTS. and nitrogenous manure in addition. It also shows— the gain of carbon, that is the increased amount of it assimilated per acre, and the gain of carbohydrates, that is the increased production of them per acre, under the influence of the nitrogenous manures ; and lastly, the estimated gain of carbohydrates for 1 of nitrogen supplied in manure. The figures show that, independently of the underground growth, there was an increased assimilation of carbon per acre in wheat— of 602 lb. by the application of 43 lb. nitrogen as ammonium-salts ; of 1234 lb. by 86 lb. applied as ammonium- salts ; and of 1512 lb. by 86 lb. applied as sodium-nitrate. Or, reckoning the increased production of the non-nitrogenous bodies — the carbohydrates, by the use of nitrogenous manures, it was estimated that there was an increase of 1240 lb. of carbohydrates per acre by the application of 43 lb. nitrogen as ammonium-salts, of 2550 lb. by 86 lb. applied as ammonium- salts, and of 3140 lb. by 86 lb. as sodium-nitrate. To put it in another way — for 1 lb. of nitrogen applied as manure, there was an increased production of carbohydrates in the grain and straw of wheat — of 28.8 lb. when 43 lb. of nitrogen were applied as ammonium-salts, of 29.7 lb. when 86 lb. were applied as ammonium-salts, and of 36.5 lb. when 86 lb. were applied as sodium-nitrate. Nitrogen It is Seen that in the case of the wheat, there was much applied in more effect from a given amount of nitrogen supplied as autvrnm. nitrate, which was always applied in the spring, than from an equal quantity as ammonium-salts, which were applied in the autumn, when the nitrogen would be subject to winter drainage. Eeference to the table will also show that there was more effect from a given amount of ammonium-salts applied to barley than to wheat ; the application having been made for the barley in the spring, and for the wheat in the autumn. Depend- It should be observed that there was such greatly increased assimilation of carbon in the wheat and in the barley as the figures show, for more than twenty years, without the addition of any carbon to the soil. It is indeed certain that, in the existing condition of our old arable soils, the increased growth of our staple starch-yielding grains is greatly dependent on an available supply of nitrogen within the soil. It is equally certain that the increased production of sugar in the gram- ineous sugar-cane in the tropics, is likewise greatly depen- dent on the supply of nitrogen within the soil. In connection with the results showing the increased assimilation of carbon, and increased production of carbo-" hydrates, under the influence of nitrogenous manures, it will further be of interest to call attention to the connection ence on WHEAT. 175 between nitrogen accumulation, chlorophyll-formation, and carbon assimilation. TABLE 49.— Eblation of Carbon assiiHilation to Niteogen ACCUMULATION, AND TO CHLOROPHYLL TOEMBD. Nitrogen in dry matter. 1 Relative amounts of chlorophyll. Carbon per acre per annum. Actual. Difference. Hay. Graminese Leguminosse Wheat. Plot 10a Plot 7 .... Barley. Plot la . Plot 4a . Per cent. 1.190 2.478 (1.227) (0.566) (1.474) (0.792) 0.77 2.40 2.00 1.00 3.20 1.46 lb. 1398 2222 1403 2088 lb. -824 -685 ■1 The figures given in parentheses are on the only partially dried substance. It should be observed that the amounts of chlorophyll recorded are as stated, relative, and not actual; and the figures show the relative amounts for the individual members of each pair of experiments, and not the comparative amounts as between one set of experiments and another. It should be further stated that the chlorophyll determinations were kindly made by Dr W. J. Eussell, F.RS., of London, in specimens collected at Eothamsted, whilst the wheat and barley were still green and actively growing. It will be seen, in the first place, that the separated mtrogm. leguminous herbage of hay contained a much higher per- <*™'^?™- centage of nitrogen in its dry matter than the separated fworo? gramineous her,bage ; and that, with the much higher per- •2'^2'M. centage of nitrogen in the leguminous herbage, there was also a much higher proportion of chlorophyll. Next, it is to be observed that the wheat plant on plot 10a, manured with ammonium-salts alone, shows a much higher percentage of nitrogen than that of plot 7, with the same amount of ammonium -salts, but with mineral manure in addition. The high proportion of chlorophyll again goes with the high nitrogen percentage ; but the last column of the table shows that on plot 10a, with ammonium - salts without mineral manure, with the high percentage of nitrogen, and the high proportion of chlorophyll in the green produce, Om-hon there was eventually a very much less assimilation of carbon. S!™^"^ crops. 176 THE EOTHAMSTED EXPEEIMENTS. The result is exactly similar in the case of the barley ; plot la being manured with ammonium-salts alone, and plot 4a with the same ammonium - salts and mineral manure in addition. It is evident that the chlorophyll formation has a close connection with the amount of nitrogen assimilated ; but that the carbon assimilation is not in proportion to the chlorophyll formed if there is not a sufficiency of the necessary mineral constituents available. No doubt there had been as much or more of both nitrogen assimilated, and chlorophyll formed, over a given area, where the mineral as well as the nitro- genous manure had been applied ; the lower proportion of both in the dry matter being due to the greater assimilation of carbon, and consequent greater formation of non-nitro- genous substance. JSffect of The next point to consider is. What is the effect of the umream- unrccovcred amount of nitrogen on succeeding crops ? This gen on sue- is illustrated by the results in the coloured columns of Table 47 (p. 168). In the table, mineral manure alone is indicated by blue, nitrogenous manure alone by yellow, and a mixture of the two by green. Plot 5 has been manured continu- ously for 43 years with mineral manure alone ; whilst plots 17 and 18 each received, alternately, mineral manure, or a quantity of ammonium-salts containing 86 lb. of nitrogen. Thus we are able, for every year, to compare a plot manured with minerals succeeding a previous application of ammo- nium-salts, with a plot receiving mineral manure alone every Increase year. It is seen that, in every case, the application of nitro- '^aenous^^^°' g^'^'^Tis manure gave a greatly increased yield, frequently mamwre. doubling that of the plot with mineral manure alone. Again, in every case, the yield of the succeeding year, when the mineral manure followed the previous application of ammo- nium-salts, was reduced approximately to that of the plot continuously treated with minerals alone. A glance down the columns of plots 17 and 18, each coloured alternately blue and yellow, and a comparison of them with the blue column of plot 5, will bring the results strikingly to view. A comparison of the averages of the periods of 8, and of 40 years, of this treatment, clearly shows the essential identity of the results of the continuous and the alternate treatment with mineral manures. The averages for the 40 years show an increase in the yield of the mineral manure after ammonia, over the yield of plot 5 with mineral manure alone every year, of only ^ of a bushel per acre per annum, in a crop of between 15 and 16 bushels. The non-effect, or the absence, of residual available nitrogen applied in the WHEAT. 177 form of ammonium-salts is evident. In other words, nitrogen Ammon- applied as ammonium-salts in. any one year was practically ™'^""'f"j exhausted that year, in the crop, or otherwise; leaving prac- moneyear. tically none for subsequent action. Lastly, in regard to plots 17 and 18, it is seen that the average produce over 40 years of the ammonium-salts succeeding the mineral manure is 30|- bushels, or exactly twice as much as that of the mineral manure succeediug the ammonium-salt. Again, plot 16 received annually for 13 years, 1852-64 ridd/rcm inclusive, mixed mineral manure and ammonium-salts con- ^'^yynan- taining a double quantity (172 lb.) of nitrogen ; then for 19 years, 1865-83, it was left unmanured ; and then, for the crop of 1884 and each year since, it has received mixed mineral manure and sodium-nitrate containing 86 lb. of nitrogen. During the 13 years of heavy manuring there was a large yield, in two cases exceeding 50 bushels, with an average for the 13 years of 39J bushels. The first 3 of the succeeding years during which no manure Result of was applied, the average yield was only 21| bushels, a de- v^*'^^- crease of nearly one-half, followed iu the succeeding two ure. periods of 8 years each by average yields of 16f and llf bushels ; against, for the corresponding periods on plot 3, con- tinuously unmanured, 12^ and 10^ bushels. Or, taking the average of the 19 years of yield without manure on plot 16, we have 14g bushels, against, over the same years, 13^ bushels on plot 5 with mineral manure only since 1852, and llf bushels on plot 3, unmanured since 1839. It is fair to presume, moreover, that some of the greater yields of plot 16 over that of plot 3 from 1865-83, were due to the residue of the mixed mineral and excessive nitrogenous manure, but perhaps mainly, as will be seen further on, to increased crop- residue. Since the re-commencement of the manuring to plot 16 for ManvHng the crop of 1884, however, the plot has given some heavy '■«^»«^- yields, notably in 1886 and 1891 ; and the average for the 8 years, 1884-91, was 37^ bushels, or only If bushel less than on plot 2, which has received 14 tons of farmyard ' manure per acre each year for the last 51 years. If, as the above results have demonstrated, there is practi- What be- caUy little or no available residue from previous application J^^ of ammonium-salts, the question arises. What becomes of nitrogen? the nitrogen of the manure not taken up by the immediate crop ? This point is illustrated by the results given in Table 50 (p. 178). The plots there tabulated all received the same amount of nitrogen in manure, but with different mineral manures, and they are given in the order of their average annual increased yield of nitrogen in the crops over plot 5, VOL. VII. M 178 THE EOTHAMSTED EXPERIMENTS. with mineral manure alone. The first column shows the estimated average annual increased yield of nitrogen per acre in the crops ; the second the estimated annual loss of nitrogen as nitric acid by drainage ; the third the estimated annual excess of nitrogen in the surface-soil over that on plot 5 with the mineral manure alone ; and the last column shows the relation which the excess in the soil bears to 100 increased yield of nitrogen in the crops. The plots were manured as follows : — - Plot lb. 10. Ammonium-salts = 86 nitrogen. 11. „ .. =^86 12. M M =86 13. ,1 .. =86 14. II ri =86 7. I. II =86 9. Sodium nitrate =86 and superptoaphate. superphosphate and soda. ri and potash. II and magnesia. II soda, potash, magnesia. If soda, potash, magnesia. and and TABLE 50. — Experiments on Wheat. Estimated Nitrogen per acre per annum, 30 years, 1851-52 to 1880-81. Plots. In crops over plot 5. Lost by drainage over plot 5. In surface-soil 9 inches deep over plot 5. Excess in surface- soil to 100 increase in crop. lb. lb. lb. lb. 10 12.4 31.2 4.8 38.7 11 17.7 28.5 11.6 65.5 12 22.2 24.5 14.6 65.8 13 23.4 25.6 17.8 76.1 14 24.1 27.5 15.5 64.3 7 25.9 19.0 19.3 74.5 9 26.5 23.7 18.5 71.2 Nitrogen in the crop. Loss of nitTogen im It is seen that the increased yield of nitrogen in the crops varied exceedingly with the same amount supplied in manure, according to the supply of mineral constituents. Plot 10, with the ammonium-salts alone, gives the smallest increased yield of nitrogen in the crop ; and plots 7 and 9, with the most complete mineral manure, each gives more than twice as much ; the other plots giving intermediate amounts. The order of the estimated loss of nitrogen by drainage is almost the converse of that of the increased yield in the crops. Plot 10, which gives the least increased yield in the crop, shows the greatest loss by drainage ; and plots 7 and 9, which yield the greatest increase in the crops, show the least loss by drainage. WHEAT. 1 79 The excess in the soils (over plot 5) is obviously much xurogm more in the order of the increased yield in the crops. Plot '"- '** *'"^- 10, with the least in the increase of crop, and the most in the drainage, shows the least excess in the soU; whilst plots 7 and 9, with the greatest increased yield in the crop, and the least loss by drainage, show the greatest excess in the son. It is clear, therefore, that whilst the excess in the soil has no direct relation to the amount supplied in the manure, it has a very obvious relation to the increased yield in the crop — in other words, to the amount of growth. The last column of the table brings this out more clearly. Excepting in the case of plot 10, with the ammonium-salts alone, there is a general uniformity in the proportion of the excess in the son over plot 5 to the increased yield in the crop over plot 5 ; and the variations, such as they are, have an obvious con- nection with the conditions of growth. Thus, plots 11, 12, and 14, all with a deficient supply of potash, show approxi- mately equal proportions retained in the soil for 100 of in- crease in the crop. Plots 13, 7, and 9, again, all with liberal supplies of potash, show higher but approximately equal proportions retained in the surface-soil for 100 of increased yield in the crop. From the various results which have been adduced, it is XUrogen obvious that the relative excess of nitrogen in the soils of ^^^' the different plots is little if at all due to the direct reten- tion of the nitrogen of the manure ; and that it is almost ex- clusively dependent on the difference in the amounts of the crop-residues (of the stubble and roots, and perhaps of weeds), of which there will be the more the greater the amount of crop grown. It may be here observed that the detailed estimates, of which the results given in Table 50 are a summary, do not account for the whole of the nitrogen applied to the experi- mental plots ; and it is believed that most, if not the whole, of the unaccounted for amounts are due to loss by drainage Loss of beyond that estimated from the pipe drainage. However, in 2rainwe'^ the use of ammonium-salts or nitrate of soda, in smaller quantities per acre than those used in the experiments, and in the course of a rotation of various crops, with varying character and range of roots, as in ordinary agriculture, there will be less loss of nitrogen by drainage than that indicated in these experiments. In the Eothamsted soil and subsoil, with chalk below affording good natural drainage, or in soils generally with good drainage, natural or artificial, it is not probable that there is any material loss by evolution as free ^ff^^ nitrogen. Where, however, nitrogen is applied in large nitrogen. 180 THE EOTHAMSTED EXPERIMENTS. I quantities as farmyard or other organic manure, there may be considerable loss by evolution as free nitrogen. Effect of The next point to consider is the differences in the ^ti'^diffr ai^o'i'it of crop with equal nitrogen, but different mineral CT Arable laid down to grass (High- ) field), spring (?), 1838 f Very old grass-land (The Park) 1867, after wheat 1874, after clover 1883, after wheat October 1866 . II 1881 . II 1866 . II 1881 . II 1865 . II 1881 . II 1866 . II 1881 March 1868 II 1882 April 1870 February 1882 . November 1881 January 1879 . September 1878 Feb. and March 1876 per cent. per cent. 0.1402 0.1372 * 0.1891 0.1882 1.836 9.8 0.1967 2.294 11.7 0.1280 1.180 9.6 0.1264 1.341 10.6 0.1119 1.039 9.3 0.1012 1.080 10.7 0.1090 0.978 9.0 0.1045 1.017 9.7 0.1202 0.1124 1.154 10.3 0.0984 0.1235 0.1509 0.1740 0.2067 2.412 11.7 0.1943 2.403 12.4 0.2466 3.377 13.7 ' Rothanisted. VARIOUS AEABLB SOILS IN GREAT BEITAIN. Mr Front's Farm Wheat soils Red Sandstone soil , rBroadfield — surface < Blackacre— surface <- Whitemoor — surface r Mid-Lothian J East Lothian •i Perthshire I Berwickshire \ England . 0.170 0.107 0.171 0.22 0.13 0.21 0.14 0.18 ^Voelolcer. K Anderson. Voelcker. UNITED STATES AND CANADIAN PEAIRIB SOILS. United States —Illinois Canada \ la\ No. 1 No. 2 . No. 3 No. 4 'Manitoba ; Portage Prairie — surface N.W. Territory ; katchewan district— J- surface J N.W. Territory; 40 miles-, from FOrt EUioe— sur- \ feoe J / Niverville— flrstl2 \ Mani-L,,„^„^ ''"'^^'•^ fnha. \ Brandon ii ^°°^ Selkirk V Winnipeg n 0.80 0.26 0.33 0.34 •• 0.247 0.303 0.250 0.261 3.42 13.1 0.187 0.618 0.428 2.66 7.58 5.21 14.2 12.3 12.2 hVoelcker. >Bothamsted. I- Rothamsted. RUSSIAN SOILS. No. 1—12 inches ' 0.607 \ No. 2— 8 II 0.467 No. 3— 6 11 0.188 No. 4— 6 ■ II 0.130 > 0. Schmidt. No. 6—11 .1 0.306 i No. 6— 17 1, 0.281 No. 7— 9 II . . 0.409 •• ; 1 Calculated on soil dried at 100° C. WHEAT. 191 0.140. Then, referring to the results obtained in the wheat- field itself, it is seen that after growing wheat with full mineral and nitrogenous manure since 1843-44, the percen- tage of nitrogen in the dry sifted surface-soil was — in 1865, 0.1230, and in 1881, 0.1264; but with mineral manure without nitrogen, it was — in 1865, only 0.1119, and in 1881, 0.1012 per cent ; and lastly, without manure from the com- mencement it was — in 1865, only 0.1090, and in 1881, 0.1045 per cent. That is to say, with mineral and nitrogenous manure, the percentage of nitrogen was the highest, and rather higher at the later than at the earlier date; the result being due, as has been proved, not to the accumula- tion of manure-residue, but of crop-residue. On the other hand, with mineral manure without nitrogen, or without any manure at aU, the percentage of nitrogen was lower than when nitrogenous manure was also used, and in each case it was lower at the later date — that is, as the exhaustion progressed. On a consideration of these various results, it may perhaps fairly be concluded that the percentage of nitrogen in the surface-soil of the wheat-field at the commencement was certainly higher than in 1865 or 1881, under the conditions of nitrogen-exhaustion with mineral manure alone, or with- out any manure at all; and that it was about as high as where nitrogenous as well as mineral manure had been annually applied ; probably, therefore, from 0.1250 to 0.1300 per cent, and probably nearer the lower than the higher figure. Looking to* the other results in the table relating to Eothamsted soils, it is seen that with barley, as with wheat, when grown year after year with mineral manures alone, the percentage of nitrogen in the surface-soil was low, with a tendency to decline from time to time ; and lastly, after roots grown with mineral manure alone, the percentage is lower still — ^indeed lower than has been found where any other crop has been grown under similar conditions. Then it is further seen, that in the ease of various arable fields laid down to permanent grass, the percentage of nitrogen increased more or less according to the time it had been laid down — the figures at the different periods being 0.1235, 0.1509, 0.1740, 0.2057, and 0.1943 ; whilst the percentage in very old grass- land was 0.2466. Next, in various arable soils in Great Britain, the percen- tage of nitrogen in the surface-soils ranged from 0.107 to 0.220. Compared with these, the percentage in various United States and Canadian prairie soils ranged from 0.187 to 0.618 ; the greater number showing about 0.30 per cent. Lastly, a num- 192 THE EOTHAMSTED EXPERIMENTS. 6ro.ss- land, rich, arable land, poor in nitro- gen amd carbon. Accumu- lated fer ber of Eussian soils ranged in percentage from 0.130 to 0.607, It is further seen that the percentages of carbon, and the amount of carbon to 1 of nitrogen, are higher in the grass- land than in the arable soils, and higher still in the rich- prairie soils. From these various results there can be no doubt that a characteristic of a permanent grass surface-soil, or of a rich virgin-soil, is a relatively high percentage of nitrogen and of carbon, and a high relation of carbon to nitrogen. On the other hand, a soil that has been long under arable culture is- much poorer in these respects ; whilst arable soils, under con- dititions of known agricultural exhaustion, show a very low percentage of nitrogen and of carbon, and a low relation of carbon to nitrogen. It has sometimes been maintained that a soil is a laboratory and not a mine. But not only the facts ascertained in our own and in other investigations, but the history of agriculture throughout the world, so far as it is known, clearly show that a fertile soil is one which has accumulated within it the residue of long periods of previous vegetation ; and that it becomes infertile as this residue is exhausted. Such ac- cumulations are truly enormous in many of the prairie lands of the American continent ; sometimes, indeed, extending to a considerable depth. But, even after the comparatively few years which most of them have been under cultivation, it is alleged by some that they are already showing exhaustion. In view of the facts both as to the percentage of nitrogen, wJ^atfrmi ^^'^ *^® annual yield of wheat without manure over 40 or 50 ^airie years in the Eothamsted experimental field, ij is indeed very difficult to believe that the rich prairie lands of the American continent, which yield so large a proportion of the wheat ex- ported from the United States and Canada, can in so much less a time have become exhausted of available nitrogen. Thus it is probable that at the commencement the surface- soil of none of these lands contained less than twice, and few of them less than three times, as high a percentage of nitrogen as the Eothamsted wheat-field soil; whilst frequently the subsoils would, to a considerable depth, be richer than the Eothamsted surface-soil. Yet it is estimated that over a period of 40 years, from 1852 to 1891 inclusive, the produce of the Eothamsted soil without manure has only reduced by an average of about ^ bushel per acre per annum due to exhaus- tion, irrespectively of fluctuations due to season ; and when we consider how much shorter a time most of the rich prairie lands have been growing wheat without manure, it seems that some other reason than exhaustion must be found for their alleged reduction in yield. Reduction land. WHEAT. 193 As to the number of years during which the greater por- tion of the rich prairie lands of America have been broken up for the growth of wheat, it may be observed that a series of unproductive seasons, not only in our own country but in Western Europe generally, which culminated in 1879, but continued for some years later, led to a more rapid reduction in our own area under the crop, and concurrently to the opening up of large wheat-growing areas in various parts of the world, and at the same time to greatly increased imports ; a much larger amount coming from the United States than from any other country, indeed generally more than from all other countries put together. Thus, the area under wheat in the United States increased from imder 21 million acres in 1872, to more than 27^ million in 1876, with an average for the 5 years of nearly 24J million. Over the next 5 years the area increased from 26J million in 1877 to 37f million in 1881, with an average over the 5 years of 33J million. Over the next 10 years, from 1882 to 1891, the area averaged 37J million acres ; and it was 39.9 million in 1891, and more than 38 J million in 1892.^ There was an increase, therefore, from less than 21 million in 1872, to an average of 37J million over the 10 years ending 1891, or by about 79 per cent. In fact, this great increase in the area under the crop took place within a period of about 20 years; the actual increase during that period amounting to about 16^ million acres, by far the greater proportion of which was rich prairie land. Of this the larger proportion was brought under cultivation within a period of about 15 years. Bearing in mind the results obtained at Eothamsted without manure for 50 years, on a comparatively very poor soil, it does in- deed seem incredible that a period of about 15 years should be sufficient to bring about palpable exhaustion of the in- comparably richer prairie soils. Within the same period of 20 years, the home consump- United tion of wheat in the United Statesj according to the rec- "^^^ ords, increased from rather under 200 million Winchester ductim bushels in 1872-73, to an average of nearly 334 million over '^'^ export. the 10 years from 1882-83 to 1891-92 ; whUst the exports have increased from 52^ million bushels in 1872-73 to an average of 146J million over the 5 years 1877-78 to 1881-82 ; but they amounted to an average of rather less than 130 mil- lion over the 10 years 1882-83 to 1891-92. The maximum amount in any one year was, however, 227J million in 1891-92. It has been estimated that, judging from the fncrease of ^ Subsequent records show that the area was reduced to 34. 6 million acres in 1893, and to 34.8 in 1894. VOL. TIL N 194 THE KOTHAMSTED EXPERIMENTS. the population of the United States in the past, the Central, Northern, and Western States, from which we now derive such large supplies of grain, will, before many years have passed, be as densely populated as the Eastern States are now; and that then the export of grain will be rapidly diminished. In this calculation, however, the essential dif- ference in the character of the land in the Eastern States, and in the prairie districts of the Central, Northern, and Western States, is not taken into account. It is true that both western meat and western wheat are materially reducing the production of them in the Eastern States ; so that the popula- tion of the east as well as of the west will consume more and more of the western produce, leaving of course the less for export. And if, in addition to, this, it be true, as alleged, that the western lands themselves are losing their fertility, there would indeed seem that there is some likelihood of material reduction in exports before very long. Certain it is, however, that large areas of formerly prairie land, which provide so much of the exports, were originally as rich as ploughed-up old grass-land in our own country, and sometimes so to a considerable depth. Hitherto the land has, as a rule, only been skimmed, practically no labour bestowed on cleaning, and compared with the pro- duce which such lands should yield if properly cultivated, very small crops of grain have been obtained. But the large crops occasionally yielded under favourable conditions are evidence of the inherent fertility, and of the possible pro- ductiveness, of the soil. Further, from what has been said, it is almost impossible to believe that such soils can have become seriously exhausted within comparatively so few years, at any rate so far as available nitrogen is concerned. Indeed, if there be palpable exhaustion at all, it would seem more likely that it is of some mineral constituents — potash, lime, or phosphoric acid, for example. However this may be, so long as wheat is grown under the conditions frequent, and indeed almost inevitable, in the case of new settlement, with sparse population — that is, growing it for several years in succession, with deficient cultivation, luxuriance of weeds, the burning of the straw, and generally the wasting of the manixre of the working stock — only low yields can be ex- pected. The practice naturally results from the fact that, under such conditions, fertility is cheap and labour dear. As population becomes more dense, however, local markets will arise for rotation products, more stock will be kept, the straw and the manure will be utilised, cultivation will be improved, and weeds will lose their ascendency. Nor can there be much doubt that under such conditions it will be EOTATION OF CROPS. 195 found that the growth of comparatively small crops of wheat, even with a fair share of weeds, for 15 or 20 years on rich prairie land has not exhausted its fertility. There will besides, for some time to come, be more rich prairie land to bring under the plough. Upon the whole, it seems prob- able that, with the improved methods which should result from increased density of population, and with the increased areas brought under cultivation, it will be longer than is sometimes supposed before the capability of the United States of production for export will be materially dimin- ished. Obviously, somewhat similar arguments are, mutatis mviaiidis, applicable to Canada. As, however, the resources of the rest of the world, taken as a whole, show no signs of diminution, it may be a question how far the range of prices will affect the production in any particular country. SECTION v.— ROTATION OF CROPS. Introduction and Historical Sketch. In the preceding sections attention has been devoted to the consideration of the influence of exhaustion, manures, and variations of season, on the amounts of produce, and on the composition, of certain individual and typical crops when each is grown separately year after year on the same land. In this way there have been discussed the characteristic re- quirements and results of growth of various cereal crops as representatives of the natural order Graminese ; of various root-crops of the orders Cruciferse and Chenopodiacese ; and lastly, of various Leguminous crops. Our subject now is the — Rotation of Crops. The mere numerical results of the field experiments made at Eotham- sted on rotation have been recorded in the annual ' Memor- anda ' ; but the first systematic discussion, either of them or of the laboratory investigations undertaken in connection with them, is that given in this paper, in this volume, and in the Journal of the Royal Agricultural Society of England (December 31, 1894); and although the present communi- cation embodies a good deal of detail, and a somewhat com- prehensive consideration of it, there still remains much which could not be included within the limits of this paper. The practice of Eotation is admitted to be the foundation importance of the improvements in our own agriculture which have taken <>f 'rotation. place during this and a considerable part of the last century. It is of great importance, therefore, carefully to consider, both 196 THE EOTHAMSTED EXPERIMENTS. Rotation crops. Persistent com-grow- Legvmirk- mis crops in early rotations. Introduc- tion of twr- nip-cul- twe. in what the practice itself consists, and how its benefits are to be explained. If the rotation of crops as followed in our own country, indeed over large portions of Europe, were to be defined in the fewest possible words, it might be said that it consists in the alternation of root-crops, and of leguminous crops, with cereals. In the United States, however, it is a gramineous crop — maize — which largely takes the place of root-crops in Europe. The cereals constituting such a very important element of human food, it was natural that they should be grown almost continuously so long as the land would yield remunerative crops. Hence, the history of agriculture, not only in our own country, ijut in others where these crops were of high relative value, shows that it very generally came to be the custom to grow them for a number of years in succession, and then to have recourse to bare fallow ; or, in some cases, to abandon the land to the growth of rough and weedy herbage, affording scanty food for domestic animals. The improvement upon these practices, attainable by alter- nating other crops with the cereals, was very much earlier re- cognised in the case of the leguminous than of the root-crops, the introduction of which is of comparatively recent date. It was, in fact, distinctly recognised by the Eomans more than two thousand years ago, not only that certain legu- minous crops were valuable as food for animals, but that their growth enriched the soil for succeeding crops — in fact, that they were of value as restorative crops grown in alternation with the cereals. There is, however, very scanty indication that root-crops were an element in their alternate cropping. As in the agriculture of the ancients, so in that of more modern times, especially in our own country, various legami- nous crops were grown in alternation with cereals long before roots were so interpolated. It was, indeed, not until about, or after, 1730 that Lord Townshend, who, as Secretary to George I., had been in Hano- ver, and there seen turnips growing as a field crop, on his return introduced them on his own estate in Norfolk, and there founded the celebrated Norfolk four-course rotation of turnips, barley, clover, and wheat. His own laud was previ- ously to a great extent a marshy or sandy waste, and its value was increased enormously under the new system. It was, however, not until towards the end of the century that it became generally adopted even throughout his own county. In this extension Mr Coke, of Holkham (afterwards Earl of Leicester), was largely instrumental, and the practice seems to have next extended into Lincolnshire. EOTATION OF CROPS. 197 It was thus that The Four-course Hotation, or, in other Four- words, the alternation of root-crops and of leguminous crops ;^**™*^ with cereals, hecame established. Such alternation is, in fact, the basis of aU the various rotations which are adopted in different pai-ts of our own country, and also to a great extent which are followed in many other countries. It is worthy of remark that, although we owe the introduc- Yield of tion of the essential elements of our rotations to the example ^^*^^ of our Continental neighbours, we, with one or two im- and/oreign material exceptions, obtain more per acre of all the staple «™«<™«- saleable products of rotation, grain and meat, under our landlord, tenant, and labourer system, than any other country in Europe, or than in America, under whatever advantages of climate, or under whatever system of holding, or of size of holdings. Thus, there is not a single country in Europe that reaches our average produce per acre of wheat ; only Belgium and Holland approach, but they do not equal, us in the pro- duce of barley ; only Belgium, Holland, and Norway exceed us in acreage yield of oats ; and no country approaches us in acreage produce of potatoes. Again, whilst several countries exceed us in number of cows to a given area, and some in the number of pigs, not one equals us in weight per acre of other cattle than cows ; and not one nearly approaches us in the weight of sheep to a given area. Nor, notwithstanding the great depression of our agriculture in recent years, the result of the low prices of produce, is there any probability that we shall soon lose our pre-eminence in production per acre. There can be no doubt that the effect of the extension of Benefidai the growth of green crops was — to a great extent to get rid of *y^^^ unprofitable fallows, greatly to increase the supply of stock crops. food, especially for winter feeding ; so to lead to a largely increased production of meat and milk, to a greatly increased supply of manure, and thus to enrich the land for the growth of grain, which, accordingly, yielded much larger crops. We have now to endeavour to ascertain how the admittedly BeaejUs of very beneficial effects of alternate, as distinguished from con- ^"^"^^^^ tinuous, cropping are to be explained. It will be well first very briefly to refer to some of the chief theoretical explana- tions that have been put forward, and afterwards to discuss the results of various direct experimental investigations con- ducted at Eothamsted on the subject of rotation. The first definite theory as to the benefits of the alternation Them-etuxU of crops assumed that the excreted matters of one description ^^""" of crop were injurious to plants of the same description, but that they were not so, and might even be beneficial, to other kinds of plants. 198 THE ROTHAMSTED EXPERIMENTS. lAebi^s Boussin- tions. Professor Daubeny's researches. Theory of excretions disproved. Rotation amd or- ganic and constitu- ents. At first Liebig pronounced this theory of rotation to be the only one having any really scientific basis. Later he seems to have modified his view considerably, and to have supposed that the explanation was — not that the excreted matters of one description of plant were injurious to another of the same description, but that, as the different plants had such very different mineral requirements, the alternation of one kind with another relieved the soil from exhaustion. In his latest work, however, after many years of controversy, he obviously more fully recognised that nitrogen probably played some important part in the matter. More than fifty years ago Boussingault published the results of an investigation, extending over a period of ten years, to determine the chemical statistics of some of the rotations actually followed in his own locality, in Alsace; and he came to the conclusion that the difference in the amounts of nitrogen taken up by the different crops constituted a very important element in the explanation of the benefits of rotation. We can only further briefly refer to the results and con- clusions of the late Professor Daubeny, of Oxford, who com- menced a series of experiments in the Botanic Garden there in 1834. One of the original objects he ha,d in view was to test the truth of De CandoUe's theory that the excretions of one description of plant were injurious to plants of the same description. He soon came to a negative conclusion on the subject; and recognised the validity of Boussingault's argu- ment, that the actual facts of vegetation in different parts of the world conclusively showed that the same description of plant may continue to grow healthily on the same land for long periods of time. On this point at is scarcely necessary to add that the experience at Eothamsted on the growth of various agricultural crops year after year on the same land for many years in succession is conclusive against the theory of injurious or poisonous excretions. But, as already said, Dr Daubeny continued his experi- ments for ten years; and although, in accordance with the prevailing ideas of the time, all his analytical results related to the mineral constituents of his soils and crops, his main conclusion was, that the benefits of rotation were probably as much connected with the available supply of the organic as of the inorganic constituents. What, then, are the indications of the results of many years of investigation of the subject, in the field and in the labora- tory, at Eothamsted ? eotation of ckops. 199 The Experiments on Eotation made at Eothamsted. The experiments have been conducted in Agdell Field. An area of 2J acres is devoted to the purpose. The ordinary four-course rotation of — turnips, barley, clover (or beans), or fallow, and wheat, was adopted. The experiments were com- menced in 1848, so that the eleventh course of four years each was completed with the harvest of 1891 ; and the wheat which has jast been sown (October 1894) is the fourth crop of the twelfth course, and wiU complete the forty-eighth year of the experiments. The area of 2J acres was divided into three main divisions, which have, respectively, been under the following conditions as to manuring : — 1. Without manure from the commencement. Manures 2. For the first nine courses, manured with superphosphate "^^y^ 1 Tji,.!,- ^^. ^, rotation ex- alone, applied only tor the turnip crop commencmg each penmmts. course ; that is, once every four years. For the tenth, and each subseq^uent course, salts of potash, soda, and magnesia, have been applied as well as superphosphate. 3. A complex artificial manure, also applied every fourth year ; that is, for the turnips commencing each course. This manure comprises — superphosphate, salts of potash, soda, and magnesia, ammonium-salts, and rape-cake ; and it supplies about 140 lb. of nitrogen per acre for the four years' course ; that is, an average of 35 lb. of nitrogen per acre per annum. The complex manure (3) was designed to be, in great measure, a substitute for farmyard manure ; and it was used instead of it, in order that the amount of the diflferent con- stituents supplied might be more accurately known than would have been the case if farmyard manure had been employed. It should be further explained, that when the land is under Removal turnips, the roots, with their leaves, are removed from one '^^'^ half of each of the three differently manured plots ; whilst, of roots. on the other half of each, the produce is consumed on the land by sheep ; or, if the weather be unsuitable for this, the roots are sliced, and both roots and leaves are spread on the land. Thus, each of the three main divisions is divided into two, making, so far, six in all. Then again, after the first course of four years, in the third year of each course the leguminous crop was grown on only half of each of the three differently manured plots, and the other half was left fallow. Lastly, as clover cannot be relied upon on such land so often as once in four years, beans have frequently been grown instead. We have finally, therefore, twelve plots instead of only 200 THE KOTHAMSTED EXPEEIMENTS. ment of plots. Method of ascertain- ing results. of tables. three. , That is to say, each of the three differently manured plots is divided into four as above described, and as indicated in the heading of the several tables ; and, as the same form of table will, as far as possible, Be adopted throughout, it is very desirable that a clear idea of the arrangement should be formed at the outset. It will be seen that under each of the three main divisions designated in the heading according to the manuring, the results are subdivided, showing first the produce obtained where the roots were carted from the land ; and secondly, where they were fed (or left) upon it. Lastly, under each of these two conditions so far as the disposal of the turnips is concerned, there is again a subdivision into two — one where in the third year of the course the land was left fallow, and the other where either clover or beans was grown. Each year the amount of produce on each of the various plots is weighed ; samples of each crop are taken ; in all the dry substance and the mineral matter (ash), and in many the nitrogen, are determined ; in many cases also complete analy- ses of the ashes of the crops have been made. Lastly, deter- minations of the total nitrogen have been made in the surface soils, and in the upper layers of the subsoils, at different periods ; and the nitrogen as nitric acid has also been deter- mined to a considerable depth. As to the results themselves, only brief reference to the main indications of these various investigations can be made. Tables 56, 57, 58, and 59, give the amounts of produce of the turnips, the barley, the leguminous crops, and the wheat, respectively, in each of the eleven years in which each was grown, in the eleven completed courses. Each table is divided into three main divisions — the upper one giving the roots, or the grain, &c., as the case may be ; the middle the leaves, or the straw ; and the lower one the total produce — roots and leaves, or grain and straw, together. Taile 56 The Swedish Turnip Crops. Preferring to Table 56, relating to the Swedish turnips, it is seen that in the first year, 1848, there was, both without manure and with superphosphate alone, much more produce than in any subsequent year ; showing that, at the commence- ment, the land was in somewhat high condition, due to pre- vious treatment. Then, again, as already said, for the tenth and eleventh courses, salts of potash, soda, and magnesia were used as well as superphosphate. For these reasons, the results of the first and of the tenth and eleventh courses are ex- cluded from the averages to which attention will chiefly be EOTATION OF CROPS. 201 TABLE 56. — Experiments on the Eotation op— Eoots, Barley, Clover (ob Beans), OR Fallow, aIjd Wheat ; in Agdell Field, Rothamsted. 11 courses, 44 years, 1848-1891. 1.— ROOTS— SWEDISH TURNIPS. UnmaDiirecl. Courses 1-9 superphosphate only. Courses 10 and 11 mixed mineral manure. Mixed mineral and nitrogenous manure. Tears. Boots carted. Boots fed. Roots carted. Roots fed. Roots carted. Roots fed. Fallow. Beans or clover. FaUow. Beans or clover. Fallow. BeaDsor clover. FaUow. Beans or clover. Fallow. Beans or clover. Fallow. Beans or clover. ROOTS. 1848 tons. cwt. 8 15i tons 3 cwt. 6} tons. cwt. S 173 tons, cwt, 6 9 tons. cwt. 14 12 tons. cwt. 11 53 tons. cwt. 17 5 tons. cwt. 11 03 tons. cwt. 19 143 tons. cwt. 10 IS tons. cwt. 21 9 tons. cwt. 11 9 1852 1856 1860 1864 1868 1872 1876 1880 1 17 2 5} 13 7i Crop 2 llj 1 11} 1 128 1 6 1 12 1 83 faUed 1 14i 17i 14 1 7i 1 14 li 9 2"9i 1 12i 1 18i 19} 1 Oi 1 8} r'9} 1 1 1 1 12 163 8 10} 1 133 2 123 7"2i 9 ISi 11 4 11 3i 6 16 1 9i 3 8 8' 101 9 8i 9 19} 13 13i 9 133 2 03 3 19} 8" 7} 10 8i 11 183 12 lOi 9 16 1 183 3 183 9 "lO} 11 63 11 33 20 8i 16 8i 4 7} 9 2} 16 "l2 15 93 22 lOi 19 16} 16 13| 4 7} 8 16i 16'l9i 17 16 21 19} 19 103 16 19i 4 7 9 5} 16"lli 18 17i 22 7i 19 6 17 li 3 12 8 8} le'io 17 19i 22 6i 1884 1888 17J 15 5 2i 1 Oi 1 3 12 8 7 193 7 ii 8 13i 10 7} 8 12} 8 6 10 6 12 9i 14 18i 21 Hi 14 63 23 12J 14 16i 21 33 14 03 20 178 iverage S") 1S62 to f 1 6 16g 1 4 15} 6 14} 6 63 7 lOi 7 10} 13 21 13 6} 13 93 13 2J 1880 ; tverage S\ courses, f 1884 and f 16i 31 1 li 10 7 Hi 9 lOi 8 9i 11 7g 18 43 IS 193 18 17 n 1888 ) 1848 193 2 53 1 0} 3 73 1 15 6 6i 1 193 4 10 2 6i 7 113 2 63 1 li 1862 1856 1860 1864 1868 1872 1876 1880 63 2i 0} 03 Crop 8} 6i Si 4i 2} (6i lb.) 03 failed 83 6 2i 4 2 Oi 03 0"7i 6} 4 3i 1} (5 lb.) 1 o"7| 5 3 1 23 8 2 4} 0'l4| 17 12} 1 Oi 7} 1} 43 "iTi 113 1 2i 12i 2 5i 0'l7 16 12 1 2 14} 13 43 o"l9} 1 7} 11 2 11} 3} 9 l"l4i 1 143 1 16 1 16} 12} 3i 8| l'l5J 2 16i 2 3i 1 173 12} 53 9} 1 13} 2 03 1 18 1 13 113 4} 8 1 19 3 3 1 183 1884 1888 7f 7} 3} li 7 71 5 3i 18} 16} 1 Oi 1 H 18 16 1 3 1 3 2 16i 1 17i 3 3i 2 63 3 6i 1 15 3 3} 2 0} verage 8\ ionrses, ( 1852 to f 1880 j verage 2 J 884 and f 1888 ) 3i n 3 2} 2i 7i 2| 4} 10} 16J 11} 1 Oi 11 171 121 1 3 1 li 2 6} 1 4i 2 14} 1 2i 2 lOi 1 H 2 12} TOTAL PRODUCE. 1848 9 15 5 Hi 9 18i 8 163 16 7 16 12 19 43 15 lOJ 22 1 18 9J 23 163 19 Oi 1862 1866 1860 1864 1868 1872 1876 1880 2 2| 2 73 li 8i Crop 8 1 163 1 16} 1 9} failed 2 2i 1 2i 16i 1 Hi 1 16 93 2 '16} 1 17i 2 2i 1 9} ri7i 1 6 1 4 13 19} 8 18} 1 163 2 17i 7"l6i 10 lOi 11 16} 12 3} 7 3} 1 10 3 12 O' 8 10 163 10 Hi 14 15} 10 6 I I 9" 4} 11 43 12 111 13 12i 10 10} \ % 10 "lO 12 13i 11 143 22 8i 16 193 4 11 9 11} 18"6i 17 4} 24 6i 21 IS 17 6i 4 103 9 6 is' 168 20 Hi 24 23 21 8} 17 113 4 123 9 15 is" 43 20 18 24 5i 20 19 17 13 3 16} 8 17i is" 9 21 2i 24 6 1884 1888 1 5i 1 2} 8} 4} 1 7i 1 108 17 Hi 8 iSi 7 18} 9 13} 11 88 9 Hi 9 2 11 9 13 12i 17 18} 23 9} 17 10 25 18} 18 2} 22 183 17 4} 22 18} eiage S) mrse.s, ( 862 to f 1880 erage 2 )urses, f 84 and f 1888 ) 1 9i 1 33 19i 6} 1 6i 1 8i 173 14} 7 4i 8 8} 6 188 10 11} 8 li 9 68 8 Si 12 lOf 14 3} 20 11} 14 10} 21 14i 14 12 20 108 14 73 20 1| 202 THE EOTHAMSTED EXPBKIMENTS. Variation with seasons. No man- ure. With sv/perphos- for tur- confined. In this table, however, as well as in those relating, respectively, to the barley and the wheat, averages are given at the foot of each division of the tables, not only for the eight intermediate courses — second to ninth, but also for the two succeeding courses — tenth and eleventh, for which pot- ash, soda, and magnesia were used as well as superphosphate. But, for the leguminous crops, the averages are, for reasons that will be explained, taken differently. The first point to notice in the results is ■ that, under each condition as to manuring, there is very great variation in the amount of produce from year to year according to the seasons. Thus, in 1868, the crop entirely failed on all the plots, although seed was sown twice. Again, whilst the complex manure con- taining nitrogen yielded more than 22 tons of roots in 1880, the same manure gave little more than 4 tons in 1860 ; the average yield over the eight courses being about 13^ tons. Against this, the average by superphosphate alone ranged from about 6| to about 7|- tons ; whilst without manure there was an average of only about 1 ton. Eeferring to this last result, it is particularly to be observed that this assumed restorative crop yields practically no pro- duce at all when grown without manure. The plot with superphosphate alone gives very much more than that without manure, but still very much less than an average agricultural crop. The increase, such as it was, was largely due to the greatly increased development of feeding- root within the surface-soil under the influence of the phos- phatic manure ; and the necessary nitrogen, beyond the small amount of combined nitrogen annually coming down in rain and the minor aqueous deposits from the atmosphere, has doubtless been gathered under the influence of the increased root-development from tlie previous accumulations within the soil itself. There is, in fact, perhaps no agricultural practice by which what is termed the condition of land, that is the readily available fertility due to recent accumulations, can be so rapidly exhausted as by growing turnips on it by super- phosphate alone — provided, of course, that the seasons are favourable. Compared with the produce with superphosphate alone, the mixed manure, supplying, besides superphosphate, not only salts of potash, soda, and magnesia, but a liberal amount of nitrogen, yielded, on the average of the eight courses, nearly twice as much, or between 13 and 14 tons of roots ; though, as already pointed out, it yielded in some seasons over 20 tons per acre. There can be no doubt that, the necessary mineral . constituents being available, there was a large in- crease of produce due to the supply of nitrogen in the manure. ROTATION OF CROPS. 203 The figures in the middle division of the table show that the produce of leaf as well as that of roots was increased by superphosphate, and that it was still further increased by the mixed manure containing nitrogen. The next point is to consider the effects of the other con- Mffects of ditions besides those of different manure supply ; that is, the ^J^^^^ removal of the root-crop, or the feeding or the spreading of it land. upon the land ; also whether, in the third year of each course, a leguminous crop was grown, or the land was fallowed. It is seen that, mthout manure, whether clover or beans were grown, or the land were fallowed, there was even rather less average produce of roots over the eight years where they had been fed on the land, than where they had been carted off ; but with such very small crops the differences are im- material, if not accidental. On the swperpho^hate plots, where the produce was much higher, and where there would, therefore, be more loss to the land by removal, the crops were materially better on the fed portions of the plots. On the mixed manure plots, on the other hand, with nearly twice as much produce as with superphosphate alone, there would be stni greater difference between the condition of the land where the roots were carted off and where they were fed on ; but there was very little difference in the average pro- duce of the root-crop. It will be seen further on, that the higher condition of the land where the more highly manured roots were fed upon it had a very marked effect on the succeeding cereal crops, and especially on the immediately succeeding barley. This was the case on both the superphosphate and the mixed manure plots. The difference of effect on the average produce of the root- Effecu of crop, by fallowing, or by growing beans or clover, in t\i,Q fnUoimng third year of each course is, in the comparable cases, prac- 1^ il^ tically immaterial under each of the three different conditions "-'^ clover. as to manuring. Before passing from Table 56 it is to be observed that there was higher average produce over the tenth and eleventh courses with superphosphate and potash, soda, and magnesia, than over the preceding eight courses with superphosphate alone. But, as there was also increase in a greater degree injiuenoe with the mixed mineral and nitrogenous manure over the "/^^asore. two than over the eight years, it is obvious that the character of the seasons had a good deal to do with the result. It is noticeable, however, that on the plots with potash, soda, and magnesia, as well as superphosphate, in the two courses, there was a higher produce of roots on the plots where beans or 204 THE EOTHAMSTED EXPEBIMENTS, Legmaes clover were grown than on those that were fallowed ; a result uiah^'^f' doubtless due to the increased growth of the leguminous crop under the influence of the potash manuring, and to accumula- tion of nitrogen in the soil thereby. It may further be observed (though not shown in the table) that in 1892 — that is, the first year of the twelfth course — the produce of the manured plots was generally higher than in either of the two preceding courses. The accompanying figures represent selected typical Swedish turnip-plants, grown in 1892 — (1) without manure, (2) with 1. Crop of roots, 1892 : Si cwts. per acre. Crop of roots, IS^' 11 t tt, C} cwts. per acre. Crop of roots, 1S92 : 24 tons 18 cwts. per acre. Itbistra- tions ex- the mixed mineral manure alone, and (3) with the mixed mineral and nitrogenous manure. Each plant was fixed upon a scaled background and so photographed, and the figures as given are about one-twentieth natural size, and strictly com- parable. The quantities of produce recorded show that with- out manure it was less, but that by each of the two descrip- tions of manure it was considerably more, than the average of the preceding courses ; and both the reversion to the un- cultivated condition without manure, and the increased EOTATION OF CEOPS. 205 growth under the influence of each of the manures, are strikingly illustrated, both by the figures and by the amounts of produce given. Indeed, the results conclusively show Ahmdauce how artificial a product is the cultivated root-crop, and how of «-vaiiaJ)ie dependent it is for its successful growth on an abundant tiaifor supply of available food — nitrogenous as well as mineral — timnips. within the soil. The Barley Crops. Table 57 (p. 206) gives the produce of barley, the second crop of the course, and therefore always succeeding the roots, in each of the eleven years in which it was grown, in precisely the same form as that of the Swedish turnips recorded in Table 51 Table 56 : the upper division giving the grain per acre, the e^i'toweii. middle division the straw, and the lower one the total pro- duce, grain and straw together. As in the case of the root-crops, so in that of the barley, the produce in the first course is excluded from the calcula- tion of the averages to which reference will chiefly be made. Indeed, the results of the first year of barley confirm the conclusion that the land was in somewhat high condition due to recent accumulations. The produce of the tenth and eleventh courses is also excluded from the averages, on account of the change of manure on the superphosphate plot for the tenth and succeeding courses. Eeferring, however, first to the results of each of the eleven Variation years, it is seen that, under each condition of manuring, or ™*'' other treatment, there is very great variation in the amount of produce from year to year, due to variations in the characters of the seasons. Thus, without manure, the average produce over the eight courses was about 30 bushels per acre, whilst in 1857 it was in each case more than 40 bushels, and in some considerably more ; but in 1869 and in 1873 it was not much over 20 bushels, and in the last two courses considerably less than 20. A glance down the columns recording the produce on the manured plots will show that in their case also there was a wide range in amount above and below the averages, according to season. Eeferring now to the average produce of the eight courses (second to ninth), the first point to notice is, that whilst the assumed restorative crop — the roots — gave practically no produce at all without manure, the barley gave, on land un- jvo maw- manured for so many years, an average of rather over 30 "'■^• bushels per acre. The truth is that the cultivation for the preceding roots kept the land clean, and as there was prac- tically no produce of roots, the soil was, in point of fact, left almost fallow for the barley during the winter preceding the 206 THE EOTHAMSTBD EXPERIMENTS. TABLE 57. — ExpERiMBH-TS on the Eotation of— Eoots, Baklet, Clover (ob Beans), or Fallow, and Wheat; in Agdell Field, Eothamstbd. 11 courses, 44 years, 1848-1891. 2. barley. Unmamired. Courses 1-9 superphosphate only. Courses 10 & 11 mixed mineral manure. Mixed mineral and nitrogenous manure. Tears. Roots carted. Boots fed. Roots carted. Roots fed. Roots carted. Roots fed. Fal- low. Beans or clover. Pal- low. Beans or olorer. Fal- low. Beans or clover. Fal- low. Beans or clover. Fal- low. Beans or clover. Fal- low. Beans or clover. DRESSED GRAIN. Bush. Bush. BusTl. Bush. Bush. Bush. Bush. Bush. Bush. Bush. Bush. Bush. 1849 33S Hi m 48 29i 29J 41 42S 37 28J 44^ m 1853 32i 341 38 28J 32 28 39i 88 37J 38i. 37i 36J 1867 48S 48 44i 40i 301 28 m 62 47i 48 661 63i 1861 36^ 88 38 291 32| SOI 401 42 60J 601 67f 64| 1865 Ui 39 35i 27 Sli 33 39i 41 44g 47} 46f 481 1869 211 24f 21 26 25^ 281 30i 33 391 42| 381 42i 1873 20S 23i 201 22 22^ 20 27 29 31} 31i 47 46i 1877 23 23} 22 23 21 24 31f 88 80S 341 441 494 1881 29i 26J 31: 26 24i 24J 28^ 28 83| 36| 471, 60i 1885 l.» m 22 16 12§ 19i m 32 19 34S 32i 44S 1889 16* 11 16J m 15^ 211 19i 29: 20 261 23} 26} Av, 8 courses 1 1853-1881 J 30 321 30i 28 271 271 861 38 401 421 481 , m Av. 2 courses ) 1886 & 1889 ]■ 16i m 191 141 14 m 18i 31i 19} 30i m 35} lb. Ih. lb. lb. lb. lb. lb. lb. lb. lb. lb. lb. 1849 2200 2983 3139 3226 1870 2111 3209 8327 2842 2088 3709 3646 1853 2187 2430 2210 2077 2008 1873 2729 276B 2695 2604 3328 2981 1867 2330 2600 2430 2312 1646 1475 2695 2780 2400 2436 3670 3405 1861 2190 2622 2018 1970, 1964 2000 2476 2563 3920 3940 4178 8940 1865 1828 2154 1809 1460 1609 1615 2043 2244 2398 2595 3274 2968 1869 1628 1948 1648 1944 1873 2026 2266 2401 3064 3309 8244 3229 1873 1374 1843 1311 1495 1370 1565 1611 1841 1626 1723 2706 2466 1877 1244 1291 1275 1341 1054 1174 1706 1994 1626 1918 2646 3126 1881 1556 1484 1668 1468 1239 1269 1600 1430 1766 1863 2993 3078 1885 1618 1270 1768 1379 1043 1441 1480 a368 1628 2461 2778 3386 1889 963 931 996 866 966 1221 1136 1618 1231 1685 1776 2030 Av. 8 courses 1 1863-1881 ) 1792 1971 1784 1758 1668 1623 2116 2260 2423 2647 3263 3146 Av. 2 courses ) 1885 & 1889 j 1235 1101 1382 1122 1004 1331 1307 1986 1380 2073 2277 2708 TOTAL PRODUCE. lb. lb. lb. lb. lb. lb. lb. lb. lb. lb. lb. lb. 1849 4149 6666 6785 6046 8676 3841 6708 5885 6026 3794 6844 6206 1863 4046 4464 4161 8817 3876 3680 6110 5068 4849 4873 5672 619U 1867 4777 6837 4912 4668 3272 3076 6326 5741 5091 6168 7261 6930 1861 4248 4718 8871 3636 3807 3776 4803 4982 7419 7891 7664 7148 1866 -8669 4182 3696 2961 3170 3394 4122 4457 4799 6148 6763 6308 1869 2881 3368 2843 8387 3328 3686 3999 4318 5414 5800 6491 670) 1873 2596 2717 2636 2844 2713 2876 3209 3676 3412 3573 .5478 5018 1877 2602 2628 2609 2673 2304 2558 3530 4167 3406 3890 6217 6963 1881 3170 2922 8297 2929 2576 2641 3083 , 3061 3661 3867 6720 6964 1885 2402 1960 3066 2236 1833 2688 2676 4193 2643 4426 4624 6946 1889 1789 1610 1898 1630 1776 2402 2248 3250. 2362 3134 3046 3409 Av. 8 courses ) 1853-1881 f 3497 ' 3790 3491 3361 3131 8196 4148 4417 4755 4962 6018 5903 Av. 2 courses ) 1885 & 1889 1 2096 1736 2477 1882 1804 2470 2412 3722 2603 3780 3835 4677 EOTATION OF GEOPS. 207 roots, duriiig the root-crop period itself, and during the succeeding winter, before the sowing of the barley. There was, therefore, very good preparation for the barley. It will be seen further on that, when grown continuously without manure, both wheat and barley yield more in proportion to their respective averages under ordinary cultivation than does either of the fallow crops — the roots or the leguminoiis crops. Yet, the produce of barley in rotation without manure BarUy m was much in excess of that when it is grown continuously : ™*2*»™ ji -t , • TIT !• 1 i. 1 T ^^^ grown the explanation doubtless being, as above referred to, that contmu- the crop had been grown after well-cultivated bare fallow. ously. Next, it is to be observed that, there having been practically no crop of roots without manure, there was no material difference between the yield of the succeeding barley where the roots were carted off or where they were fed on the land. Turning now to the produce on the four plots with super- With phosphate alone, it is seen that whilst the average yield of ^JP^^- barley on the two portions from which the roots had been carted off was under 28 bushels, that on the portions where they had been fed on the land was, in one case more than 35|-, and in the other 38 bushels. The effect on the one hand of the removal of the larger crop of roots, and on the other of the retention on the land of the greater part of its constituents, is thus very evident. It is further to be re- marked, that the produce of barley where the roots grown by superphosphate had been removed from the land was even less than on the two corresponding portions of the un- manured plot. Thus, there is confirmation of the supposi- tion that the higher crop of barley without manure was due to the previous preparation, and conservation of constituents, by fallow : and that the lower produce on the superphosphate plot where the roots had been removed was largely due to so much greater exhaustion, especially of the available nitrogen, of the surface soil. ZsText it is seen that, on the plots where the mixed manure Mixed containing nitrogen had been applied for the preceding '"■™«^«- turnips, the produce of barley was on a much higher level ; and it was much higher on the portions where the turnips had been fed on the land than on those from which they had been removed. It may be observed that the produce, even on the plots ^/scfe of with superphosphate alone, was, where the roots had been fed ^|^"™-^' on the land, about the average of the country at large under roots m, ordinary rotation — namely, from 36 to 38 bushels; whilst, ^'™'^' on the full manured plot, the produce was much more than this — namely, in one case 40f , and in the other 42f bushels, where the roots had been removed; and where they had 208 THE KOTHAMSTED BXPEKIMENTS. Table 58 explained. Intervals between clover crops. been fed on the land, in one case 48f , and in the other 47|- bushels. Thus, then, the effect on the succeeding barley of the full mineral and nitrogenous manure applied for the preceding turnips is very obvious ; whilst the effect on the one hand of the removal of the root-crop, and on the other of the reten- tion on the land of most of its constituents, is also very marked. The experimental results relating to the second crop of the course — the barley — so far fully confirm, there- fore, the explanations which have been given of the beneficial effects of root-crops grown under the ordinary conditions of manuring, on the succeeding cereal grown in alternation with them. Examination of the results relating to the quantities of straw, and of total produce (grain and straw together), as given in the middle and lower divisions of the table, will show that they fully bear out the general conclusions that have been drawn from a consideration of the produce of the grain alone. The Leguminous Crops (or Fallow). Table 58 (p. 209) gives for the third element of the typical four-course rotation — the leguminous crops — the results ob- tained in each of the eleven years of the forty-four in which they were grown, in exactly the same form as those previ- ously recorded for the turnips and for the barley. Bat as in some of the years clover, and in others beans, were grown, the averages are here taken, not for the eight and for the two courses, as with the other crops, but, respectively, for the four years of the eleven in which clover was grown, and for the seven in which beans were grown. A glance at the table brings to view some of the difficulties connected with the growth of these crops. Thus, although the scheme of the four-course rotation supposes the growth of red clover as the third crop of each course, that is once in four years, it has in fact only been grown four times in the forty-four years — namely, in the first, seventh, ninth, and tenth courses ; and when it failed beans were grown instead. It is, indeed, a matter of general knowledge and experience, that it is only on a few descriptions of soil that clover can be grown so frequently as every fourth year ; and in many cases it is not attempted to grow it more than once in eight years. The difficulty of growing red clover qr beans frequently on ordinary arable land has been very fully illustrated in our experiments on the growth of leguminous crops. On the other hand, it has been found that red clover may be grown for many years in succession on rich garden soil ; and, further, ROTATION OF CKOPS. 209 TABLE 58. — Experiments on the Rotation op — Roots, Barley, Clover (oe Beans), or Fallow, and Wheat; in Agdell Eield, Rothamstbd. 11 courses, 44 years, 1848-1891. 3. clover (OR BEANS), OR FALLOW. Un manured. Courses 1-9 superphosphate only. Courses 10 and 11 mixed mineral manure. Mixed mineral and nitro- genous manure. Years. Boots carted. Roots fed. Roots carted. Roots fed. Roots carted. Roots fed. Fal- low. Beans or clover. Fal- low, Beans or clover. Fal- low. Beans or clover. Fal- low. Beans or clover. Fal- low. Beans or clover. Fal- low. Beans or clover. BEANS; DRESSED CORN— 1854, '58, '62, '66, '70 '78, and '90. (CLOVEE-1850 , '74, '82 and '86 ) Bush. Bush. Bush. Bush. Bush. Bush. 1850 (clover) (clover) (clover) (clover' (clover; (clover) 1854 6i 6, 5 101 H 131 1858 ei 6i 6 3S 121 14i 1862 29 27 29 30 *1J 1866 lOi 81 7 10 20§ 24f 1870 181 irj 16 j 15| 24i 26§ 1874 (clover) (clover' (clover) (clover (clover^ 'clover) 1878 81 7» 71 m 20i 261 1882 (clover) [clover; (clover) (clover; (clover' (clover] (clover) 'clover) 1886 (clover) 'clover] (clover) 'clover) 1890 7 Si 24g 24 iH Ui Average 7 " courses.besms. 1864, '58, '62, - •66, -70, '78, and '90 J 111 118 13S 16i m 23i BEANS ; STRAW— 1854, "68, 62, '66, •70, '78, and '90. (CLOVER— 1850, '74, •82, and ■86.) lb. lb. lb. lb. lb. lb. 1860 (clover; (clover; (clover) (clover' (clover) (clover) 1864 1055 953 1103 1378 1356 1605 1868 1100 966 1166 1320 1520 1760 1862 1840 1845 2150 ' 2155 3280 2945 1866 1013 905 978 1835 1990 2155 1870 788 710 708 878 1056 1008 1874 (clover; (clover! (clover) (clover; (clover) (clover) 1878 740 775 1046 1360 1656 1880 1882 (clover; (clover (clover) (clover' (clover) [clover) (clover) [clover) 1886 (clover; (clover [clover) [clover) 1890 60S 633 1764 1630 1102 1059 Average 7 courses,beans. 1854, '68, '62, J- 1013 969 1280 1507 1708 1773 '66, '70, '78, and '90 J CLOVER (AS HAT)— 1860, '74, 82, and '86. BEANS (CORN and STRAW)— 1854, '58, '62, '66, 70, '78, and '90. lb. lb. lb. lb. lb. lb. 1 lb. lb. lb. lb. lb. lb. 1860 (6440) (5920) (7027) (6413) (6799) (6329) (6739) (6580) (7697) (6920) (7275) (6753) 1864 1446 1367 1534 2124 2066 2544 1858 1515 1307 1606 1895 2367 2764 1862 3661 3646. 4040 4027 5990 5520 1866 1689 1485 1463 2481 3343 3782 1870 1591 1864 1778 1867 2664 2746 1874 (2838) (2497) (5093) (6186) (7904) (7708) 1878 1301 1255 1657 2241 2963 3617 1882 (2935) (1286) (2492) (1305) (6700) (7927) (4696) (8882) (3266) (9374) (3645) 1886 (4925) 1890 1079 1197 3441 3269 2146 2196 Average 7 "^ courses,beana. 1854, '68, '62, - 1754 1716 2203 2558 3075 3308 •66, ^70, •78, and '90 J Average 4 ) courses, clover, ( 1850, '74, '82, f 3245 2927 5762 6097 6740 6870 and '86 ) VOL. VII. 210 THE EOTHAMSTED EXPEKIMENTS. that on ordinary arable land where clover had entirely failed, some other Leguminosae, having more extended root range, or more powerful root habit, grew luxuriantly, and yielded large crops, containing large amounts of nitrogen, for a number of years in succession. Lastly, in another field, where beans had frequently failed, red clover was afterwards sown, and gave unusually large crops. * Eeferring to the results in Table 58, it is seen that when clover was grown in 1850, that is in the iirst course, and when it had not been grown on the same land for many years, large crops were obtained on all the plots ; though the larger where the mixed manure including potash, and also nitrogen, had been applied for the root-crop three years previously. Por the second, third, and fourth courses, clover was sown with the preceding barley, but in all three it failed in the winter, and beans were grown instead ; that is, in 1854, 1858, and 1862. After these repeated failures, clover was not sown for the fifth and sixth courses, but beans were taken instead, in 1866 and in 1870. In the seventh course, clover was sown again, with the barley, and gave three cuttings in 1874 ; that Effects of is, twenty-four years since the last good crop. Without dov^"nd m^'imre, the produce was, however, not much more than one ieans. ton per acre ; with superphosphate it was much more ; and with the mixed manure, including potash, much more still — corresponding to about 3^ tons of clover hay. For the eighth course clover was not sown, but beans were taken in 1878. For the ninth and tenth courses, however, clover was again sown, yielding in the ninth (1882) even more than in 1874; but in the tenth (1886) very much smaller crops, though more with mineral manure alone, now including potash, than with the mixed manure containing nitrogen also. Lastly, for the eleventh course, clover was again sown with the barley, but failed, and in 1890 beans were grown instead ; the crops, as. in the case of the clover in the tenth course, being greater with mineral manure alone (now including potash) than with the mixed manure containing nitrogen also. Faiimes of Thus, in only four out of the eleven years in which clover should have been grown, was any crop obtained, and beans had to be taken in the other seven. The produce of clover is given in the lower division of the table, side by side with the- total produce (corn and straw) of the leans; and the results- for the clover are entered in parentheses. - Briefly to summarise the results given in the table, it may dove^ aM ^^ stated that the average produce of clover, reckoned as hay,, was, without manure, rather over 3000 lb. ; with the super- phosphate (in the last year with potash, soda, and magnesia, also) nearly 6000 lb. ; and with the mineral and nitrogenous clover. ROTATION OF CEOPS. 211 manures together for each course, about 6800 lb. With the mineral manure alone, therefore, there was about twice as much, and with the mineral and nitrogenous manures together, considerably more than twice as much, as without manure. Compared with these amounts of clover reckoned as hay, the seven bean crops (corn and straw together) gave an average of about 1700 lb. without manure, of nearly 2400 lb. with mineral manure alone, and about 3200 lb. with the mineral and nitrogenous manures together. Not only, therefore, was the average produce of the bean crop very much less than that of the clover, but in point of fact it was only in one year, 1862, that anything like a really good crop of beans was obtained. It may be added, though the point will be further illustrated presently, that the crops of the four years of clover contained, even without manure, about as much nitrogen as, and with each of the two manures considerably more than, those of the seven years of beans. In fact, the average produce of the bean crop, and of nitrogen Nitrogen in it, was very much less than in the case of the clover, i^kgumes Nevertheless, even the average yield of nitrogen was much more in the beans than in either of the cereals with which they were grown in alternation. Thus, without manure, the four clover crops gave an average of 60.2 lb. of nitrogen per acre, and the seven bean crops 34.9 lb. ; but over the eleven courses the barley gave an average of only 28.0 lb., and the wheat of only 31.7 lb. "With mineral manure alone, the average yield of nitrogen was, in the clover 119.2 lb., in the beans 49.2 lb., in the barley only 27.7 lb., and in the wheat only 39.3 lb. Lastly, with mineral and nitrogenous manure together, the clover gave an average yield of nitrogen of 134.6 lb., the beans of 64.1 lb., the barley 41.2 lb., and the wheat 43.5 lb. There can, indeed, be no doubt, that the leguminous Zegumm- crops, and especially the clover, growing on land in the same "«« "'■op* condition, and similarly manured, have the power of taking nitrogen. up much more nitrogen over a given area from some source, than the cereals with which they are interpolated ; and that the beneficial effects of the growth of such crops in rotation with the cereals are intimately connected with this capability. Before passing from the results in Table 58 it may be Zegmmn- observed that, both with mineral manure alone, and with "^l^^ mineral and nitrogenous manure together, there is rather more consimp- produce, both of the clover and of the bean crop, where the *"^ "f roots had been fed upon the land, than where they had been lamd. carted off ; that is the higher the condition of the land. Thus, then, the effects of the treatment of the iirst crop of the course — the roots — on the produce of the third or leguminous crop are clearly shown. 212 THE EOTHAMSTED EXPERIMENTS. Legwm/in- ous crops as a svisti tutefor fallow. As already referred to, in the second and subsequent courses, when the third year came round each plot was divided, clover or beans being grown on one half, and the other half left fallow. We have, therefore, the means of com- paring the effects on the other crops of the rotation — of fallow on the one hand, which of course removes nothing (though there may be the more loss by drainage), and of growing beans or clover on the other, a characteristic of which is the assimilation, and consequently the removal in the crops, especially of large amounts of nitrogen, but of other consti- tuents also ; at the same time, however, leaving in the land more or less of nitrogenous crop-residue. Such a comparison obviously has a special interest, since it is chiefly as a substi- tute for fallow that the growth of leguminous crops has been introduced into our rotations. Va/riations with The Wheat Grops. Table 59 (p. 213) records the results obtained with the fourth element of the rotation — the wheat — exactly in the same form as in the case of the other crops. Looking first to the figures relating to the individual years, it is seen that, under each condition of manuring or other treatment, there is an enormous variation in the amount of produce in the different years, according to the seasons. Thus, taking for illustration the results in the first column under each of the three main conditions as to manuring, that is where the roots were carted from the land, and where in the third year of the course it was left fallow, there was, without manure, only 10^ bushels of, wheat in 1879, but 45 bushels in 1863 ; on the superphosphate plot there was in 1879 only 14f bushels, and 46 bushels in 1863 ; and on the mixed manure plot only 12| bushels in 1879, but 52| bushels in 1863. Or, comparing the quantities of total produce, corn and straw to- gether, which more directly represent the amounts of growth, we have, on the same plots, without manure, 2162 lb. per acre in 1879, and 7446 lb. in 1863 ; on the superphosphate plot 2905 lb. in 1879, and 7626 lb. in 1863 ; and lastly, on the mixed manure plot, only 2478 lb. in 1879, but 8837 lb. in 1863. The cases cited are those of the most extreme fluctuations due to season ; but a glance at the columns will show that there were very considerable variations in other years, under each condition as to manuring, or other treatment ; whilst the amounts of the variations differ more or less under the different soil conditions. It _ will be obvious, therefore, that if we would fairly compare with one another the effects of EOTATION OF CROPS. 213 TABLE 59. — Experiments on the Eotation of — Roots, Baelby, Clover (or Beans), or Fallow, and Wheat; in Asdell Field, Eothamsted. 11 courses, 44 years, 1848-1891. 4. WHEAT. XJnmanured. Courses 1-9 superphosphate only. Courses 10 and 11 mixed mineral manure. Mixed mineral and nitrogenous manure. Tears. Roots carted. Roots fed. Roots carted. Roots fed. Roots carted. Roots fed. Fal- low. Beans or clover. Fal- low. Beans or clover. Fal- low. Beans or clover. Fal- low. Beans or clover. Fal- low. Beans or clover. Fal- low. Beans or clover. DRESSED GRAIN. Bush. Bush. Bush. Bush. Bush. Bush. Bush. Bush. Bush. Bush. Bush. Bush. 1851 30J 28^ 3U 30J 31« 28 S2i 32 30i 28J 27i 813 1865 371 .35i 37i 34^ S8J 36i 37 36| 38 871 378 403 1859 36| 36i 36J 30i 37i 343 39i 37i 42 393 m 38J 1863 i& 34i 42 30i 46 34J 49i 41S 62 46§ 49 441 1867 27 21 23§ 16i 264 191 27 25 22 233 19i 21i 1871 14 20 14J 2ii 16i 23i 16! 23 17 24 17i 25^ 1876 24; 21 , 24^ 19g .28 28^ 30, 313 29 31i 30 30i 1879 10 10 111 Si 14: 14i 14 16i 12 13 lOi 14 1SB3 33 29 34i 263 38^ 36J 40^ 40 37J 46i 39i 50i 1887 34 26 88^ 27f 41| 42i 40 443 89 42i 41 48i 1891 32 29 31i 26i 36 m 40 50i 41 441 46i 42 Av. 8 oonrses 1 1855 to 1883 ! 28i 26 27i 23i 303 28i 313 311 31J 32g 30i 33i Av. 2 courses ) 1887 and 1891 j 33i 27i 32| 261 38i 42i 40i m 40i 431 434 42| lb. lb. lb. lb. lb. lb. lb. lb. lb. lb. lb. lb. 1851 3278 3431 8498 3760 3497 8871 3834 4014 3610 8552 3969 4036 1866 4295 3619 4070 3361 4286 8626 4492 3611 4962 3942 6107 4370 1859 4816 4030 4045 3866 4310 3930 4720 4320 6830 4610 6646 4955 1863 4563 3468 4296 3008 4690 3390 6051 3888 6496 4698 6638 4919 1867 2664 2143 2698 1624 2774 1966 2989 2648 2860 3003 2906 1654 1871 2075 2799 1946 2656 2128 3048 2240 2980 2628 3440 2863 3644 1875 2833 2430 2861 2863 8230 3636 3526 3928 8628 4686 4085 4385 1879 1493 1824 1612 1219 1966 1771 1843 1771 1691 1668 1426 2138 1883 2994 2280 8231 2060 8686 3021 4110 3276 3689 4024 4028 4605 1887 2605 1869 2666 1844 3466 3298 3480 3468 8808 3423 8768 3645 1891 2941 2598 5898 2318 3686 3995 4103 6017 4288 4676 4938 4809 Av. 8 courses! 1855 to 1888 / 3163 2762 3081 2441 3383 3023 3621 3308 3782 3758 3960 3821 At. 2 courses \ 1887 and 1891 / 2728 2229 2777 2081 3526 3647 8792 4243 3798 8999 4360 3977 TOTAL PRODUCE. lb. lb. 11). lb. lb. lb. lb. lb. lb. lb. lb. lb. 1851 6290 5389 5684 6855 6617 6263 6062 6176 6642 6600 5801 6169 1855 6736 6859 6473 6626 6766 6789 6961 6921 7428 6371 7499 6992 1859 6582 6262 6270 6266 6671 6120 7242 6689 8066 7164 8136 7417 1863 7446 6621 6999 4941 7626 5619 8194 6662 8837 7627 8747 7721 1867 4830 3473 4126 2606 4420 3222 4702 4242 4828 4667 4180 3023 1871 3004 4092 2840 3994 3133 4621 3198 4404 3747 4942 3926 6286 1875 4412 3784 4396 3642 5065 5328 6443 5964 6448 6699 5942 6292 1879 2162 1987 2361 1800 2906 2729 2756 2781 2478 2493 2100 3034 1883 5140 4175 6446 3741 6208 6400 6778 6901 6132 6921 6536 7743 1887 4689 3483 4811 3560 6103 6994 6105 6882 6894 6103 6410 6409 1891 4868 4371 4763 3921 5742 6546 6509 8034 6748 7260 7610 6811 At. 8 courses \ 1855 to 1883 f 4976 4407 4863 3927 6348 4841 5668 6307 6808 5847 6883 6932 Av. 2 courses 1 1887 and 1891 f 4779 3927 4787 3736 6923 6270 6307 7183 6321 6677 7010 6610 214 THE EOTHAMSTED EXPEEIMENTS. the varying conditions, it is important to take the average results of a sufficient number of years to eliminate the influ- ence of the varying seasons. Most of our illustrations will, therefore, be drawn from the average results over the eight years of wheat in the second to the ninth courses ; but some reference will also be made to the averages for the tenth and eleventh courses. Let us first compare the average amounts of produce of grain under the three main conditions as to manuring, exclud- ing, however, those obtained on the portion of the unmanured plot where the roots were fed on the land, and where beans or clover were grown in the third year of each course ; as the crops, especially of the barley and of the wheat, were some- what adversely affected by a dell on one side of the plot, the surface-soil being in consequence comparatively shallow. The figures show that, on the three portions, the produce ranged, without manure, from 26 to 28^ bushels; with super- phosphate, from 28 J to 31 1; and with the mixed manure. Effects of from 30|- to SSJ bushels. Or, taking the amounts of total manwres. pjQ^mjg (grain and straw together), the range of amounts is — without manure, from 4407 to 4976 lb. ; with superphosphate, from 4841 to 5658 lb. ; and with the mixed manure, from 5808 to 5932 lb. There is, therefore, both in grain and in total produce of the fourth crop of the course, an obvious difference, but certainly less than might have been expected, due to the varying conditions as to manuring in the first year, separated from the fourth by the growth and removal of the intermediate crops. Wheat and Next, Comparing the effects on the fourth crop — the wheat the con- — gf ^^ removal of the first — the turnips — or the retention of roots on of them, or of most of their constituents, on the land, it is seen that without manure, under which conditions there were practically no roots grown, the difference of result from removal or otherwise is quite immaterial, and is probably accidental. With superphosphate alone, and more roots grown, the nitrogen of which was doubtless obtained from previous accumulations within the soil, the removal or the retention on the land of the constituents of the turnips should, therefore, more materially affect the condition of the soil for the growth of the succeeding crops. It was shown that the effect was very marked on the barley which imme- diately succeeded the roots. There was also somewhat less produce, both of clover and of beans, where the roots had been removed ; and now, in the case of the fourth crop — the wheat — there is still distinct effect. Thus, taking the fallow portions, there was an average of 30f bushels of wheat where the roots had been removed, and 31f bushels where they EOTATION OF CROPS. 215 were fed or retained on the land ; the corresponding amounts of total produce being 5348 lb. and 5658 lb. Or, taking the produce on the bean and clover portions, there were 28| bushels of grain where the roots had been removed, and Slf bushels where they had not been removed, the corresponding amounts of total produce being 4841 lb. and 5307 lb. Lastly, with the mixed manure, including nitrogen, the average pro- duce was, on the fallow portions, 31| bushels after the removal of the roots, but only 30| where they had not been removed, the amounts of total produce being, however, 5808 lb. and 5883 lb. On the bean or clover portions, the results were 32f bushels where the roots were carted, and 33J bushels where they were not removed ; and the amounts of total produce were 5847 and 5932 lb. Eeference to the average produce of the last two courses, the tenth and eleventh, the wheat years of which were of more than average productiveness, shows, iu the case of the manured plots, more striking difference in the amount of the fourth crop due to the removal or the retention on the land of the constituents of the first crop — the roots. The roots of those courses were, however, more than average in amount. The results, both with superphosphate alone and with the mixed manure, afford, therefore, distinct evidence of the effect of the removal or otherwise of the first crop of the course — the turnips — not only on the second and third crops, but on the fourth crop — the wheat — also. The next point is to illustrate the difference of effect on the Effects of other crops of the rotation, on the one hand of the growth and ^Jt^^^ removal of the highly nitrogenous leguminous crop, and on /allow on, the other of fallowing which removes nothing; and first as ««|««*«»'* to the wheat, which we are now specially considering, and ceeding which immediately succeeds the leguminous crop or the '"■"•?'*• fallow. A careful examination of the average results over the eight courses (second to ninth) will show that, both without manure and with superphosphate alone — that is, under conditions of exhaustion, especially of available nitrogen — the wheat crops were in every case higher after fallow, with its supposed accumulation, than after the leguminous crops, which removed much more nitrogen than the succeeding wheat would require. On the other hand, on the mixed manure plots, where the condition of the land, and especially its nitrogenous condition, was not exhausted, but fairly maintained — there was even rather more average produce of wheat after the removal of the highly nitrogenous leguminous crops than after the accumulations of the fallow. It is unsafe to form general conclusions from the results of 216 THE KOTHAMSTED EXPERIMENTS. individual years, since the characters of the seasons may have so much influence. But it may be observed that, after the heavy crops of clover on the superphosphate plots in 1882, and more where the roots were fed than where they had been removed, the wheat crops of the next year, 1883, which were higher than average, were lower after the leguminous crop than after fallow; whilst, on the highly manured plot, they were much the higher after the leguminous crop. In the tenth course, however, after the use of potash as well as superphosphate, there were fair but by no means such heavy crops of clover as in the very favourable season of the preced- ing course, and there was less where there had then been the larger crop ; and in the eleventh course also there was less total produce of beans where the heavier crop of clover had been grown in the ninth course. The result was, that on the average of the last two courses the wheat gave less instead of more total produce after fallow than after the leguminous crops; but more where the roots had been fed than where they had been carted — that is, more where the land was the less exhausted. The general result is, that where there was not exhaustion, but accumulation due to manure and to increased crop residue, the growth and removal of the leguminous crops not only gave large amounts of nitrogen in the removed crops, whilst the fallow yielded none, but also left more available nitrogen for the succeeding wheat than was rendered available (and remained) from the resources of the soil after the fallow. In other words, not only were the nitrogen and other con- stituents obtained in the leguminous crops an entire gain compared with the result of fallow, but, on the average of years, a somewhat larger succeeding wheat crop was obtained as well. Here, then, is a striking illustration of the advantages of the interpolation of leguminous crops instead of fallow with the cereals in our rotations ; and it is seen that the benefit may be the greater if the land be not abnormally exhausted, as was the case on the continuously unmanured and on the superphosphate plots. Although there was thus great difference between the effects, on the one hand, of the growth and removal of a leguminous crop, and on the other of fallo\^, so far as the third year of the course is concerned ; yet, where the manurial conditions were not defective, there was even more wheat succeeding the leguminous crop than succeeding the fallow. The influence of the conditions of the third year of the course does not, however, seem to extend in any marked degree to the crop succeeding the wheat — that is, to the roots com- EOTATIOX OF CROPS. 217 mencing the next course, and to the barley succeeding the roots. So far as the roots are concerned, the average results over the eight courses show, both without manure and ^vith super- phosphate alone, that is on the most exhausted plots, that the advantage, if any, is more with the fallow than with the legu- minous plots ; whilst, with the full manure, there is scarcely any difference of result clearly traceable to the treatment of the land in the third year of the preceding courses. Over the last two courses, again, without manure no benefit accrued to the root-crop by the growth of the leguminous crop as compared with fallow. On the superphosphate plots, how- ever, now with potash, soda, and magnesia, as well, and doubtless more leguminous produce accordingly, there were more roots on the leguminous than on the fallow plots ; but, with the full manure, there was practically no difference in the produce of roots on the fallow compared with the legu- minous crop plots. Obviously, the fact that there was not materially less produce of roots where the leguminous crops had been grown and removed, as compared with where the land had been fallow, is of itself evidence of the beneficial rather than exhausting effect of their ^owth and removal, so far as the requirements of the succeeding crops are concerned. Xor is the effect of the growth and removal of a leguminous crop, compared with fallow, very definite on the barlej' suc- ceeding the manured roots. It is, however, over the eight courses, in favour of the growth of the leguminous crops ; and, though with very small crops, it is, excepting without manure, miuch more so over the last two courses. From the results as a whole it may be concluded that, where the land was the most exhausted, the growth of the leguminous crop w£is correspondingly limited, and, being at the expense of the little accumulation that there was, its removal further exhausted the immediately available sup- plies; whilst, where the accumulations were greater, the growth was dependent on a more extended root-development, and therefore greater range of collection ; the luxuriance was much greater, and the surface-soil at any rate gained by an increased amount of highly nitrogenous leguminous crop- residue. It has farther been seen, that the effects of the manuring and treatment of the first crop of the course — the turnips — were manifest in the produce of the fourth crop — the wheat ; and also that the effects of fallowing, or of grow- ing and removing a highly nitrogenous crop, in the third year, were clearly traceable on the crop of the fourth year, and to some extent, though in a much less degree, on the subse- quent crops commencing the next course. 218 the eothamsted expeeiments. The Amounts of Peoduoe geown in Eotation, and in THE VAEIOUS CeOPS GEOWN CONTINUOUSLY. Obviously, when considering what are the benefits arising from rotation as distinguished from the growth of the indi- vidual crops continuously, it is desirable, as far as practicable, to compare the results of the two methods in regard to their yield per acre of some of the more important constituents of the crops. For the purposes of such a comparison, it will be of interest to illustrate the point by reference specially to the amounts of dry matter, nitrogen, total mineral matter {ash), phosphoric acid, and potash {and in some cases of lime), in the crops grown in rotation, and in those grown continuously, under as far as possible parallel conditions as to manuring. Methods of Accordingly, so far as results obtained under rotation are IvT^f!.^"" concerned, the amounts of each of the above constituents are plained. Calculated m the produce per acre of the respective crops, m each of the eight courses (second to ninth), under each of the t\^elve different conditions as to manuring, or other treat- ment ; and the average amounts of these per acre per annum are compared with th9se in the individual crops grown con- tinuously, as a rule in the same seasons as those in which the rotation crops were obtained, and under the same, or nearly parallel, conditions as to manuring. The amounts of the constituents removed per acre in the rotation crops are calculated from the results of actual an- alyses ; and in the case of the continuously grown crops the amounts of dry matter and ash, and sometimes those of nitro- gen, are also calculated from direct determinations ; but generally the nitrogen, and always the phosphoric acid, potash, and lime, are calculated from the percentage compo- sition of the rotation crops grown under parallel conditions as to manuring. It may be stated that, for the purposes of the illustrations given, the results of 60 complete analyses of the ashes of representative samples of the rotation crops, and of 8 of the ashes of the bean plant taken at different stages of its growth, have thus contributed ; and it may be added, that the ash-analyses were executed by Mr E. Eichter, for- merly in the Eothamsted Laboratory, but now for some years of Charlottenburg, Berlin. The Amownts of Dry Matter produced in the Rotation, and in the Continuous Crops, Table 60 (p. 219) shows the average annual amount of dry matter produced per acre, in each of the four crops — roots, barley, leguminous crop, and wheat — ^grown in rotation, and ROTATION OF CEOPS. 219 TABLE 60. — Experiments on the Rotation of — Roots, Baklet, Clover (or Beans), or Fallow, and Wheat ; in Agdbll Field, Rothamsted. 8 Courses, 32 Years, 1852-1883. AVERAGE amounts OF DRT MATTER PER ACRE PEE ANNUM, GROWN IX ROTATION, COMPARED WITH THOSE IS THE CROPS GROWN CONTINUOUSLY. Unmannred. Saperpta ospliate. Mixed mineral and nitrogenous manure. Roots carted. Roots fed. Roots carted. Roots fed. Roots carted. Roots fed. Fal- low. Beans or clover B«,ns l""- clover Fal- low. Beans or clover Fal- low. Beans or clover Fal- low. Beans or clover Fal- low. Beans or clover SWEDISH TURNIPS. fRotaiion . . . Roots ■{ Continnonsi . . lb. 359 236 lb. 228 236 lb. 323 236 lb. 205 236 lb. 1724 945 lb. 1631 945 lb. 1918 945 lb. 1901 945 lb. 3081 1876 lb. 3128 1876 lb. 3107 1876 lb. 3069 1876 VRotn.-l-or-cont. 123 -8 87 -31 779 686 973 956 1205 1252 1231 1193 fBotation . . . Leaves-j Continnonsi . . 56 49 49 49 62 49 45 49 161 142 176 142 179 142 200 142 310 345 355 345 333 345 354 345 ^.Rotn.-^or-cont. 7 3 -4 19 34 37 58 -35 10 -12 9 Botation . . . Total ■ Continuoasi . . 415 285 277 285 375 285 2502 285 1885 1087 1807 1087 2097 1087 2101 1087 3391 2221 3483 2221 3440 2221 3423 2221 I- Botn. -f or - cont. 130 -8 90 -35 798 720 1010 1014 1170 1262 1219 1202 BARLET. (Rotation . , . Grain - Continaons . . 1396 875 1489 875 1399 875 1307 875 1284 1128 1294 1128 1665 1128 1780 1128 1917 2298 1987 2298 2262 1 2273 2298 2298 V Botn. -I-or - cent. 521 614 524 432 156 166 537 652 -381 -311 -36 1 -25 TBotation . . . Straw -i Continnons . . 1493 947 1647 947 1486 947 1459 947 1307 1052 1355 1052 1765 1052 1879 1052 2029 2489 2129 2489 2701 2613 2489 1 2489 V.Botn.-f-or-cont. 546 700 , 539 612 255 303 713 827 -460 -360 212 1 124 TBotation . . . Total 4 Continuous . . 2889 1822 3136 1822 2885 1822 27662 1822 2691 2180 2649 2180 3430 2180 3659 2180 3946 4787 4116 4787 4963 4787 4886 4787 l-Botn.-l-or-cont. 1067 1314 1063 944 411 469 1250 1479 -841 -671 176 j 99 BEANS (6 COURSES), CLOVER (2 COURSES), OR FALLOW. (•Rotation . . . Com J Continuous . . 631 234 1 625 234 640 265 1 769 ' 1 265 1147 581 1 1292 581 t Rotn.-{-or-cont. 397 i 391 375 1 1 504 566 711 r Rotation . . . Straw ■( Continuous . . 879 i 422 835 422 978 524 i ! 524 1487 799 1640 799 I Rotn.-f-or - cont. i 457 1 1 413 454 \ i 689 688 1 741 /- Rotation . . . Total } Continuous . . 1 1510 i i 656 ; 1 1460 1 656 1618 7S9 1 1082 1 ■ 789 2634 1380 2832 1380 ( Rotn. 4- or - cont. t 854 [ 1 804 829 i 1193 1254 1 1452 *''°'^"t Continuous . . , 2309 1 ? ! 1 19962 4717 7 5645 ? 6714 ? 6833 » Average of 8 courses, ) beans and clover / 1710 1 15942 2393 2897 3664 ! 3832 'Rotation . . Grain • Continnons . 1516 647 1368 647 1483 647 1235 647 1636 766 1514 766 1702 766 1668 766 1685 1238 1740 1238 1599 1238 1752 1238 VRotn.-for-cont 869 721 836 1 588 870 748 936 902 447 502 361 514 rBotation . . Straw < Continuous . 2636 1082 2296 1082 2573 1 2036 1082 ! 1082 2844 1204 2513 1204 3021 1204 2767 1204 3158 2142 3137 2142 3273 3186 2142 2142 lBotn.-^or-cont 1654 1214 1491 1 954 1640 1309 1817 1563 1016 995 1131 1044 (Rotation . . Total J. Continuous . 4162 1729 3664 1729 4066 1 32713 1729 1 1729 4480 1970 4027 1970 4723 1970 4435 1970 4843 3380 4877 3380 4872 4938 3380 ' 3380 ^ Botn. -For - cont 2423 1935 2327 1 1542 2610 2057 2753 2465 1463 1497 1492 1 1658 1 Average per acre, 19 years, 1849-52 and 1856-70. 2 Probably crop too low owing to a dell. 220 THE EOTHAMSTED EXPERIMENTS. Mawu/rial treatment. No man- ure. With superphos- Mixed manwres. contimiously, as above described. It shows the amounts, sep- arately in the roots, leaves, and total produce, of the turnips ; in the grain, straw, and total produce, of the barley, and of the wheat; in the corn, straw, and total produce, of the beans; and in the clover. It will be seen that the arrangement and headings of the columns are exactly the same as in the tables of produce already considered ; and that, for each description of crop, or part of the crop, the first line shows the amounts obtained under rotation, the second those in the crop grown continuously, and the third the difference between the two. The Dry Matter in the Turnip Crops. — Eeferring first to the upper division of the table, relating to the Swedish turnips, it should be stated that results for the crops grown continuously are not available for the same eight years as those grown in rotation ;' but for each of the three conditions as to manuring, the average for 19 years of growth is taken. So far as manuring is concerned, the unmanured and the superphosphate conditions were the same for the rotation and for the continuous crops. But, in the case of the mixed manure, the rotation plots received a larger amount of nitro- gen for the roots ; in fact, enough to carry the four crops of the course. The continuous plot, on the other hand, received a less amount each year ; but, unlike the rotation plots, with no intermediate crops to use up any available residue from the previous application. The figures show that — without manure — the difference in the amounts of dry matter produced in rotation and in con- tinuous growth are immaterial. The utter failure in both cases without manure is confirmatory of the absolute depend- ence of this valuable rotation crop on supplies within the soil itself, either from accumulations or from direct manuring. The less produce of the continuous than of the rotation crops with superphosphate is also quite consistent with the supposition that, under such conditions, the crop greatly exhausts the available nitrogen of the soil, and especially of the surface-soil. With the mixed mineral and nitrogenous manure, again, there is also considerably less production of dry substance when the crop is grown continuously than when it is grown in rotation. The result is, however, due partly to the larger amount of nitrogen directly supplied by manure to the rota- tion crops as above referred to, but partly to the fact that when the same description of root-crop, with the same char- acter and range of roots, is grown year after year on the same land, the surface-soil becomes close, and a somewhat impervious pan is formed below ; conditions which are very unfavourable for a crop which pre-eminently requires a good EOTATION OF CROPS. 221 tilth for great development of fibrous root within the soil. The resulte with the mixed manure are, of course, the most Greater comparable with those of ordinary practice; and it is clear ^^^^'■" that,, however explained, much more produce is obtained under rotation than with continuous growth. It need only further be remarked that, of the total dry matter produced, there are many times as much in the edible root as in the leaf which almost wholly remains only for manure again. The, Dry Matter in the Barley Crops. — The second division of Table 60 compares the amounts of dry matter yielded in barley, grown, respectively, in rotation, and continuously — that is, year after year on the same land. The results for the continuously grown crops relate to the average produce of the same eight seasons as those in which the rotation crops were obtained. The unmanured and the superphosphate conditions MarmridL were also quite parallel in the two series of experiments. In t^^tmemt. the case of the mixed manure results, it should be borne in mind that in the rotation experiments a quantity of manure was applied for the preceding crop — the turnips — which is supposed to carry the whole of the crops of the four years' course ; whilst, in the continuous experiments, the quantity of nitrogen, for example, which is applied each year for the immediate crop, amounts to rather more than one-fourth of that applied for four years in the rotation experiments. The figures show that — without manure — there was much Xo man- less dry matter in grain, straw, and total produce, in the crops ""■ grown continuously than in those grown in rotation ; in fact, in the total produce only about three-fifths as much. The much higher amount under rotation is quite consistent with the explanation that in the rotation experiments without manure, the roots having failed, the barley crop had, in point of fact, the benefit of the preparation which bare fallow is known to confer. With superphosphate alone, the continuously grown barley With crops yielded more dry matter in grain, straw, and in total ^^^^"' produce, than those without manure ; the excess being largely due to increased capability of utilising the available nitrogen of the surface-soil, under the influence of the phosphatic man- ure. Both sets of the superphosphate rotation crops yielded more dry matter than the continuous ones, the excess being, however, much less where the rotation roots had been removed than where they had been consumed or spread upon the land. The effect of the growth and accumulation by the previous drop root-crop, and of the more or less available manurial residue ''*»«'"«• left under the different conditions, as compared with the re- sult when the barley is grown year after year on the same land, is thus very evident. 222 THE EOTilAMSTED EXPEEIMENTS, Mixed vmnures. Mffects of the con- of roots on the land. Dry matter vfi grain and straw of la/rley. for barley- growing. As already said, the amount of nitrogen annually applied on the mixed manure plot was, for the continuous crops, somewhat more than one-fourth of that applied for the pre- ceding root-crops in the case of the rotation ,plots. Under these circumstances, the amounts of dry matter in grain, straw, and total produce, were considerably less 'in the barley grown in rotation where the roots and leaves of the turnips had been removed than in that grown continuously; but where in the case of the rotation barley the root-crops had been consumed or spread upon the land, the average yield of dry matter per acre was much more nearly identical ujider rotation and under continuous cropping; though upon the whole it was more under rotation. The effects on the second crop of the course, of the manurial and other treatment of the first crop, are here, then, further illustrated. Lastly, it is to be observed that a larger proportion of the total dry matter of the crop is, on the average, accumulated in the straw which is generally retained on the farm, than in the grain which is, as a rule, exported from it. Thus, both the actual and the comparative results clearly show, that the successful growth of the barley was directly dependent on the supplies within the soil, and that the object may be gained, either in a properly manured rotation, or by the direct application of suitable manures, including a liberal supply of nitrogen for the immediate crop. Having regard to the general economy of the farm, the former plan is as a rule the most advantageous ; though, owing to the success with which the crop can be grown by direct artificial manures, such manures are often used as supplements ; or, sometimes, a barley crop is taken after another cereal, by the aid of arti- ficial manures alone. The Dry Matter in the Leguminous Crops. — The third divi- sion of the Table (60) shows the average amounts of dry matter per acre per annum in the corn, straw, and total produce, of the six crops of beans grown in rotation in the eight years ; also the average' amounts in the same six years when the crop was grown continuously in another field. Be- low the bean results are given the average amounts per acre per annum in the clover grown in rotation in the remaining two of the eight years ; and there are also given the average amounts over the eight years, in the six crops of beans and two of clover. It will be seen, however, that there is no entry in the line for continuous crops of clover, for the simple reason that, as has been shown in various papers, it was found impossible to grow clover year after year on ordinary arable land. The figures show that, meagre as was the average produce ROTATION OF CROPS. 223 of dry matter in the crops of beans, even when grown in rotation, they were much less still when grown continuously. This was the case whether we look to the amounts in the corn, the straw, or the total produce. Indeed, the lines of Effects of total produce show that the average amounts in the contin- '^''"'■«*- uously grown crops were, under each condition of manuring or other treatment, less than half as much as those grown in rotation. In both cases, there was somewhat more with superphosphate than without manure, and more still with the mixed manure, including both potash and nitrogen, but even under these conditions, and in rotation, the produce was very small. Under each condition as to manuring, the produce of dry matter in the clover grown in rotation was more, and in some very much more, than in the beans so grown. Without manure, it averaged only about 1 ton per acre per annum ; with superphosphate, in one case more than 2, and in the other more than 2\ tons ; and with the full manure, including potash and nitrogen, it averaged more than 3 tons. Lastly, the average production of dry substance in the six crops of beans and two of clover taken together was — with- out manure only about | ton ; with superphosphate, in one case little more than 1 ton, and in the other rather more than IJ ton ; and, with the mixed manure, in both cases less than If ton. These amounts in the leguminous crops with the mixed manure were, however, greater than those obtained in the turnip crops, but less than those in either the barley or the wheat grown in rotation. The significance of the amounts grown in the leguminous crops wUl, however, be the more clearly recognised when we come to consider the quantities of nitrogen in the different crops ; and also the fact of the large proportion of the manurial constituents of the legumin- ous crops grown in rotation, that will generally be retained on the farm. ITie Dry Matter in the Wheat Crops. — The bottom division of the Table (60) shows the average amounts of dry substance in the wheat — grain, straw, and total produce — grown in rotation, and those obtained in the same years in another field under as far as possible parallel conditions as to manur- ing, but grown continuously — that is, year after year on the same land. A glance at the figures shows that, both without manure Less dry and with superphosphate- alone, the amount of dry matter '»«**?^ ™ produced was, both in gram and straw, in each case consider- tiian in ably less than half as much in the crops grown continuously '■<'to'«»» as in those grown in rotation ; and that, even with the mixed "^'^'^ manure, supplying both mineral constituents and nitrogen, it 224 THE ROTHAMSTED EXPERIMENTS. was considerably less in the continuous than in the rotation crops. So far as the unmanured and the superphosphate crops are concerned, it is obvious that the growth year after year must be much more exhausting, both of nitrogen and of certain essential mineral constituents, in a condition of com- position and of distribution within the soil and subsoil avail- able to one particular crop, than when the crop is grown in alternation with others, of different requirements, habits, and root-ranges. It has been explained that in the case of the mixed manure rotation plots there was applied for the first crop of the course, besides a full supply of mineral constituents, about 140 lb. of nitrogen ; at the average ra.te, therefore, of 35 lb. per acre per annum over the four years. But, in the case of the continuously grown wheat crops, not only a full supply of mineral manure, but 43 lb. of nitrogen as ammonium-salts, was directly applied every year. The fact of the greater amount of produce on the rotation plots would indicate, therefore, that notwithstanding the growth and removal of the intermediate crops since the application of the manure for the roots, there was more nitrogen, and more of other con- stituents also, in a condition of composition and of distribu- tion available for the wheat, than in the case of the annual direct supply. Of course, the proportion of grain and of straw in a wheat crop varies, as it also does in barley, according to variety, soil, season, and other circumstances. It is seen that, in the ex- perimental crops, whether grown in rotation or continuously, there was always much more of the produced dry matter accumulated in the straw than in the grain. Indeed, there was in some cases nearly twice as much. On the assump- tion, therefore, that as a rule the grain will be sold, and the straw retained on the farm as food and litter, very much more than half of the produced dry matter will be so re- tained. pry matter Comparing the amounts of dry matter accumulated in the oM^T^to different rotation crops, and taking as the most normal the crops. quantities obtained under the influence of the mixed manure, including nitrogen, it is seen that, on the average, the two cereal crops — the barley and the wheat — produced approxi- mately equal amounts ; and each considerably more than either of the fallow crops — the roots or the Leguminosse. ROTATION OF CEOPS. 225 The Amounts of Nitrogen in the Rotation, and in the Continuous Crops. Table 61 (p. 226) shows the average amounts of nitrogen per acre per annum, over the eight years, in the rotation, and in the continuous crops, respectively. The Nitrogen in the Boot-crops. — Without manure, with Noman- extremely small crops, but very abnormally high percentage "''^' of nitrogen in them, the amounts per acre were, in the continuously grown crops only about twice as much as annually comes down as combined nitrogen in the rain and minor aqueous deposits from the atmosphere ; whilst, even in the rotation crops, the amounts averaged but little more than in the continuous. With superphosphate alone, much larger* crops, but much wuh lower percentages of nitrogen, there was very much more *^^*''*' nitrogen taken up than without manure ; in fact, when grown in rotation from three to four times as much, and when grown continuously more than twice as much. There was, too, very much more in the rotation than in the continuous crops. The detailed results published elsewhere, relating to the con- tinuous growth of root-crops afford conclusive evidence that the increased amount of nitrogen taken up by the crop under the influence of phosphatic manures is derived from the re- sources of the soil itself, by the aid of the greatly enhanced development of fibrous feeding root induced by such manures. With the mixed manure containing nitrogen there was, as Mixed with superphosphate alone, much more nitrogen taken up "«"»»'''«• under rotation than with continuous growth. But, under rotation, there was about twice as much taken up with the mixed manure containing nitrogen as with superphosphate without nitrogen; and with continuous growth there was nearly three times as much taken up as with superphosphate without nitrogen. It is clear, therefore, that the crops, whether grown in rotation or continuously, took up much of the nitrogen supplied by the manure. Indeed, it cannot Sources of be doubted that, beyond the small amount of combined nitro- ?^7S gen annually coming down from the atmosphere in rain and the minor aqueous deposits, the source of the large amount of nitrogen of root- crops is the store of it within the soil, whether this be due to accumulations, or to direct supply by manure. On the other hand, the large amounts of produce Sources of obtained by the aid of nitrogenous manures on land to which ^l^^'^'^ no carbonaceous manure has been applied for about fifty years is evidence that the atmosphere is at any rate the chief, if not the exclusive, source of the carbon of the crops. Lastly, as to the results in the table relating to the Swed- VOL. VII. p 226 THE KOTHAMSTED EXPEEIMENTS. TABLE 61.— Experiments on the Rotation of— Boots, Babley, Clover (or Beans), or Fallow, and Wheat ; in Agdbll Field, Eothamsted. 8 courses, 32 years, 1852-1883. AVERAGE AMOUNTS OP NITBOGEN PER ACRE PER ANNUM IN THE ROTATION CROPS, COMPARED WITH THOSE IN THE CROPS GROWN CONTINUOUSLY. Unmanured. Superphosphate. Mixed mineral and nitrogenous manure. Roots carted. Roots fed. Roots carted. Roots fed. Roots carted. Roots fed. Pal- low. Beans or clover Fal- low. Beans or clover Fal- low. Beans or clover Fal- low. Beans or clover Fal- low. Beans or clover Pal- Beans l"^- clover SWEDISH TURNIPS. Rotation . . Roots -i Continuous! . Rotn.-)-or-cont lb. 9.4 6.8 lb. S.8 6.8 lb. 8.5 6.8 lb. 6.3 6.8 lb. 28.7 13.6 lb. 26.8 13.6 lb. 32.9 13.6 lb. 32.2 13.6 lb. 66.3 40.1 lb, 66.7 40.1 ' lb. 68.2 40.1 lb. 66.5 40.1 2.6 • 1.0 1.7 -1.5 16.1 13.2 19.3 18.6 26.2 26.6 28.1 25,4 I Rotation . . Leaves -j Continuous i . 2.1 2.0 1.8 ■ 2.0 1.9 2.0 1.6 2.0 . 6.1 5.8 6.5 6.8 6.9 6.8 7.6 6.8 12.2 14.1 13.9 14.1 13.0 14.1 13,9 14,1 \.Rotn.-{-or-cont 0.1 -0.2 -0.1 -0.4 0.8 0.7 1.1 ,1.8 -1.9 -0.2 -1.1 -0.2 (■Rotation . . Total < Continuous i . 1L5 8.8 7.6 S.8 10.4 8.8 6.9 2 8.8 34.8 19.4 33.3 19.4 39.8 19.4 39.8 19.4 78.5 54.2 80.6 54.2 81.2 642 79,4 54.2 iRotn.-|-or-cont 2.7 -1.2 1.6 -1.9 15.4 13.9 20.4 20.4 24.3 1 26.4 27.0 25.2 BARLEY. (Rotation . Grain X Continuous . 21.6 13.6 23,0 13,5 21,6 13,6 20.1 13.6 17.8 15.5 17.8 16,6 22,9 15,5 24,6 15.5 29,7 36,2 30.7 35.2 35.0 35.2 34.9 85,2 \.Rotn.-for-cont 8,0 9,6 8.0 6.6 2.3 2,3 . 7,4 9.1 -6,5 -4.5 -0.2 -0,3 (Rotation . . Straw i. Continuous . 6,6 4,2 7,4 4,2 6.6 4.2 6.0 4.2 6.6 4.6 • 5,7 4,5 7,6 4,5 7.9 4,6 9,5 11.4 10.0 11.4 12.6 11.4 n,9 11,4- \.Rotn.+or-cdnt 2.4 3,2 2.4 2,4 1,0 1,2 8,0. 3,4 -1.9 -1.4 1.1 0.6 ("Rotation . . Total -j Continuous . 28.1 17.7 30,4 17,7 28.1 17.7 26,7 3 17.7 23,3 20.0 23,6 20,0 30.4 20.0 32.5 20.0 39.2 46.6 40.7 46.6 47.6 46.6 46,8 46,6 lRotn.-1-or-cont 10.4 12,7 10.4 9.0 3.3 3,6 10.4 12.5 -7.4 -5.0 0.9 0,2 BEANS (6 COURSES), CLOVER (2 COURSES), OR FALLOW. ( Rotation . Corn ■ Continuous . . 27,5 9,7 27.2 9.7 30.4 10.5 30. 6 10.6 49.6 21,4 65,7 21,4 t. Rotn. -f-or - cont. 17,8 17.6 19.9 26,1 28.2 343 (Rotation . . . Straw J. Continuous 9,4 4,6 8.9 4.6 10.1 6.5 12,4 6,5 14.0 7.1 14,5 7.1 ^ Rotn. +or- cont. 4,8 4,3 4.6 6,9 6.9 7.4 (Rotation . . . Total < Continuous i . 36,9 14,3 36.1 14.3 40.6 16.0 49,0 16,0 63.6 28,5 70,2 28,6 \. Rotn. -}- or - cont. 22,6 21.8 24.5 33.0 35,1 41,7 rinvPT Rotation . . . ^^°^''^ ■ Continuous . . 66,0 ? 47.0 2 ? 124.6 ? 144.6 ? 167,0 ? 168,4 Average of 8 courses, ) Beans and Clover | * 41,6 38.9 2 61.5 72.9 89,5 94,7 WHEAT, ("Rotation . . . Grain -{ Continuous lRotn.+or-cont. (Rotation . . . Straw < Continuous . . \ Rotn. -For - cont. ( Rotation . Total i Continuous . . V. Rotn. 4-or - cont. 26.2 11.6 10.4 6.4 17.0 23.7 11.6 9,1 6,4 3,7 32,8 17.0 15.8 26.6 11.6 36.4 17,0 21,5 11,6 27,2 13,9 11,8 6,9 2,8 ( 5,9 29,7 2J 39,0 17,0 I 19,8 12,7 I 19,2 25,4 13,9 11,6^ "10.5 6.9 4,6 36,9 19,8 28,6 1S,9 14,7 12,3 5,9 40,9 19.8 28.2 13.9 11.7 6.9 89,9 19,8 6,0 13,2 10,1 42,1 34,0 13.6 10,1 43,7 34,0 27.7 23.9 3.8 13.8 10,1 3,7 41,5 34,0 7,5 1 Calculated on average produce of 19 years, 1849-62 and 1866-70. 2 Probably crop too low owing to a delL ROTATION OF CROPS. 227 isli turnips, it is seen that by far the greater part of the nitro- gen of the crops was accumulated in the edible root The Nitrogen in the Barley Crops. — The second division of Table 61 shows the average amounts of nitrogen per acre per annum over the eight years in the rotation and in the contin- uous barley crops respectively. Eef erring to the results chiefly in their bearing on the . Rotn. +or- cont. 93.6 83.6 86.5 69.1 92.1 88.1 93.2 92.7 51.2 62.2 64.0 60.11 (■Rotation . . . Total -l Continuous . . 194.2 88.0 182.6 88.0 186.6 88.0 166.62 88.0 2U.0 106.6 201.6 106.6 212.5 106.6 213.1 106.6 218.6 161.7 282.2 161.7 220.2 161.7 229.9 161.7 (Rotn. -for- cont. 106.2 94.5 98.6 77.6 106.4 95.9 106.9 107.6 66.8 70.5 68.5 68.2 1 Average per acre, 19 years, 1849-52 and 1856-70. 2 Pfobably crop too lo* o-wing to a dell. ROTATION OF CROPS. 233 TABLE 63. — Experiments on the Rotation op— Boots, Barley, Clovek (or Beans), or Fallow, and Wheat ; in Agdell Field, Kothamsted. 8 courses, .32 years, 1852-1883. average amounts op phosphoric acid per acre per annum in the rotation CROPS, compared with those in the crops grown continuously. Unmanured. Superpliospliate. Mixed mineral and nitro- genous manure. Roots carted. j Roots fed. Boots carted. Roots fed. Roots carted. Roots fed. Fal- low. Beans or clover iFal- 1 low. Beans or clover Pal- low. Beans or Clover Fal- Bean.^ 1-- cCer Fal- low. Beans or ' clover' j,^j. Beans 1™- clover SWEDISH TURNIPS. (Rotation . . . Roots i Continnons ^ . . lb. 1.26 0.88 lb. 0.77 0.88 ' lb. 1.11 0.88 lb. 0.71 0.88 lb. 7.91 4.14 lb. 1 7.68 1 4.14 lb. 8.83 4.14 lb. 8.78 4.14 lb. 1 lb. 16.67 1 17.02 *91 1 9.91 6.76 ! 7.11 2.79 1 3.17 3.07 1 3.07 lb. 18.14 9.91 lb. 17.12 9.91 (.Rotn.-for-cout. 0.38 -0.11 0.23 -0.17 3.77 3.64 i 4 69 4.64 8.23 7.21 (■Rotation . . . Leaves-j Continnons i . . 0.29 0.26 0.26 0.25 0.28 0.26 0.25 0.25 1.27 1.16 1.44 i 1.16 1.38 1.16 1.62 1.16 3.04 3.07 3.16 3.07 lRotn.^-o^-cont. 0.04 0.00 0.03 0.00 0.11 0.28 0.20 0.46 -0.28 0.10 -0.03 0.09 fRotation . . . Total 4 Continuous i . . 1.55 1.13 1.U2 1.13 1.39 1.13 0.962 1.13 9.18 5.30 9.12 6.30 1 10.19 5.30 10.40 5.30 19.46 20.19 12.98 12.98 21.18 ' 12.98 20.28 12.98 lRotn.-For-cont. 0.42 -0.11 0.26 -0.17 3.88 3.82 j 4.89 5.10 6.48 7.21 1 8.20 7.30 BARLEY. 'Rotation . . . Grain Continuous . . 11.24 6.95 11.59 6.95 11.02 6.95 9.89 6.95 12.29 10.00 11.91 10.00 15.52 10.00 16.16 10.00 18.34 21.31 18.63 21.31 21.04 21.31 20.90 2L31 LRotn. -I-or - cont. 4.29 4.64 4.07 2.94 2.29 1.91 5.52 6.16 -2.97 -2.68 -0.27-0.41 CBotation . . . Straw < Continuous . . 1.87 1.10 2.U3 1.10 1.82 1.10 1.74 1.10 1.80 1.33 1.85 1.33 2.32 1.33 2.38 1,33 2.87 3.30 2.96 3.30 3.68 3.30 3.53 3.30 lnotn.-For-cont. 0.77 0.93 0.72 0.64 0.47 0.5-2 0.a9 1.05 -0.43 -0.34 0.88 0.23 (■Rotation . . . Total -J Continuous . . 13.11 8.05 13.62 8.08 12.84 8.05 11.632 8.05 14.09 11.33 13.76 11.33 17.84 11.33 18.54 11.33 21.21 24.61 21.69 24.61 24.72 24.61 24.43 24.81 V Rotn. -{-or - cont. 5.06 5.67 4.79 3.58 2.76 2.43 6.51 7.21 -3.40 -3.02 0.11 -0.18 BEANS (6 COURSES), CLOVER (2 COURSES), OR FALLOW. fRotation . . . Com i Uontiunous . . I. Rotn. -f-or - cont. 5.15 2.11 6.14 2.11 6.81 ! 8.16 ! 8.18 3.16 11.49 6.76 13.05 6.75 3.04 3.03 3.66 1 6.02 4.74 6.311 •Rotation . . . Straw < Continuous . . . Rotn. -l-or- cont. - 1.17 0.63 1.17 0.63 1.78 1 0.95 i 1.97 0.95 1.99 1.24 2.06 1.24 0.64 0.6* , 0.83 1.02 0.76 11.82 (Rotation . . . Total < Continuous . . .Rotn.-hor-cont 6.82 2.74 6.31 2.74 { 8.69 1 4.11 1U.15 4.11 13.48 7.99 15.11 7.99 3.58 3.57 1 4.48 6.04 6.49 7.12 Clover -T^*!-'''" ■ ■ ■'■ . VContinuous . . i 8.04 ? 6.961 ? j 20.30 1 ■! 22.69 9 31.09 ? 34.29 ? Average of 8 courses, ) beans and clover ) 6.76 6.482 11.52 13.36 18. OS 19.90 'Rotation . . . Grain < Continuous . . 12.63 6.45 11.18 6.45 12.19 6.45 10.50 6.45 14.48 7.99 14.23 7.99 14.68 7.99 15.25 7.99 15.12 12.40 16.50 12.40 14.68 12.40 16.43 12.40 iRotn.-l-or-conl. 6.08 4.73 5.74 4.05 6.49 6.24 6.69 7.26 2.72 4.10 2,18 4.03 (Rotation . . . Straw i Continuous . . 2.87 1.27 2.73 1.27 2.76 1.27 2.48 1.27 3.87 1.88 3.76 1.88 3.84 1.88 3.95 1.88 4.94 3.62 5.46 3.62 6.00 3.62 5.31 3.62 vEotn. -For -cont 1.60 1.46 1.49 1.21 1.99 1.87 1.96 2.07 1.32 1.84 1 1.38 1.69 (Rotation . . . Total i Continnons . . 16.40 7.72 13.91 7.72 14.95 7.72 12.982 7.72 18.35 9.87 17.68 9.87 18.52 9.87 19.20 9.87 20.06 16.02 2L96 16.02 ' 19.58 , 21.74 16.02 ' 16.02 C Rotn. -l-or -cont. 7.68 6.19 7.23 5.26 8.48 8.11 8.66 9.33 4.04 6.94 3.56 5.72 J Calculated on average produce of 19 years, 1849-52 and 1856-70- 2 Probably crop too low owing to a dell. 234 THE EOTHAMSTED EXPERIMENTS. Rotation and contin' uous crops. amount of nitrogeu as well, there was, although the supply of phosphoric acid by manure was exactly the same, now about twice as much of it taken up, as a coincident of the greatly increased growth, due partly to the other mineral constituents at the same time added, but especially to the influence of the increased available supply of nitrogen. Still, only a small proportion of the phosphoric acid applied was taken up, considering the recognised importance of its application for turnips, and its undoubted specific effects on their growth as above described. Comparing the amounts of phosphoric acid in the rotation crops with those in the continuous ones, the equally small, or even smaller, amount taken up without manure by the latter, is further confirmation of the incapability of this assumed restorative crop to yield any practical amount of produce without adequate soil supplies. With superphos- phate alone, as also with the mixed manure, the continuous crops took up little more than half as much phosphoric acid as the rotation ones under the assumed fairly parallel con- ditions as to manuring. The deficiency is, however, obviously not due to any deficiency of supply within the soil, but is only a coincident of the less total growth, attributable to a great extent, as has been explained, to the unfavourable mechanical condition of the soil induced by the continuous growth of the crop. Lastly, in regard to the phosphoric acid in the turnip crops, it is to be observed that in all cases much more was accumulated in the edible roots than in the leaves which re- main only for manure again ; indeed, in the case of the most normal crops, those grown in rotation with the full mixed manure, there was five or six times as much accumulated in the roots as in the leaves. The Phosphoric Acid in the Barley Crops. — Looking first to the amounts in the total produce, grain and straw together, and to the portions of the rotation plots from which the pre- vious root-crops had been- removed, it is seen that, without manure, rather more than 13 lb. of phosphoric acid was, on the average, annually removed in the barley crops ; and where superphosphate had previously been applied for the roots, the succeeding barley took up only about 14 lb., that is scarcely any more than without the supply of it ; but where the mixed rfianure, including nitrogen, had been applied for the roots, there was about one-and-a-half time as much, or rather over 21 lb. of phosphoric acid in the succeeding barley Removal of crops. Then, where the root-crops had not been removed from the land, the amounts of phosphoric acid in the suc- ceeding barley crops were, without manure, about 12 lb. per [fnfawur- dble me- chanical condition of soil. Pliosphoric acid in ! root. No man- wre. With superphi phate. Mixed manure. root-crops. EOTATION OF CROPS. 235 acre, with superphosphate about 18 lb., and with the mixed manure nearly 25 lb. In the case of the phosphoric acid, therefore, as in that of the nitrogen, the influence of the manu- ring, and other treatment, of the preceding crop of the course, is clearly reflected in the amounts taken up in the succeeding barley. Comparing the amounts of phosphoric acid in the rotation Rotation barley crops with those in the continuously grown ones, it is ^^"^j' seen that, both without manure and with superphosphate, the rotation crops took up considerably the most phosphoric acid ; and this was the case notwithstanding that the continuously grown crops were annually manured with superphosphate, whilst for those grown in rotation the application had only been for the preceding crop — the turnips. The less assimila- tion in the case of the continuous crops was doubtless due to the diminished total growth, which in its turn was due to the greater exhaustion of the available nitrogen of the soil with the annual growth. Consistently with this view, where the mixed manure supplying an abundance of nitrogen was ap- plied, and the crops, both rotation and continuous, were pretty full averages for the particular soil and the seasons of growth, the amounts of phosphoric acid in the rotation crops where the roots had not been removed were almost identical with those in the continuous crops. Where, however, the rotation roots had been removed, carrying off therefore the whole of the nitrogen that had been taken up, the succeeding barley crops were accordingly not full for the seasons of their growth, and the amounts of phosphoric acid in them were less than in the continuously grown crops. The figures relating to both the rotation and the continuous Phosphoric barley further show, that about six-sevenths of the total phos- '^i^^^ phoric acid of the crops is accumulated in the grain which is straw of supposed to be sold off the farm. There was, indeed, even a *«»■%• somewhat higher proportion where phosphoric acid was sup- plied in the manure. Lastly, as in the cases of the total pro- duce, the dry matter, and the nitrogen, there is much less difference between the amounts of phosphoric acid taken up under the three different conditions as to manuring than in the case of the turnips. That is, the assumed restorative crop Depend- is much more dependent on direct manuring to yield any ^^^/;^ crop at all than is the cereal crop, which is assumed to be manure. benefited by the interpolation of it. The Phosphoric Acid in the Leguminous Crops. — Eeferring to -E/scfa of the third division of Table 63, it is seen that the amounts of '™««'-^- phosphoric acid in the total produce of beans (corn and straw together) were more where superphosphate was supplied than without manure, and more still under the influence of the 236 THE KOTHAMSTED EXPEKIMENTS. Rotation and contin- uous crops. Clover. Beans. Effects of manures. Rotation and contin- uous crops. mixed manure, containing, besides superphosphate, salts of potash, soda, and magnesia, and nitrogen also. But, under all three conditions as to manuring, the continuously grown crops take up much less than those grown in rotation. Whether, however, grown in rotation or continuously, three, four, or more times as much of the phosphoric acid is finally- accumulated in the corn as remains in the straw. In refer- ence to all the results with beans, however, it is to be borne in mind that under none of the conditions were good crops obtained. The clover took up, without manure, little more phosphoric acid than the rotation beans ; but, with superphosphate, the clover took up more than twice as much as the beans ; and with the mixed manure it took up more still, and also more than twice as much as the beans grown under the same conditions. Taking the average of the six crops of beans and two crops of clover grown in the eight courses, there was, both without manure and with superphosphate, much less phosphoric acid taken up than in either, the preceding barley or the succeed- ing wheat ; and even with the mixed manure, which gave the most normal crops, the average amount of phosphoric acid taken up in the beans and clover was less than in either of the two cereals under the same conditions. The Phosphoric Acid in the Wheat Crops. — The bottom di- vision of Table 63 shows that the rotation wheat, as did the rotation barley, took up very much more phosphoric acid without manure than did either of the so-called fallow crops — the turnips or the leguminous crops. With superphosphate, again, both the wheat and barley took up more than either the turnips or the average of the leguminous crops. With the fiiU mixed manure, however, when each of the four descrip- tions of crop grew more normally, the amount of phosphoric acid taken up was more nearly uniform in the four cases ; the barley, however, then yielding more than the wheat, more than the turnips, more than the average of the leguminous crops, but all considerably less than the average of the two years of clover. Comparing the amounts of phosphoric acid in the total produce of the rotation with those in the continuously grown wheat, it is seen that there is, without manure, only about half as much taken up in the continuous as in the rotation crops ; with superphosphate, again, only about half as much in the continuous as in the rotation ; but with the more nor- mal growth, when the full mixed manure was annually applied to the continuously, grown crops, there was, with the fuller produce, proportionally much more phosphoric acid taken up EOTATION OF CROPS. 237 — indeed, on the average, about three-fourths as much in the continuous as in the rotation crops. Lastly, the figures show that by far the larger proportion of PiMspiwrk the total phosphoric acid in the wheat crops is stored up in the ^^^'^^^ grain, which is assumed to be sold off the farm. Thus, without straw of manure more than four-fifths, and with superphospate nearly four-fifths, of the total phosphoric acid of the crops was in the grain. "With the mixed manure, however, with rather larger total amounts taken up than with superphosphate alone, there was comparatively little more stored up in the grain, the ex- cess for the most part remaining in the straw. The larger amount of total phosphoric acid taken up with the mixed manure than with superphosphate, the amount supplied by manure being the same in the two cases, is to be attributed to the coincident supply of other constituents in the mixed manure, inducing greater luxuriance, and with it greater activity of collection. The Amounts of Potash in the notation, and in the Continuous Crops. The results relating to the amount and distribution of potash in the rotation and in the continuous crops are re- corded in Table 64 (p. 239). The Potash in the Boot-crops. — Before referring to the de- Swgairin tails on this point, attention should be recalled to the facts ™<'<-<^'?P*- fully illustrated in other papers^ — that root-crops are essen- tially sugar crops ; that the very characteristic effect which nitrogenous manures exert on their increased growth is mainly represented by a greatly increased production of the non- nitrogenous substance — sugar ; that, however the action is to be explained, it is certain that the presence of potash is an important condition of the formation in plants of carbohy- drates generally ; and that, in the case of root-crops, the pro- duction of the carbohydrate — sugar — is greatly dependent on a liberal available supply of potash. Eeferring to the upper division of the table, and for the Potash in purpose of the first illustrations to the rotation results, it is T""'* "'l'^ ^1 .,, , T - ,, ,, ' leaves of seen that, without manure and very abnormally small crops, turnips. there were only three, four, or five times as much potash in the roots as in the leaves ; with superphosphate, on the other hand, and greatly increased root development, there were eight or nine times as much potash in the roots as in the leaves ; and with the mixed manure (including potash), there were, with the further greatly increased actual amount of roots and of potash in them, seven or eight times as much in the, roots as in the leaves. That is, there was the greatest 238 THE KOTHAMSTED EXPEEIMENTS. accumulation of potash with the greatest accumulation of sugar, Effects of Looking to the actual amounts of potash in the total pro- manures. (j^gg^ roots and leaves together, of the rotation crops, it is seen that, without manure, there was only from 4 to 6 lb. of potash per acre per annum ; but with superphosphate, without potash supply, from 25 to 28 lb. That is, without any supply by manure the plants were able to gather about 20 lb. more potash per acre per annum from the soil itself, by virtue of the greatly increased development of fibrous feeding root under the influence of the phosphatic manure. With the mixed manure, however, containing potash, there was about three times as much of it taken up as with superphoshate alone. But, with the supply of potash there was also a liberal supply of available nitrogen, to which the greatly increased growth is largely to be attributed ; and with the increased luxuriance much more potash was of course required if there were to be a correspond- ingly increased formation of the characteristic non-nitrogenous product of the cultivated root-sugar. Thus, we have — without manure only 4 to 6 lb. of potash taken up, with superphos- phate (without potash) from 25 to 28 lb., and with the mixed manure, supplying besides phosphoric acid both nitrogen and potash, nearly 80 lb. of potash per acre per annum in the crops. Rotation Comparing the amounts of potash in the rotation crops "^"cTO^r ^'i*^ those in the continuously grown ones, it is seen that— without manure, and practically no growth, there was but little difference in the amounts taken up ; with superphos- phate there was little more than half as much taken up in the continuous as in the rotation crops ; whilst with the mixed manure, with full supply of potash, and much larger amounts of it in both the rotation and continuous crops, there wa,s rather less than two-thirds as much in the con- tinuous as in the rotation crops. The deficient amounts in the continuous crops are, however, as in the case of the other constituents, coincidents of the less amounts of produce of the continuous crops ; which, as has been pointed out, were, in the case of the superphosphate plot, due partly to the greater exhaustion of available nitrogen of the surface soil with the continuous growth, but partly also to the unfavour- Vnfcmcmr- able mechanical condition of the soil induced by such th^al growth ; and this was probably the chief cause of the deficient . condition produce in the case of the mixed manure crops also. ^fsoil. rpj^^ Potash in the Barley Crops. — The second division of Table 64 records the results on this point. ■ In the case of the turnips it was found that much more potash was .aeeumulated in the roots than in the leaves • and EOTATION OF CROPS. 239 TABLE 64. — Experiments on tece Kotahon op — Boots, Bablet, Ciover (oe Beaius), ob Fallow, axd Wheat ; nr Agdell Fihu), Rothausted. 8 courses, 32 years, 1852-1883. AVERAGE AMOUNTS OF POTASH PER ACRE PER ASXUII IX THE ROTATION CROPS, COMPARED WITH THOSE IN THE CROPS GROWN CONTINTTODSLT. UnmanTired. SapeTphosphate. Mixed roineial and nitro- genotis manure. Roots carted. Roots fed. Roots carted. Roots fed. Roots carted. Roots fed. JB^ns I'"'- clover p^. JB^ l""- elovet Fal- i^^"^ 1"- Closer j^,. JBeans 1»- !cl°ve, Fal- B^»°' Fal- l"-- cl^veri 1°^- Beans or clover SWEDISH TUENIPa rBotation . . . Roots - Continaoas 1 . . lb. 5.00 3.48 lb. 3.04 3.48 lb. 4.40 3.48 lb. 2.82 3.48 lb. 22.49 12.08 lb. 1 21.67 1 12.08 ' lb. 25.05 12.08 lb. 24.86 12.08 lb. 66.62 39.51 lb. 67.99 39.51 1 lb. 1 lb. 72.48 1 68.53 39.51 39.51 v Botn. -^or - cont. 1.52 -0.44 0.92 -0.66 10.41 9.59 1 12.97 12.78 27.11 2S.4S 32.97 1 29.02 9. 89 , 10.25 9.98 i 9.98 Botalaon . LeaTes Contmoons i . . 1.07 0.94 0.95 0.94 1.04 0.94 0.93 0.94 2.60 2.38 2.96 . 2.38 ' 2.77 3.31 2.88 2.38 8.66 1D.S2 9.98 ; 9.98 vBotn--{-or-cont. 0.13 0.01 0.10 1 -0.01 0.22 0.58 ; 0.39 1 0.93 -1.32 0.34 -0.091 0.27 82.37 i 78.78 49.49 : 49.49 32. 8S 29.29 rBotation . Total J Continnoas 1 . . 6.07 4.42 3.99 4.42 5.44 1 3.755 4.42 4.42 25.09 14.46 24.63 1 14.46 : 27.82 1 28.17 14.46 ! 14.46 75.28 49.49 78.31 49.49 vBotn. -f or - cont. L65 1-0.43 1.02 -0.67 10.63 ! 10.17 I 13.36 1 13.71 25.79 2S.62 BABLET. (Rotation . . . Grain -j Continnoas . . 8.13 5.03 8.38 li 7.97 5.03 1 5.03 7.15 5.03 8.09 6.59 7.85 1 6.59 ! 10.23 6.59 10.65 6.59 12.33 12.52 14.32 14.32 14.14 14.04 14.32 14.32 V Botn. -{-or - cont. 3.10 3.35 |: 2.94 2.12 1.50 1.26 1 3.64 1 4.06 -1.99 -1.80 -0.18 -0.28 TBotation . . . Straw -c Continnons . . 10.83 6.45 11.81 [ 6.45 10.52 6.45 10.09 6.45 9.32 7.03 9.50 il 12.10 ! 12.54 7.03 ,1 7.03 7.03 18.41 18.97 21.00 21.00 • 23.48 23.31 21.00 21.00 lBotn.+or-cont. 4.3S 5.36 4.07 3.64 2.29 2.47 5.07 5.51 - 2.59 i- 2.03' 2.48 2.31 (Rotation . Total -J Continnons . . V. Botn. -f or - cont. 18.96 11.48 20.19 11.48 18.49 11.48 17.242 11.48 17.41 13.62 17.35 13.62 22.33 23.19 13.62 13.62 30.74 1 31.49 1 35.32 1 35.32 | 37.62 37.35 35.32 ' 35.32 7.48 8.71 7.01 6.76 3.79 3.73 8.71 9.57 -4.58-3.83 2.30 2.03 BEANS (6 COUBSES). CLOVER (2 COURSES), OB FALLOW. (Rotation . Com - Continnons . . ( Botn.-1-or-cont. 7.26 2.98 i 7.23 2.98 7.35 3.46 8.79 3.46 15.20 8.94 17.25 8.94 4.28 4.25 3.89 1 6.33 6.26 8.31 (Botation . . . Straw • Continnons . . I. Botn. -For -cont 2.87 1.54 , 2.87 1.54 3.47 1.82 4.01 1.82 6.96 4.33 7.21 4.33 1.33 i 1.33 1.65 2.19 2.63 2.88 (Botation . . . Total -I Continnons . . 10.13 4.52 10.10 4.52 10.82 5.28 i 12.80 5.28 22.16 13.27 24.46 13.27 V Botn. -for - cont. 5.61 1 6.58 6.54 1 7.52 8.89 11.19 m«„«_ f Botation . . . *^°^^i Continnons . . 34.18 ( 29.672 ? 57.63 9 65.48 123.12 ? 132.62 ? Average of 8 conises, '■ beans and clover / 16.14 14.992 22.52 i 25.96 47.40 5L50 (Rotation . Grain - Continnons . . 8.65 I 8.08 4.45 1 4.45 8.42 4.45 7.26 4.45 9.55 5.27 9.39 ' 5.27 1 9.69 5.27 10.06 5.27 9.90 : 10.82 8.12 1 8.12 9.55 10.78 8.12 j 8.12 I Kotn. +0T- cont. 4. 20 3.63 3.97 2.81 4.28 4.12 4.42 4.79 1.78 2.70 i.43| 2.66 (Botation . . 19.12 17.94 Straw J Continnons . . , 8.49 8.49 ' lBotn.-For-cont. ! 10.63 9.45 18.30 8.49 16.31 8.49 20.25 10.00 19.14 10.00 j 20.45 10.00 20.21 10.00 25.85 27.47 18.81 1 18.81 26.21 27.12 18.81 18.81 9.81 7.82 10.25 9.14 10.45 10.21 7.04 8.66 7.40 8.31 (Rotation . . . Total -j Continnons . . 27.77 26.02 12.94 12.94 26.72 12.94 23.572 12.94 29.80 15.27 28.53 1 15.27 ■ 30.14 15.27 30.27 15.27 35.75 26.93 88.29 26.93 35.76 ; 37.90 j 26.93 26.93 ^Rotn.-ror - cont. 14.83 13.08 13.78 10.63 14.53 13.26 14.87 15.00 8.82 11.36 1 8.83 10.97 1 Calcnlated on average prodnce of 19 years, 1849-52 and 1856-70. ' Probably crop too low owing to a delL 240 THE EOTHAMSTED EXPERIMENTS. Potash in grain and straw of harley. Effects of manwes. Rotation and contin- uous crops. this fact was assumed to be connected with the greater amount of the carbohydrate — sugar-^in the roots than in the leaves. The results relating to the barley show, however, that there was in every case more, and in some much more, potash in the straw than in the grain. On this point it is to be observed, not only that the root-crop is taken up when still in the vegetative stage, and its contents are still in the con- dition of reserve or migratory material, whilst in the case of the cereal the crop is ripened, and its constituents are, there- fore, more fixed. Further, whilst in the turnip-crop there was several times as much dry substance in the roots as in the leaves, in the barley there was even more dry organic substance in the straw than in the grain. Again, in both crops, by far the larger proportion of the dry substance con- sists of carbodyrates — in the one chiefly sugar, and in the other almost exclusively starch and cellulose — the latter mak- ing up by far the greater portion of the dry substance of the straw. It is obviously quite consistent that under these circumstances there should be more of the total potash of the barley crop accumulated in the straw than in the grain. It must at the same time be observed that, whilst the potash in the grain is comparatively fixed and bears a fairly uniform relation to the amount of dry substance, the quantity which remains in the straw is subject to great variation in propor- tion to the dry matter, according to the variation in the supply of it within the soil — a great excess above. the amount in other cases being sometimes found in the straw. Indeed, the figures show a considerably greater proportion of the total potash of the crop accumulated in the straw where there was a liberal supply of it in manure. Eeferring to the amounts of potash taken up in the rota- tion barley crops on the different plots, according to the manuring or other treaitment, the figures show that there was not much difference between the amounts without man- ure and with superphosphate alone. There was,- however, distinctly more taken up on the portions of the superphos- phate plot where the roots had not been removed than where they were ; and where, therefore, there was conservation for the succeeding crop. With the mixed manure, however, with its supply of potash as well as of phosphoric acid and nitrogen, the amount of potash in the crops is greatly in- creased, the increase corresponding closely with the increased amount of produce. Lastly in regard to the potash, whilst without manure and with superphosphate alone the rotation barley has gathered much more than the continuously grown, with the mixed manure and full supply of all constituents, the amounts of ROTATION OF CROPS. 241 potash taken up were, as were those of nitrogen and phos- phoric acid, nearly the same in the rotation and the continu- ous crops where in rotation the preceding roots had not heen removed ; but where they had been removed, the amounts of potash ia the succeeding barley were less, as were the crops themselves. Hie Potash in the Zeguminous Crops. — Of aU. the mineral constituents of the crops, perhaps potash and lime are the most generally recognised as having some distinctive effects when applied as manure for leguminous crops. We have now to refer to the records relating to the potash in these crops, as given in the third division of Table 64. The figures show that, in the case of the beans, unlike that Potash in of the cereals, there is much more potash iu the corn than iu '^^'^ the straw ; indeed, more than twice as much of the potash of Ugwnea. the crops was accumulated in the corn as in the straw ; indi- cating, therefore, a special requirement of it for the formation of the final and most fixed product of the plant — ^the seed. Looking to the amounts of potash per acre in the total pro- EffecU of dnce, corn and straw together, of the rotation beans, it is seen '»""'«■«*■ that they take up very little more under the influence of the superphosphate than without manure ; the quantities averag- ing about 10 lb. per acre without manure, and scarcely 12 lb. with superphosphate. With the mixed manure, however, directly supplying potash for the previous root-crop, the amounts of it taken up were, in the one case 22.16, and in the other 24.46 lb., or about twice as much as with the super- phosphate alone. The influence of the previous supply of potash on the amounts of it taken up in the beans was, in fact, much greater than was that of the supply of phosphoric acid on the amounts of it taken up. But, as in the case of the phosphoric acid, so also in that of Rotatim the potash, the continuously grown beans took up only about ^^'^"' half as much as those grown in rotation ; proportionally more, however, where it had been supplied than where it had not. It will be remembered that, when discussing the amounts of produce of the bean crops, attention was called to the fact that throughout the experiments a really good agricultural crop was scarcely ever obtained ; and this of course must be taken into account when considering the amounts of the several constituents of the crops. Comparing the amounts of potash stored up in the rotation Glover and clover with those in the rotation beans, it is seen that, even ^^^'"^' without manure and with very small produce, the clover, with its greater root- range and longer period of growth, gathered up about three times as much potash as the beans — about 30 lb. against only about 10 lb. in the beans. VOL. vn. Q 242 THE EOTHAMSTED EXPERIMENTS. Potash mcmwres for Ugu- mmous Potash in gram cmd stroAV of With, superphosphate alone, whilst the beau crops contained only 10.82 and 12.80 lb. of potash, th6 clover contained 57.63 and 65.48 lb. That is, under the influence of the phosphatic manure, probably partly on the plant and partly on the soil, the clover had accumulated in the removed crop five or six times as much potash as the beans, from the soil itself; whilst, of the phosphoric acid itself, little more than twice as much was taken up in the clover as in the beans under the influence of the superphosphate without potash. It would thus appear that the beneficial effects of the phosphatic manure on the clover were largely connected with the increased capability of the plant to take up more potash. With the mixed manure, supplying a large amount of potash, the amount of it found in the clover crops was, how- ever, much greater still. Both in the beans and in the clover the amount of potash in the crops grown under the influence of the direct supply of it was about twice as much as those grown with superphosphate without potash. But whilst,, under the influence of the supply of it, the shorter-lived, more meagrely rooting, and less successfully grown bean crops stored up only 22.16 and 24.46 lb. of potash, the clover crops contained in one case 123.12 lb., and in the other 132.62 lb. The very much larger proportion of the totaj. potash of the bean crops which is found in the corn than in the straw would seem to indicate its greater importance in connection with the maturing than with the merely vegetative and accumulating tendencies of growth ; yet the increased amount of it taken up by the beans coincidently with increased .growth, and the much larger amounts of it in the clover with its much greater amounts of growth and produce, and harvested as it is in the unripened condition, are on the other hand indications of a direct connection between potash supply and the luxuriance of growth or vegetative activity of these leguminous crops. Indeed, as already referred to, potash manures are well known to be frequently beneficial to such crops. To these points further reference will be made presently, when calling atten- tion to the amount of lime taken up by leguminous crops. The Potash in the Wheat Crops. — The results on this point are given in the bottom division of Table 64. It has been seen that by far the larger proportion, both of the nitrogen and of the phosphoric acid of the wheat crops, was accumulated in the grain. But the figures relating to the potash show that of it there was very much more in the straw than in the grain. There was also much more, but not in so great a degree more, in the straw than in the grain of the other cereal — the barley, It has been pointed out that potash is at any rate essentially connected with the formatioa ROTATION OF CEOPS. 243 of the carbohydrates. Consistently with this it was found that by far the larger proportion of the potash of the turnip crop was in the roots, where was the great accumulation of sugar. Again, of the total potash of the barley crop, the larger proportion was found in the straw where there was the greatest accumulation of carbohydrate — as cellulose ; and now, in the wheat, with a larger proportion of straw to grain, and a proportionally larger amount of the total carbohydrates accumulated in the straw, we have in it a still larger pro- portion of the total potash of the crop. It is, however, to be borne in mind, as has been pointed out, that the straw of both barley and wheat frequently contains, besides the mineral constituents actually essential for the organic formations and changes, a more or less surplus amount taken up as the result of liberal supply, and retained by the plant. Although there is doubtless clear foundation in fact for the Functions conclusion that the role of phosphoric acid is more in con- ^^ ^^j. nection with the formation and activity of the nitrogenous phone bodies, and that of the potash with those of the non-nitro- "'^• genous compounds, yet it is obvious that in such a view we have only a partial and imperfect explanation of the function of these mineral constituents. Thus, in the case of the beans there was, consistently enough, much more phosphoric acid in the corn than in the straw — that is, the more where there was the more nitrogen ; but there was also by far the larger proportion of the potash accumulated in the corn, although the greater part of the dry matter of the crop, and with this of its carbohydrates, was iu the straw. Indeed, although the leguminous crops are pre - eminently highly nitrogenous, a liberal supply of potash is essential for their luxuriance; whilst they contain a higher proportion of it in their dry substance than do the cereals, with their higher proportion of carbohydrates. Eeference to the figures shows that the application of super- phosphate, without potash, enabled the wheat plant, whether grown in rotation or continuously, to take up an increased, but not a much increased, amount of potash, compared with that in the unmanured crops ; and that the direct application of it increased the assimilation of it still further, though the increased amount of it stored up represented only a small proportion of that supplied in the manure. Without manure, the rotation wheat crops contained an Rotation average of about 27 lb. of potash, but the continuously grown ""^ contin- ones scarcely 13 lb., or only about half as much. With super- and the phosphate, without potash, the rotation crops gave an average ?^^* »/ of nearly 30 lb., and the continuously grown ones little more "^""''*'' |;han 15 lb.; or, again, only about half as much. That is, 244 THE EOTHAMSTED EXPERIMENTS. Other mm- eral con- stituents. when the growing crops had to rely for their potash exclu- sively on the stores of the soil itself, the rotation crops took up about twice as much as the continuous. Lastly, with the mixed manure supplying potash, the rotation wheat crops gathered nearly 36 lb. after fallow, but about 38 lb. after the leguminous crops ; whilst the continuously grown ones yielded an average of only about 27 lb. That is, although in the case of the rotation wheat crops three other crops had been grown since the application of the manure, they took up more potash than the continuously grown ones for which potash was annually supplied. So much for the results relating to the amounts of the two important and typical mineral constituents — phosphoric acid and potash — taken up by the different crops when grown, respectively, in rotation and continuously, under different con- ditions as to manuring, and other treatment. Similar results relating to other mineral constituents of the crops have been got out, and the discussion of some of them brings to view points of considerable interest, but neither time nor space will admit of their consideration here. It must suffice to refer briefly to the amounts of lime taken up by the legu- minous crops under different conditions ; a point which has an interesting relation to the results as to the potash taken up by those crops, and to the questions which arose in the discussion of them. aym and Strom of Rqtation arid contin- uous crops. The Amounts of Lime in the Rotation, and in the Continuous Legv/minous Grcps. The following Table (65) gives, for the leguminous crops alone, the amounts of lime in the rotation and in the con- tinuous crops, in the same form in which the phosphoric acid and potash have been given for each of the four crops of the rotation. Very different from what was found to be the case with the potash, it is seen that in the rotation bean-crops a very small proportion of the total amount of lime is accumulated in the corn ; ten, twelve, or more times as much being found in the straw. Then, the amounts of lime in the total crops were — without manure between 15 and 16 lb. ; with super- phosphate, which of course supplied some lime, the quantity was raised to 18.68 and 20.71 lb. ; and with the mixed manure, also supplying the same amount of lime in its super- phosphate, it was further raised to 26.57 and 27.71 lb. It is further seen, that the continuously grown beans contained — ■ in corn, straw, and total produce — in some cases only aboutj ROTATION OF CROPS. 245 and in others not much more than, half as much lime as the rotation ones. It is remarkable, however, that whilst without manure the Effect of rotation hean-crops contained only from 15 to 16 lb. of lime, '"<™«''^' the clover contained 67.84 and 59.10 lb. ; with superphosphate the beans gave 18.68 and 20.71 lb., and the clover 158.62 and 184.52 lb. or about eight times as much as the beans; and lastly, with the mixed manure, the bean-crops contained 26.57 and 27.71 lb., and the clover 181.75 and 195.14 lb. of lime, or about seven times as much as the beans. TABLE 65. — Average amounts op Lime per acre per annum in the Rotation, and in the continuously grown, Leguminous Crops. Unmannxed. Superphosphate. Mixed mineral and nitrogenous manure. Roots carted- Roots fed. Roots carted. Roots fed. Eloots carted. 1 Boots fed. Pal- low. Beans or clover Fal- low. Beans or clover Fal- low. Beans or clover Fal- low. Beans or clover Fal- low. 2. « .ill Fal- low. Beans or clover BEANS (6 COURSES), CLOVER (2 COURSES), OR FALLOW. /Rotation . . . Com •< Cont™""™ lb. 1.16 0.47 lb. 1.14 0.47 lb. 1.10 0.52 lb. 1.32 0.62 lb. 2.10 1.24 lb. 2.38 1.24 V Rotn. -F or - con t. 0.6S ' 0.67 0.68 0.80 0.86 1.14 (•Rotation . . . Straw • Continuous 14.61 7.85 j 14.66 ' 7.85 17.68 9.36 19.39 9.36 24.47 15.08 25.33 16.08 .Rotn.-)-or-cont. 6.75 6.81 8.22 10.03 9.39 10.25 Rotation . . . Total ■< Continuous . . 15.76 8.32 15.80 8.32 18.68 9.88 20.71 9.88 26.57 16.32 27.71 16.32 Botn. -fop - cont. 7.44 7.48 8.80 10.83 10.25 11.39 Total J Rotation . . . '■°^^ t Continuous . . 67.84 ? 59.101 168.62 1 184.52 ? 181.76 ? 196.14 Average of 8 courses, ) beans and clover j" * 28.78 1 26.631 53.67 61.66 j 65.36 69.57 1 Probably crop too low owing to a dell- An increased amount of lime is, therefore, even more Fwnotion directly connected with increased luxuriance and increased "•C^'?**" production, than is an increased amount of potash taken up. grmath. Then, again, the increased amount of potash was apparently more or less directly connected with tendency to maturation or seed-formation ; but the lime is found chiefly in the straw of the beans, and to be enormously increased in amount in the clover, which does not ripen, but is cut whilst still in the vegetative condition. The indication is, therefore, that the lime is, both actually and as compared with the potash, much 246 THE KOTHAMSTED EXPERIMENTS. Propor- tions of lime, potash, amd cwr- ionic acid in the ash of plants, and thei/r relation to the assimil- ation of more directly connected with the accumulative or vegetative, as distinguished from the maturing processes of the plant. Certain it is, at any rate, that a largely increased accumula- tion of lime is a coincident of increased luxuriance in both crops ; and it is especially so in the case of the crop the amount of which depends on the extension of the vegetative stages of development, and the production of a large amount of crude or unripened vegetable substance. Thus, then, the actual and relative importance of potash and lime in the growth of the highly nitrogenous leguminous crops is clearly illustrated in the acreage amounts given, of potash in the third division of Table 64, and of lime in Table 65. But the study of the percentage composition of the ashes of the crops, and especially of both the percentage composition of the ashes, and the amount of the constituents per acre, in the bean plant taken at different stages of its growth, and of somewhat similar results relating to the first, second, and third crops of clover, affords further c6nfirmation of the conclusions which have been drawn from the results already considered. It will be impossible to go into any detail here in regard to these further results, and it must sufSce to state very briefly their general indications. The bean-plant ash analyses showed that, on the average, about 75 per cent, and at the time of pod formation nearly 80 per cent, of the total ash consisted of lime, potash, and carbonic acid. Compared with these results, those relating to the more highly nitrogenous clover, which is not allowed to ripen, but is cut when it reaches the blooming stage, so inducing re -growth and extension of the more] specially vegetative stages, show that from about 80 to about 84 per cent of the total ash consisted of lime, potash, and carbonic acid. But whilst in the a&h of the ripened corn-yielding bean-crop there was about one and a-half time as much potash as lime, in that of the merely vegetating unripened clover there was twice or even three times as much lime as potash. Further, in the ash of the first and third crops of clover, which would be the most succulent and unripe, the relative excess of lime., over potash is much greater than in that of the second crop, which develops at the period of the season when the seed-forming tendency is much the greater. Again, in the clover ashes there was about one and a-half time as much carbonic acid as in the ash of the ripened bean plant. It is thus further illustrated that a peculiarity of the composition of these pre-eminently nitrogen-assimilating ele- ments of rotation is, that their ashes consist chiefly of lime, potash, and carbonic acid; that the potash predominates in the ripened and less nitrogen-yielding bean-crop ; and that EOTATION OF CROPS. 247 the lime and carbonic acid predominate in the continuously vegetating and much more largely nitrogen -accumulating clover. Referring to the probable or possible significance of these • facts, it is obvious that, so far as the nitrogen of the plant is taken up as nitrate of a fixed base, that base, so far as it does not pass back into the roots, -will remain in the above-ground parts of the plant, most probably in combination with an organic acid, which will be converted into carbonic acid in the incineration, and be found as such in the ash, if not ex- pelled by an excess of fixed acid, or by silica. In the case of the cereals of the rotation, it is probable that Nitrogen most, if not all, of their comparatively small amount of nitro- ^**^^ gen is taken up as nitrate. Potash is by far the predominat- as nitrate. ing base in the ash of the grain, straw, and total produce ; lime is in much less amount, both actually and in equival- ency; and magnesia is in less amount still, though it is a characteristic constituent of the grain-ashes. There is prac- tically no carbonic acid in either wheat or barley grain-ash, and but little in the straw-ash ; and if there have been organic acid salts formed with the base of the nitrate, the carbonic acid may have been expelled in the incineration, by the excess of fixed acid in the grain-ash, or by silica in the straw- ash. Taking the produce by the mixed manure as the most normal, the root-crops of the rotation come next in amount of nitrogen assimilated over a given area. Potash and lime are the predominating bases. There is much more potash than lime in the more definite product — the root; but the pro- portion of lime to potash is much greater in the leaf-ash, as would be expected if the nitrogen had been taken up chiefly as calcium nitrate, and the nitric acid subjected to decomposition in the leaves. Lastly come the Legumiuosae, with their much higher amounts of nitrogen assimilated. These plants also doubtless derive at any rate much nitrogen from nitrates in the soil and' subsoil ; and it has been shown that their great assimilation of nitrogen is associated with very large amounts of lime and carbonic acid in their ashes. Eeferring to the results with the rotation beans grown by the mixed manure, calculation shows that, taking the total crop, corn and straw together, it contained very much less lime than would be required if the whole of its nitrogen had been taken up as calcium nitrate ; so that either part of the nitrogen must have been taken up as nitrate of some other base, .or in some quite difi'erent state of combination, or as free nitrogen ; or some of the lime must- have been elimid- 248 THE KOTHAMSTED EXPEEIMENTS, ated from the above-ground parts of the plant into the roots, and possibly some of it passed from them into the soil. Again, the amount of carbonic acid found in the ashes of the crop for 100 of nitrogen in it would require about one and a- half time as much lime as was found in association with it ; indicating the probability that part of the nitrogen taken up as nitric acid was as the nitrate of some other base — potash, and possibly to some extent soda also. Turning to the results with the rotation clover grown by the mixed manure, calculation shows that in the case of this continuously vegetating, unripened, and much higher nitrogen- yielding crop, there was very much more of both lime and carbonic acid in the ash for 100 of nitrogen assimilated than in the total bean-crop. If, however, the whole of the nitrogen of the clover crops had been taken up as calcium nitrate, it would have required nearly twice as much lime as the amount found, provided the whole of it remained ; nor would the amounts of potash and soda found suffice to make up the balance. Again, the amount of carbonic acid found is little more than two-thirds as much as would be required to repre- sent organic acid equivalent to the amount of nitric acid subjected to change. Either, therefore, fixed base, partly in combination with organic acid, must have been eliminated from the above-ground parts of the plant, and passed into the roots, and possibly into the soil, or a good deal of the nitro- gen must have been taken up in some other form than as nitrate ; possibly in part as organic nitrogen taken up from the soil by the agency of the acid sap ; or, in part as free nitrogen, probably brought into combination under the in- fluence of micro-organisms within the nodules found on the roots of leguminous plants, the resulting compound being .either directly available as a source of nitrogen to the host, or it may be so only after it has itself suffered change, hum as a However this may be, considering the very characteristic Z^kamd. •differences in the mineral composition of the different crops of rotation according to the amounts of nitrogen they assimi- late, the fact that undoubtedly the highly nitrogenous Legum- inosse do take up at any rate a large proportion of their nitrogen as nitrate, and that the greater the amount of nitro- gen assimilated the more is the ash characterised by contain- ing fixed base, and especially lime, in combination with carbonic acid, it seems very probable, if not indeed established, that the office of the lime, and partly that of the other bases also, is that of carriers of nitric acid ; which, when trans- formed, and the nitrogen assimilated, leaves the base as a residue, presumably in combination with organic acid. Fur- ther, the power of these plants to assimilate so very much KOTATION OF CEOPS. 249 more nitrogen over a given area than the other crops may, at any rate in part, be dependent on their being able, by virtue of the range and character of their roots, to gather up more nitrogen in the form supposed than the plants with which they are alternated. Such a view does not, however, exclude the supposition that some of their nitrogen is derived in other ways, as above referred to. In connection with the foregoing results of direct experi- mental investigation into the mineral composition of legum- inous crops, it may be observed — that clover at any rate grows AppUca- more favourably on land that has recently been chalked or ^^^nd limed ; that chalking or liming of the mixed herbage of grass potash far land also favours the development of the leguminous herbage ; ^^3^™*^ and that the application of gypsum to clover has been found very effective on some lands, especially in America, though it has not proved to be at all generally useful when it has been so applied in this country. Indeed, the direct application of potash as manure is certainly more frequent, and is more generally recognised as effective for leguminous crops, than is that of lime, notwithstanding its obvious importance, and its great influence on the luxuriance of growth of such crops. This may perhaps be partly explained by the fact that, in many, if not in most, soils, the immediately available supply of potash within the root-range of the plant will probably be sooner exhausted than will that of lime. Stjmmahy and General Conclusions. It remains, in conclusion, very briefly to summarise the facts brought out in this extended inquiry on the subject of rotation, and to endeavour to draw from them an explanation of the benefits arising from the practice of it. At the commencement it was pointed out, that although Vanatiims many different rotations are adopted, they may for the most *^ '^°^' part be considered as little more than local adaptations of the system of alternating root-crops and leguminous crops with the cereals. Thus, there are rotations of five, six, seven, or more years. But these variations are, in most cases, only adaptations of the principle to variations of soil, altitude, aspect, climate, markets, and other local conditions ; and they consist chiefly in the variation in the description of the root- crop, and perhaps the introduction of potatoes ; in growing a different cereal, or it may be more than one cereal consecu- tively ; in the growth of some other leguminous crop than clover ; or the intermixture with the clover of grass seeds ; and perhaps the extension of the period allotted to this ele- ment of the rotation to two or more years. 250 THE EOTHAMSTED EXPERIMENTS. Removal or home con- smnption of crops. Excep- tional rota- tions. Self-sup- porting rotations. Mineral constitu- rotation crops,. It is true, also, that, under any specific rotation, there may be deviations from the plan of retaining the whole of the root- crop, the straw of the grain crops, and the leguminous fodder- crops, on the farm, for the production of meat or milk, and, coincidently, for that of manure to be returned to the land. But it is also true that, when under the influence of special local, or other demand— proximity to towns, easy railway or other communication, and so on — the products which would otherwise be retained on the farm are exported from it, the import of town or other manures is generally an essential condition of such practice. Indeed, this system of free sale very frequently involves full compensation by purchased manures of some kind. In our own country, such deviations from the practice of merely selling grain and meat have been much developed in recent years ; and they will doubtless con- tinue to increase under the altered conditions of our agriculture, dependent on very large imports of grain, increasing imports of meat and other products of feeding, and very large imports of cattle-food and other agricultural produce. Already much more attention is being devoted to dairy products, not only on grass farms, but on those that are mainly arable ; and there will doubtless be some, but probably by no means so great an extension as some suppose, in the production of other smaller articles required by town populations. It is further true-, though the remark applies in a very limited degree to our own country, that there are other devia- tions which have more the character of exceptions to the general rule of rotation, such as the introduction of flax, hemp, tobacco, or other so-called industrial crops. But, in these cases, as with potatoes, the growth involves special expenditure for manure instead of conservation of it. Indeed, the induce- ment is the high price of the product, rather than the main- tenance, or the improvement, of the condition of the land for future crops. Still, as such deviations from regular rotation practice as have been referred to, do, as has been said, generally involve more or less, and frequently full, compensation by manure from external sources, we may, in endeavouring to explain the benefits which accrue from the practice of rotation, confine attention, for the purposes of illustration, to what may be called the self-supporting system, and to the simple four-course one which has been selected for investigation at Eothamsted, and from the results relating to which the illustrations which have been brought forward have been drawn. : It will be well first briefly to refer to the evidence relating to some of the more important mineral constituents found in the different crops of the four-course rotation. EOTATION OF CHOPS. 251 Oi phosphoric acid, the cereal crops take up as much as, or Phosphoric more than, any of the other crops of the rotation, excepting '"'*^" clover ; and the greater portion of what they take up is lost to the farm in the saleable product — the grain. The remainder, that in the straw, as well as that in the roots and the legu- minous crops, is supposed to be retained on the farm, except- ing the small amount exported in meat and milk. Oi potash, each of the crops takes up very much more than Potash. . of phosphoric acid. But much less potash than phosphoric acid is exported in the cereal grains, much more being re- tained in the straw ; whilst the other products of the rotation — the roots and the Leguminosae — which are also supposed to be retained on the farm, contain very much more potash than the cereals, and comparatively little of it is exported in meat and milk. The general result is, that the whole of the crops of rotation take up very much more of potash than of phos- phoric acid, whilst probably even less of it is eventually lost to the land. Of lime, very little is taken up by the cereal crops, and by Ume. the roots much less than of potash ; more by the Leguminosae than by the other crops, and, by the clover especially, some- times much more than by all the other crops of the rotation put together. Of the lime of the crops, however, very little goes in the saleable products of the farm under the conditions supposed of a self-supporting rotation. There is, however, frequently a considerable loss of lime in land- drainage. Although the facts relating to other mineral constituents of the crops are not without significance, reference can be made here to .only one other of these constituents — namely, the silica. The interpolated crops of rotation — the roots and the Legu- silica. minosse — take up scarcely any silica ; but the cereals take up a very large amount of it. Indeed, the large amount of silica taken up by these crops when grown under ordinary condi- tions, is as characteristic a chemical phenomenon of rotation as is the very large amount of lime taken up by clover and other Leguminosae. Very little silica, however, is lost to the land in the assumed saleable products. Thus, then, although diflferent, and sometimes very large. Loss and amounts of these typical mineral constituents are taken up by »•«*»■«■ 0/ the various crops constituting the rotation, there is no mate- constitu- rial export of any in the saleable products, excepting of phos- «»'*• phoric acid and of potash ; and, so far at least as phosphoric acid is concerned, experience has shown that it may be ad- vantageously supplied in purchased manures. But, although the eventual loss to the land of mineral con- stituents is, in a self-supporting rotation, comparatively so 252 THE EOTHAMSTED EXPEKIMENTS. Importance small, the very fact that the different crops require for their fJtt^tT^ growth, not only very different amounts of individual consti- enu. tuents, but require these to be available within tne sou m very different conditions, both of combination and of distri- bution, points to the conclusion that, in any explanation of the benefits of an alternation of crops, the position, and the rdle, of the mineral constituents must not be overlooked ; and the less can it be so, when their connection with the very important element — the nitrogen of the crops — is considered. Nitrogen As to the nitrogen : — It has been seen that, although very ^ms"'*^^ characteristically benefited by nitrogenous manures, the cereal crops take up and retain much less nitrogen than any of the crops alternated with them. In fact, the root -crops may contain two, or more, times as much nitrogen as either of the cereals, and the leguminous crop, especially the clover, much more than the root-crops. The greater part of the nitrogen of the cereals is, however, sold off the farm ; but perhaps not more than 10 or 15 per cent of that of either the root-crop, or the clover, or other forage leguminous crop, is sold off in animal increase or milk. Thus, most of the nitrogen of the straw of the cereals, and a very large proportion of that of the much more highly nitrogen-yielding crops, returns to the land as manure, for the benefit of future cereals and other crops. Indeed, it is, as a rule, only a comparatively small proportion of the very much increased amount of nitrogen obtained in rotation compared with that in continuous cereal -cropping (chiefly due to the interpolated crops), that is lost to the land in the saleable products. Assimila- -A.S to the source of the nitrogen of the so-called " restora- tion, of tive crops," it has been shown that certainly in the case of roofaf^" ^ the roots it is not, as has sometimes been assumed, that such plants take up nitrogen from the air by virtue of their ex- tended leaf- surface. Both common experience and direct experiment demonstrate that they are as dependent as any crop that is grown on available nitrogen within the soil, which is generally supplied by the direct application of nitrogenous manures — natural or artificial. Under such conditions of supply, however, the root-crops, so to speak, gross feeders as they are, and distributing a very large amount of fibrous feed- ing root within the soil, avail themselves of a much greater quantity of the nitrogen supplied than the cereals would do under similar circumstances ; this result being partly due to their period of accumulation and growth extending even months after the period of collection by the ripening cereals has terminated, and at the season when nitrification within the soil is the most active, and the accumulation of nitrates in it is the greatest. Lastly, full supply of both mineral con- ROTATION OF CEOPS. 253 stituents and nitrogen being at command, these crops assimi- late a very large amount of carbon from the atmosphere, and produce, besides nitrogenous food products, a very large amount of the carbohydrate — sugar — as respiratory and fat- forming food for the live-stock of the farm. Yerj' much the same may be said of maize as grown as a AssimUa- fodder-crop in America, as of roots as grown in rotation in ^^^^^ j^ other countries. Thus, there can be no doubt that the maize maize, derives its nitrogen from the soil, collecting some time longer than wheat, and availing itself of the nitrates formed after the collection by the wheat has ceased. But, so far as the product is consumed on the farm, much of its nitrogen is recovered in the manure — the more when the food is used for growing or fattening stock, and the less when for the produc- tion of milk. The still more highly nitrogenous leguminous crops, on Legvmin- the other hand, although not characteristically benefited by "^l^^ nitrogenous manures, nevertheless contribute much more supply of nitrogen to the total produce of the rotation than any of the ?^™^|" *" other crops comprised in it. It is also certain that, at any rate a large proportion of the nitrogen of these crops is obtained from the soil and subsoil ; though recent investiga- tions have proved that some of their nitrogen, and sometimes much of it, may be derived indirectly from the free nitrogen of the atmosphere, brought into combination under the influ- ence of micro-organisms within the nodules on the roots of the plants. It is the leguminous fodder crops, and among them espe- cially clover, which has a much more extended period of growth, and much more extended range of collection within the soil and subsoil, than any of the other crops of the rota- tion, that yield in their produce the largest amount of nitro- gen per acre. Much of this is doubtless taken up as nitrate ; yet, the direct application of nitrate of soda has compara- mtrate of tively little beneficial influence on their growth. The nitric *?** "•'^ acid is probably taken up chiefly as nitrate of lime, but pro- bably as nitrate of potash also, and it is not without signifi- cance that the high nitrogen-yielding clover takes up, or at least retains, very little soda. The general result is, then, that although undoubtedly the clover takes up a good deal of its nitrogen as nitrate, this would seem to be derived from Sowces of accumulations within the soil, which are brought into suit- ^g^™^^ able conditions of combination, and distributed through a wide range of soU and subsoil. So much, then, for the benefits of rotation, so far as the requirements, the habits of growth, and the capabilities of 254 THE EOTHAMSTED EXPEEIMENTS. Relation of rotation to economical Rotation and sale of prodMce. Rotation bution of laiou/r. Rotation and clean- ing land. gathering and assimilating the various mineral constituents, and the nitrogen, of the different crops, are concerned. It cannot be doubted, that the difference in the amounts, in the conditions of combination, and in the distribution within the soil, of the various mineral constituents, is at least an element in the explanation of the benefits of alternation ; nor, on the other hand, can there be any doubt that the facts relating to the amount, and to the sources, of the nitrogen of the different crops, are of still greater significance than are those in regard to the mineral constituents. But, it is not only the conditions of growth, but the uses of the different crops when grown, that have to be taken into account. Thus, the cereals, when grown in rotation, yield more produce for sale in the season of growth than when grown continuously. Again, the crops alternated with them accumulate very much more of mineral constituents and of nitrogen in their produce, than do the cereals themselves ; and, by far the greater proportion of those constituents re- mains in circulation in the manure of the farm, whilst the remainder yields highly valuable products for sale in the forms of meat and milk. Further, independently of the benefits arising from the difference in the requirements and results of growth of the different crops, of the increased amount of manure available, and of the increased sale of highly valuable animal products, there are other elements of advantage of considerable import- ance. Eor example, with a variety of crops, the mechanical operations of the farm, involving horse and hand labour, are better distributed over the year, and are therefore more economically performed. Last, but by no means least, the opportunities which alternate cropping affords for the clean- ing of the land from weeds is a prominent element of ad- vantage. Thus, then, the benefits of rotation are very various; and the explanation of them, though largely dependent on the facts which have been ascertained by scientific investigation, also largely involves considerations connected with the general economy of the farm ; and since, as has been seen, so large a proportion of the produce grown is retained on the farm, as stock-food or litter, it is obvious that the benefits cannot be fully appreciated without arriving at some definite idea of the importance to the farmer of the saleable animal products, and of the manure obtained. This subject will be considered in Section VI., which now follows. FEECrNG OF ANIMALS. 255 SECTION VI.— THE FEEDING OF ANIMALS FOR THE PRODUCTION OF MEAT, MILK, AND MANURE, AND FOR THE EXERCISE OF FORCE. Inteoduction and History. It was shown in the last Section (V.), on the Eotation of stock-feed- Crops, that any explanation of the benefits of rotation is ^^f^^' quite inadequate which does not take into account the results img. of the feeding of the animals on the farm. Thus, in the discussion of the amounts of the produce of the various crops grown in alternation with one another, and of the amounts of the various constituents of the individual crops, or of their separate parts, it was pointed out that only certain portions of them were at once available as saleable products ; a large proportion remaining for use on the farm in some way, and only eventually yielding a profitable return. The extent to which the retention on the farm of the ConstUu- constituents accumulated in the crops may take place, may ^^„f^g. usefully be illustrated by reference to a particular example, ■mmiedfram which will convey a clearer conception of the importance f'^JS' of the subject than any mere general statement can do. land. Accordingly, in Table 66 is given an approximate estimate of the proportion of certain selected constituents of the crops grown in the typical four-course rotation of Swedish turnips, barley, leguminous crop, and wheat; which would be at once sold off the farm, and of the amounts retained upon it ; supposing that only the grain of the cereals is sold, and that the root -crop, the leguminous crop, and the straw of the cereals, are retained for further use. The estimates are TABLE 66. — Illusteation of the Peopoetion or the Constituents OF Ceops geown in Eotation, at once sold off the Faem, AND OF THOSE RETAINED UPON IT FOE FDRTHEE USE. Per cent of total in the crops. At once sold off the farm. Retained on the farm for farther use. Diy matter Mtrogen Total mineral matter (asli) Phosphoric acid .... Potash per cent. 30.6 43.4 14.5 56.2 20.0 per cent 69.4 56.6 85.5 43.8 80.0 256 THE EOTHAMSTED EXPERIMENTS. founded on the average amounts of produce obtained, over eight courses, in the fully manured rotation, the particulars of which were given and discussed in the paper on Eotation above referred to. It is true that the exact figures given in the table have only reference to a particular case, and that in practice there will sometimes be larger and sometimes smaller proportions of these constituents of the crops at once sold, or retained on the farm. Nevertheless, the illustrations may be taken as essentially typical, and as so far conveying a very useful impression on the subject. Produce Eefcrring to the figures, the question arises — To what bene- for^HoO:- ^^^^^ 0^ profitable purposes are about two-thirds of the total feeOimg. vegetable substance grown, more than half its nitrogen, nearly half its phosphoric acid, and about four-fifths of its potash, retained on the farm? Briefly stated, it is for the feeding of animals for the production of meat, milk, and manure, and for the exercise of force — that is, for their labour. It is, then, the facts, and the principles, involved in the feed- ing of the animals of the farm for these various purposes, that we have now to consider. Increased It is obvious that, SO long as a country is only sparsely ^d^'^ populated, and the needs of the people are amply supplied nomicai under a comparatively rude system of agriculture, in which feeding. extended area precludes the necessity for improved methods, there would be little either of scope or of inducement to study economy in the feeding -of animals, or to systematic practice in regard to it. But as population increases in proportion to area, there arises the necessity for increased production over a given area. It was pointed out in our paper on Eotation that, in our country, gradually a greater variety of crops came to be grown; that first leguminous crops, and then root-crops, were introduced, and finally the system of rotation became general. Thus, a much greater variety, and a much greater quantity, of home-produced stock-foods became available; and in time foods of various kinds were imported from other countries. Somewhat similar changes in their food resources occurred in various'parts of the Continent of Europe ; and with these came the inducement, if not the necessity, to pay more atten- tion to the subject of feeding. The end was, however, sought to be attained by somewhat characteristically different methods Improve- in our own country and on the Continent. With us, more ^rlteA. special attention was paid to the improvement of the breeds of the farm animals themselves — not only to enhance the development of the most valuable characters in the final product, but to secure early maturity, and thus materially FEEDING OF ANIMALS. 257 to economise the expenditure of food in the mere mainten- ance of the living meat-and-manure-making machine. As to the use and adaptation of different foods, but little systematic inquiry was undertaken in regard to it, each feeder relying largely on his own judgment, or on the unwritten rules adopted in his locality, as the result of practical experience. On the Continent, however, and especially in Germany, Gontin- much more attention was paid to the character of the food ^*«V«e of the nitrogenous, and of the non-nitrogenous constituents, bodies and in the bodies of the animals themselves, and in their increase ««'"«™*»^- whilst fattening ; and it also involved that of the composition of the excrements, that is, of the manure. Thus, the inquiry embraced the following points : — Points em- 1. The amount of food, and of its several constituents, con- braced in sumed in relation to a given Hve- weight of animal, within a ^■^*'^ given time. ments. 264 THE EOTHAMSTED EXPERIMENTS. 2. The amount of food, and of its several constituents, consumed to produce a given amount of increase in live- weight. 3. The proportion, and relative development, of the different organs or parts of different animals. 4. The proximate and ultimate composition of the animals, in different conditions as to age and fatness ; and the probable composition of their increase in live-weight during the fatten- ing process. 5. The composition of the solid and liquid excreta (the manure) in relation to that of the food consumed. 6. The loss or expenditure of constituents by respiration and the cutaneous exhalations — that is, in the mere susten- ance of the living meat-and-manure-making machine. 7. The yield of milk in relation to the food consumed to produce it ; and the influence of different descriptions of food on the quantity, and on the composition, of the milk. As already said, several hundred animals, oxen, sheep, and pigs, have been submitted to experiment. The amount, and the relative development, of the different organs or parts were determined in 2 calves, 2 heifers, 14 bullocks, 1 lamb, 249 sheep, and 59 pigs. The percentages of water, mineral matter, fat, and nitro- genous substance, were determined in certain separated parts, and in the entire bodies, of ten animals — namely, 1 calf, 2 oxen, 1 lamb, 4 sheep, and 2 pigs. Complete analyses of the ashes, respectively of the entire carcasses, of the mixed in- ternal and other " offal " parts, and of the entire bodies, of each of these ten animals, have also been made. From the data provided as above described, as to the chemi- cal composition of the different descriptions of animal in differ- ent conditions as to age and fatness, the composition of the increase whilst fattening, and the relation of the constituents stored up in the increase to those consumed in food, have been estimated. To ascertain the composition of the manure in relation to that of the food consumed, oxen, sheep, and pigs, have been experimented upon. The loss or expenditure of constituents, by respiration and the cutaneous exhalations, has not been determined directly — that is, by means of a respiration apparatus, but only by difference — that is, by calculation, founded on the amounts of dry matter, ash, and nitrogen, in the food, and in the (in- crease) faeces and urine. Independently of the points,of inquiry above enumerated, the results obtained have supplied data for the consideration of the following questions : — FEEDING OF ANIMALS. 265 1. The sources in the food of the fat produced in the incidental animal body. «"^'^- 2. The characteristic demands of the animal body (for nitrogenous or non-nitrogenous constituents of food) in the exercise of muscular power. 3. The comparative characters of animal and vegetable foods in human dietaries. Food Consumed and Inckease Produced. It is proposed, first, to consider the results illustrating the amounts of food, and of its nitrogenous and non-nitrogenous constituents respectively, consumed by a given live-weight of animal within a given time, and the amounts required to produce a given amount of increase in live -weight. The illustrations will be drawn from experiments with sheep and with pigs. The Experiments with Sheep. Table 67 (p. 266) shows, for each of three series of experi- Quantity ments with sheep, in the first three columns the amounts of "•^■^°°'??2f nitrogenous, of non-nitrogenous, and of total organic substance, sheep. consumed per 100 lb. live-weight per week, and in the last three columns the amounts consumed to produce 100 lb. increase in live^weight. The figures represent the quantities of the crude constituents consumed — that is, the amounts of nitrogenous substance calculated by multiplying the nitrogen by 6.3, which implies that the whole of it exists in the foods as albuminoids, which admittedly is not the case. It will be Table 67 seen, however, that this method is quite sufficient for the ^P^'^"'^- purposes of the illustrations at present in view, though it is frequently far from accurate in the case of unripened vege- table products ; and it is especially so in that of succulent foods, such as feeding roots. The amounts of crude non- nitrogenous substance are calculated by deducting those of the mineral matter, and of the crude nitrogenous constitu- ents, from those of the total dry matter consumed. Here again, it is admitted that the results are only approximations to the truth ; but it will be seen that, as in the case of the nitrogenous constituents, they are nevertheless quite sufficient for the purposes of our present illustrations. The crude total organic matter is simply the sum of the nitrogenous and non- nitrogenous calculated as above. Eeferring to the results, it is impossible to go into any Expiana- detail here. A glance at the figures in the first three columns ^^^ of the Table (67) relating to the amounts of the constituents consumed per 100 lb. live-weight per week is sufficient to show 266 THE BOTHAMSTED EXPERIMENTS. that, in all comparable cases, tBere was much more uniformity in the amounts of the non -nitrogenous than in those of the nitrogenous substance consumed for a given live-weight of the fattening animal within a given time. The deviations from this general regularity in the amount of non-nitrogenous substance consumed are, indeed, in most cases such that, when they are examined, they tend clearly to show that the uni- formity would be considerably greater if the amounts of only the really available respiratory and fat-forming constit- uents had been represented, instead of those of the crude or total non-nitrogenous substance consumed. TABLE 6V.— ExPBEiMENTS with Sheep made at Eothamstbd in 1850. NiTEOGBNOtrS AND NoN-NITROGENOTJS CONSTITUENTS CONSUMED PEE 100 LB. LIVE- WEIGHT PER WEEK ; AND TO PEODUCE 100 LB. INCREASE IN LIVE-WEIGHT. Limited food. Ad Uhitum food. Per 100 lb. Jive-weight per week. Nitro- ge- Non- nitrp- genoiis. Total organic. To produce 100 lb. in- crease in live-weight. Ni- troge- nous. Non- nitro- genous. game. SBEIES 1. 5 SHEEP IN EACH PEN (14 WEEKS). 1 2 3 i Linseed-cake . Oats .... Clover-cliaflF . Oat-straw chaff ) Swe- ( V dish 4 j turnips 2.46 1.S7 1.64 1.07 9.85 11.36 13.12 10.17 12.31 12.93 14.76 11.24 167 103 102 102 650' 684 736 913 817 787 , 838 ,1015 Mean . 1.68 11.13 12.81 118 746 : ; 864 SERIES 2. 5 SHEEP IN EACH PEN (19 WEEKS). 1 2 3 4 Linseed-oake . Linseed .... Barley .... .Malt .... ( Clover- J r chaff ■) 3.78 3.21 2.58 2.52 12.93 12.66 13.79 14.02 16.71 15.87 16.37 16.55 821 289 235 266 1103 1144 1269 1457 1424 1433 1504 1723 Mean 3.02 13.35 16.38 278 1244 1621 SERIES 3. 51 SHEEP IN EACH PEN (10 WEEKS). Barley .... Malt and malt dust . Barley (steeped) Malt and dust (steeped) . Malt and dust (extra quantity) Man- 1.70 1.64 2.08 1.77 1.89 10.59 10.12 12.60 10.70 11.63 12.29 11.76 14.68 12.47 13.52 118 111 121 136 126 731 677 730 821 776 850 788 851 958 903 Mean 1.82 ,11.13 12.94 123 747 870 * Only four sheep in pens 1, 3, and 4. FEEDING OF ANIMALS, 267 As pointed out in our earlier papers, in reading the figures allowance has obviously to be made, both for those of the non-nitrogenous constituents which would probably become at once effete, and also for the different respiratory and fat- forming capacities of the portions which are digestible. Thus, comparing series with series, the amounts are higher in Series II. where the ad libitum, food was clover-chaff containing a large amount of indigestible fibre, than in either of the other series where it consisted of Swedish turnips or mangel-wurzel. Then, the quantity consumed was higher in the third pen of Series I., with clover-chaff, than in the other pens of the same series ; and it was lower in pen 1 of Series I. with linseed- cake containing much oil, and it was again lower in pens 1 and 2 of Series II., also with much fatty matter in the food, than in the other pens of the same series with cereal grain. Indeed, when we bear in mind the various circumstances which must tend to modify the indications of the actual figures, it will be admitted that the coincidences in the amounts of available respiratory and fat -forming constitu- ents consumed by a given weight of animal within a given time, are much more striking and conclusive than, consider- ing the views prevalent on the subject at the time, could have been anticipated. With this general uniformity in the amounts of the non- nitrogenous substance consumed by a given live - weight within a given time, the amounts of the nitrogenous con- stituents so consumed are, on the other hand, seen to vary under the same circumstances in the proportion of from one to two, or three, or more. Let us now refer to the last three columns of Table 67, which show the amounts of the respective constituents con- sumed to produce 100 lb. increase in live-weight. In consider- ing these results we must, as when discussing those relating to the consumption by a given live-weight within a given time, read the indications of the actual figures as modified by the obviously different capacities for the purposes of the ani- mal economy, of the substances the amounts of which they are assumed to represent. It must also be borne in mind, that the proportion of real dry substance in the increase of the animal will vary to some extent, according to the char- acter of the food. For example, it will be rather the less, the Food de- more succulent the food, and the greater, the greater the pro- ^^^^ portion of fat in the increase. Again, as in the case of the ance and results showing the consumption for a given live-weight of the ^^^^^ *^ fattening animal within a given time, the figures represented the demand — not only for respiration, and for maintenance in other respects, but also that for increase in live- weight, so 268 THE EOTHAMSTED EXPEEIMENTS. now those specially arranged to show the relation of con- sumption to increase, at the same time include the amounts required by the exigencies of respiration and maintenance. Taking a general view of the results, which is all that can '<»^°f be done here, it is seen that where clover-chaff, with its large '°™ *' amount of indigestible woody-fibre, was used as the ad libitum food, the total amount of non-nitrogenous substance consumed to produce a given increase in live- weight was' much greater than where the ad libitum food consisted of roots. Due allowance must, therefore, be made for this in comparing the results of one series with those of another. Doing this, it is obvious that the amounts of really available non-nitrogenous substances consumed were, at any rate much more nearly uniform in the different series, and in the different pens, than were those of the nitrogenous substance. Of the dif- ferences that would still remain, most would be again reduced by making allowance for the different respiratory and fat- forming capacities of the remaining available non-nitrogenous constituents; since, for example, much less of fatty matter would be required than of starch or sugar, or of the pectine compounds of the roots. Again, as in the case of the consumption by a given live- weight within a given time, so now in that of the consump- tion to produce a given amount of increase, there is a much wider range of difference in the amounts of the nitrogenous than of the non-nitrogenous constituents consumed ; and the differences are, as before, much greater than can be explained by the differences in the character of the nitrogenous sub- stances which the figures represent in the different cases. Former Thus, then, the results of these experiments with sheep, condusions -^i^en interpreted with due regard to the known differences in the character of the nitrogenous and non-nitrogenous con- stituents in the different foods, fully justify the conclusions drawn from them more than forty years ago — namely, that taking food-stuffs as they go, it is their supply of the digestible non-nitrogenous, that is of the more specially respiratory and Fat-form- fat-forming constituents, rather than that of the nitrogenous TevTti °^ specially flesh-forming ones, that regulates, both the factors. amount of food consumed by a given live-weight of animal within a given time, and the amount of increase in live- weight produced. But, as it seems to have been tacitly assumed in recent years, since much attention has been paid to the investigation of the digestibility of the different constituents of foods, that conclusions founded on the determined amount in the foods of the crude substances only cannot be relied upon, we have had the whole of our early results, both with sheep and with FEEDING OF AOTMALS. 269 pigs, re-calculated, making correction, as far as practicable, Re-aOcu- for the amounts of the constituents in the different foods ^^jf% which are assumed to he indigestible, or otherwise not of Wolff's food-value, according to the tables given by Emil von Wolff ^^• in the edition of his work published in 1888. He there gives for nearly 200 different articles of stock foods — ^the percent- ages of water, mineral matter (ash), crude protein, crude fibre, non-nitrogenous extractive matters, and crude fat ; and then the percentages of digestible albuminoids, digestible carbo- hydrates, and digestible fat. In applying his data to our results, the amount of the crude substance in each description of food is reduced in the proportion which his figures show of crude to digestible in the same description of food. Further, in the case of the so estimated amounts of digestible fatty matter, the figure obtained has been multiplied by 2.4 to bring it approximately to its equivalent of carbohydrate, the amount then being added to the other digestible non-nitro- genous substance, so reckoning the whole as carbohydrate. Lastly, as Wolff makes no correction for the non-albuminoid condition of a large portion of the nitrogen in succulent roots, it has been assumed, in accordance with results obtained at Eothamsted and elsewhere, that in Swedish turnips only 45 per cent, and in mangels only 40 per cent, of the total nitro- gen will exist as albuminoids. There are obvious objections to some of the modes adopted for the determination of the digestible constituents of the various foods, which render them inapplicable without con- siderable reservation, to the estimate of the amounts of the constituents which will probably be actually digested in the case of ordinary liberal rations. But, if accepted as approxi- mations only, they undoubtedly afford useful data for some general conclusions. Neither space nor time wOl permit of either the record or Re-calm- the discussion of the re-calculated tables. It must suflBce ^^^f*^' here to say that the results as so re-calculated, that is making fmvier correction as far as present knovrledge permits, for the prob- ^°ft^ different conditions, for the purpose of estimating the com- Uve-weight. position of their increase in passing from one condition to another. First referring to oxen, the composition of their increase during the feeding process has been estimated in the case of the recorded results of actual practical feeding, in some cases of large numbers of animals, and over considerable periods of time. Other cases have been those of results obtained at Eothamsted, or under Eothamsted superintendence, mostly in direct feeding experiments, but sometimes in the feeding of animals in the ordinary practice of the farm. Eeviewing the whole of the results, the indication was, that the composition of the increase of moderately fattened oxen during a final fattening period of several months will contain about, or little more than, 1^ per cent of mineral matter, seldom more than 7 to 8 per cent of nitrogenous sub- stance and seldom as little as 60, and generally nearer 65 per cent of fat J whilst the total dry substance of the increase 282 THE HOTHAMSTED EXPEKIMENTS. Difference in growing and fatten- will generally range from 70 to 75 per cent. In the case, however, of oxen fattened very young, and the feeding period extending over a much longer time, similar calculations lead to the conclusion that the growing and fattening increase of such animals may contain perhaps 2 J per cent or more of mineral matter, against only about 1 J per cent over a limited final period of more purely fattening increase ; about 10 per cent of nitrogenous substance, against only 7 to 8 per cent in the only fattening increase ; and perhaps only from 50 to 55 per cent of fat, against from 60 to 65 per cent in the more exclusively fattening increase. In fact, whilst the growing and fattening increase would consist of about two-thirds dry substance and one-third water, that of the more purely fatten- ing increase would consist of nearly three-fourths dry sub- stance and only about one-fourth water. Similar results relating to sheep, lead to the conclusion that during a final period of some months of feeding on good fattening' food, their increase will generally contain not less than 2 per cent of mineral matter, and frec[uently more ; that is distinctly more than in the case of oxen, the quantity largely depending on the amount of wool. Of nitrogenous substance, the final fattening increase of sheep will probably seldom contain more than 7 per cent, and frequently some- what less. In other words, notwithstanding the large amount of nitrogen in the wool of sheep, their fattening increase will probably generally contain less nitrogenous substance than that of oxen. On the other hand, the increase of well fed and moderately fattened sheep will generally contain nearly, and sometimes more than, 70 per cent of fat, against an average of less than 65 per cent in the case of oxen ; and in the case of very fat sheep the percentage of fat in the increase may even reach 75 per cent. Upon the whole, it may be assumed that the increase of liberally fed and moderately fattened sheep, over several months of final fattening, will probably consist of about 2 per cent of mineral matter, about, or less than, 7 per cent of nitrogenous substance, from 65 to 70 per cent of fat ; and in all, of from 75 to 80 per cent of total dry substance ; whilst the increase over the final period of excessive fattening may contain from 70 to 75 per cent of fat, and from 80 to 85 per cent of total dry substance. Eeferring to pigs, the increase of those liberally and suit- ably fed for fresh pork will probably, on the average, contain — an immaterial amount of mineral matter, only from 6^ to 7J per cent of nitrogenous substance, from 65 to 70 per cent of fat, and from 70 to 75 per cent of total dry substance. The increase over the last few months of high feeding of pigs fed FEEDING OF ANIMALS. 283 for curing will, however, probably contain lower percentages of nitrogenous substance, but higher, and sometimes consider- ably higher, percentages of both fat and total dry substance. The tendency of the demand in recent years has, however, been for less excessively fat bacon than formerly. Thus far, then, it has been shown that the amounts of food, Mtrogen- or of its various constituents, consumed, both for a given live- ^^^^o- weight of animal within a given time, and to produce a given genous sub- amount of increase, were very much more dependent on the S^"^^ "-^ quantities of the non-nitrogenous, than on those of the nitro- increase in genous constituents, which the food supplied. It has been '^e-we^At said, that when the large requirement for non-nitrogenous constituents of food to meet the expenditure by respiration is borne in mind, it need not excite surprise that consumption in relation to a given live- weight within a given time should be so largely measurable by the amount of digestible and available non-nitrogenous substance which the food supplies ; but that, at first sight, it was less intelligible that the quan- tities consumed to produce a given amount of increase in live- weight should also be much more dependent on the supplies of the non-nitrogenous, than on those of the nitrogenous, constituents of the food. The results relating to the chemical composition of the dif- PropoHi^n f erent animals, in different conditions as to age and maturity, °^fi'' "'^ , , , ' _ . _ *=* . •' * nitrogenous nave shown, however, that even store animals may contain as matter in much, or even more, of the non-nitrogenous substance — fat — *^ ^' „ -than of nitrogenous substance ; whilst the bodies of fattened animals. animals may contain two, three, four, or even more times as much dry fat as dry nitrogenous matter. It has further been shown, that the proportion of fat to nitrogenous substance in the increase in live-weight of the fattening animal, is much higher than in the entire bodies of the fattened animals. If, therefore, the non-nitrogenous substance of the increase — the fat — is derived from the non-nitrogenous constituents of the food, the relatively large demand for such constituents for the production of fattening increase, would seem to be amply accounted for. The important question arises, therefore, What are the Animport- sources in the food of the fat of the fattening animal ? In ti^^^"^' other words, from what constituent or constituents in the food is the fat produced ? 284 THE EOTHAMSTED EXPERIMENTS. Sovirce of fat. of Boussin- gault and others. Rothatn- sted experi- ments. Fat in ani- mals and in food. Fat derived from carbo- SOUECES IN THE FoOD OF THE FAT PKODUCED m THE Animal Body. Prior to the publication of Liebig's work on Organic Chem- istry in its Applications to Physiology and Pathology, in 1842, it seems to have been assumed that the Herbivora derived their fat from ready-formed fatty matters in their food ; and that the Carnivora derived theirs from the ready-formed fat of the animals they consumed. Liebig argued that, as a rule, the food consumed by the Herbivora did not contain suffi- cient fatty matter for the purpose ; and he maintained that, although fat might be formed from the nitrogenous substance of the food, its main source was the starch, sugar, and other carbohydrates, which the food supplied. Dumas and Boussingault ^ at first called in question the view that fat was produced in the animal body, and assumed that the food of the Herbivora supplied sufficient fatty matter to account for the whole of the fat stored up. Subsequently, however, Dumas and Milne-Edwards,^ from the results of ex- periments with bees, Persoz^ from experiments with geese, and Boussingault * from those with pigs, geese, and ducks, concluded that fat was formed from the carbohydrates of the food. At the same time Boussingault considered that, in normal feeding, the amount of albuminoids consumed would generally supply sufficient carbon for the production of the fat formed by the animal. Next came the evidence of the Eothamsted experiments, the majority of which were conducted within the years 1848- 1853 inclusive; and they involved feeding experiments on between 400 and 500 animals, with foods of known composi- tion ; the slaughter, determination of the weights of the parts, and noting on the character as to fatness, &c., of more than 300 animals ; and finally, the chemical analysis of ten animals. In the first place, it was clearly demonstrated that much more fat was stored up in the bodies of the fattening animals than could be derived from the ready-formed fatty matter in their food. Secondly, from a careful study of the enormous amount of experimental data obtained, as well as of the known facts of practical experience in feeding, it was con- sidered no doubt whatever could be entertained that much, if not the whole, of the fat formed in the bodies of the herbivora fed for the production of meat was derived from the carbohydrates of the food. ^ Balance of Orgcmic Nature, 1844, p. 116 et seq. ^ Corwpt. Bend., vol. xvii. p. 531. ^ Aim. Chim. Fhys., vol. xiv. p. 408 et, seq. ^ Ibid., vol. xiv. p. 419 et seq.; xviii. p. 444 et seq. FEEDING OF ANIMALS. 285 In fact, the experimentally determined relation of the non- nitrogenous, and of the nitrogenous constituents of the food, respectively, to the amount of increase produced ; the com- position of fattening increase generally ; the relatively greater tendency to grow in frame and to form flesh with highly nitrogenous food; the greater tendency to form fat with food comparatively rich in non-nitrogenous substances, and especially in carbohydrates ; and conimon experience in feeding — all pointed in the same direction. For some years there was little or no discussion on the Liebig's subject, and it seemed to be tacitly admitted, both on the ™^^^' Continent and in this country, that the views of Liebig, as to the formation of at any rate much of the fat of the herbivora from carbohydrates, were correct. In 1865, however, at a meeting of a Congress of Agri- views of cultural Chemists, held at Munich, in August of that year, p°j(g^^ Professor Voit, from the results of experiments made in ko/er. Pettenkofer's respiration apparatus, with dogs fed chiefly on flesh, maintained that fat must have been produced from nitrogenous substance ; and that this was probably the chief, if not the only, source of the fat even of herbivora. Pettenkofer and Voit further maintained, that to establish the formation of fat from the carbohydrates, experiments must be brought forward in which the fat deposited was in excess of that supplied by the food, plus that which could be derived from the transformation of albumin. Of course, the mere fact that the food consumed contained enough nitrogenous substance for the formation of all the fat that had been produced, would of itself be no proof that that substance had been its exclusive source. On the other hand, if the amount of fat stored up in the animal was in excess of that which could be derived from the ready-formed fatty matter of the food, and from the transformation of the nitrogenous substance, it would be proved that at any rate some of the stored-up fat must have had another source — and this could only be the carbohydrates. Accordingly, the results of many of the Eothamsted feeding experiments were calculated, to ascertain whether or not the ready-formed fat, and the nitrogenous substance of the food, were sufficient to account for the whole of the fat estimated to have been stored up. None of the experiments had been specially arranged with a view to the elucidation of this question. In some of them, however, what may be called minimum amounts, and in others excessive quantities of nitrogenous substance, had been consumed. Some of the Rotham- results seemed to us to afford clear evidence on the point, ^(edresvzu. and we gave a paper on the subject in the Physiological 286 THE KOTHAMSTED EXPERIMENTS. Section, at the meeting of the British Association for the Advancement of Science, at Nottingham, in 1866 ; and it was published, in abstract, in the 'Eeport of the British Association' for 1866, and in full in the 'Philosophical Magazine ' for December of that year. And, as it is upon the results as then given that any subsequent discussion of our conclusion has been founded, it is proposed, in the first place, to consider the evidence afforded by those results ; but afterwards to adduce certain modifications of some of them, in order to bring them more into accord with recent know- ledge on some points, and to meet more effectively objections that have been raised against the conclusions drawn from them. The first point to consider was — What description of animal is likely to yield the most direct and conclusive results on the subject 1 Obviously the one which is fed more especially with the view to the production of fat ; which consumes in its most appropriate fattening food a comparatively low propor- tion of nitrogenous substance, and a comparatively high' pro- portion of carbohydrates ; and which yields a large proportion of fat, both in relation to the weight of its body within a given time, and to the amount of food consumed. The fol- Tabie69 lowing Table (69) briefly summarises the results of very evplamed. inxmerous experiments with oxen, sheep, and pigs, so far as they illustrate the comparative characters of the different descriptions of animal in regard to the points above enumerated. Fattening In the first place it is to be observed, that although the qualities of proportion of intestines and contents is greater, that of the stomach and contents is very much less in the pig than in either of the ruminants, as also is that of the stomachs and contents, and intestines and contents, taken together ; the percentage of these collectively being, in oxen 14.3, in sheep 10.9, and in pigs only 7.5 of the weight of the body. The fact is, that the appropriate fattening food of the pig consists of ripened seeds, and highly starchy roots, containing but little indigestible fibre, whilst that of the ruminants contains a considerable amount of slowly digestible or indigestible cellulose, and often a much greater amount of indigestible or unassimilable nitrogenous substance. The result is, that a less proportion of the live-weight of the pig consists of more or less effete matter retained in the alimentary organs. Then, the second division of the table shows, that with the much higher character of its food, and the much less propor- tion of it -indigestible and effete, the pig both consumes very much more, and yields very much more increase, for a given live-weight within a given time. FEEDING OF AiJIMALS. 281 Lastly, as is shown in the third division of the table, for 100 of dry substance of food consumed, the pig yields very much more, both of fat and of dry substance in increase ; and, on the other hand, voids very much less of dry substance in urine and in faeces. TABLE 69.- -Showing the Compabative Fattening Qualities OF DIPFEBBNT AnIMALS. Oxen. Sheep. Pigs. RELATION OF PAETS IN 100 LIVE-WEIGHT. Average of 16 249 59 Stomach and contents . Intestines and contents 11.5 2.8 7.4 3.5 1.3 6.2 Internal loose fat Heart, aorta, lungs, windpipe, Uver, gall- bladder and contents, pancreas, spleen, and blood Other offal parts 14.3 4.6 7.0 13.0 10.9 7.0 7.3 15.0 7.5 1.6 6.6 1.1 Total offal parts Carcass Loss by evaporation, &c 38.9 59.3 1.8 40.2 59.7 0.1 16.8 82.6 0.6 Total 100.0 100.0 100.0 PER 100 LIVE-WEIGHT. Dry substance consumed in food per week . Increase yielded per week . . . . 12.5 1.13 16.0 1.76 27.0 6.43 PER 100 DRY SUBSTANCE OF FOOD. Fat in increase .... Total dry substance in increase . Total dry substance in excretions 15.7 17.6 16.7 AVERAGE FAT PER CENT. In lean condition In fat condition . In increase wliilst fattening 16.0 18.0 30.0 33.0 60.0 65.0 22.0 44.0 70.0 Thus, as compared with either oxen or sheep, the pig offers Figs most many advantages as a subject for the consideration of the »«»*«6.^e /or- relations of food and increase, and consequently for that of rmnts. the source in the food of the fat which he yields. He has a 288 THE KOTHAMSTED EXPEBIMENTS. less proportion of alimentary organs and contents ; he con- sumes more food in proportion to his weight; he yields a larger proportion both of total increase and of fat ; and fin- ally, much less of his food is effete and voided. The general result is, that changes in his live- weight are in a much less proportion influenced by variations in the contents of the ali- mentary organs, and are, therefore, much truer indications of change in the substance of the body ; and hence the range of error in calculating the amount and composition of his increase in relation to the amount and composition of the food consumed, is much less. The Eolyperiments at Bothamsted with Pigs. In the selection of the experiments with pigs, for calculat- ing whether more fat was stored up than could possibly have been derived from the ready-formed fat and the nitrogenous substance of the food, some have been taken in which the proportion of the nitrogenous to the non-nitrogenous constitu- ents of the food was abnormally high, and others in which it was fairly normal, or even low. In all cases, the experiments were conducted for periods of not less than eight or ten weeks ; and the amounts, both of total increase, and of fat stored up, were so large in proportion both to the original weight of the animal, and to the amount of food consumed, that the data obtained may safely be relied upon for the settlement of the question at issue. Table 70 ^ In the upper portion of the next Table (70) are recorded some particulars of the nine experiments selected for calcula- tion — namely, the description of the food, the number of animals experimented upon, the duration of the experiment, the original and final liv^-weights, the increase per head and on 100 original weight, the percentage of carcass in fasted live-weight, and the amount of crude non-nitrogenous to 1 of crude nitrogenous substance in the food. The middle division of the table shows, for 100 increase in live-weight — the amount of nitrogenous substance consumed in the food, the amount of it estimated to be stored up in the increase, and the quantity remaining, and therefore possibly available for the formation of fat. Next, there are given — the estimated amount of fat in the increase, the amount ready-formed in the food, and the difference — that is, the amount newly-formed. There are then given — the amounts of carbon in the estimated newly-formed fat, the amounts in the available nitrogenous substance minus that in the urea formed, supposing the whole of the nitrogen not stored up in increase to contribute to such formation ; and lastly, the dif- FEEDING OF ANIMALS. 289 ference, that is, the amount of carbon available from the nitro- genous substance for the formation of fat, more or less than that required for the amount of fat produced. Then, in the bottom division of the table are shown, /or 100 of carbon in the estimated jprodv^ed fat — the amount available from the nitrogenous substance, and the amount not available from that source, in each experiment; the amount not so available representing, of course, the proportion required from other sources. It is hardly necessary to point out, that according to the above mode of illustration, the figures show, not only the ut- most proportion of the stored up fat which could possibly have had its source in the nitrogenous substance of the food, but notably more than could possibly have been so derived. Thiis, to say nothing of other considerations, it has been as- sumed, for simplicity of illustration, and for the sake of argument, that the whole of the nitrogenous substance of the food not stored up as increase would be perfectly digested, and be available for fat-formation ; and that, in the breaking up of the nitrogenous substance for the formation of fat, no other carbon compounds than fat and urea would be produced ; and lastly, that the whole of the ready-formed fatty matter of the food has contributed to the fat stored up. It is obvious, however, that these assumptions are in part improbable, and in part quite inadmissible ; whilst the tendency of the error is, in each case, to show too large a proportion of the stored up fat to have been possibly derived from the ready-formed fat, and the nitrogenous constituents, of the food. It is obvious, therefore, that where the figures show an Aviowntof excess of carbon available from nitrogenous substance over ■^^^^^^^^ that which would be required if the produced fat had been smn-ces formed from it,, the excess is over-estimated; and, on the^^^^^ other hand, that where they show a deficiency of nitrogenous /^wes substance for such formation, the deficiency is under-esti- **<""• mated; so that, in fact, the amount of fat required to be de- rived from other sources would be greater than the figures indicate. Indeed, according to the mode of calculation adopted, 100 of nitrogenous substance would yield 62 parts of fat ; Ijut it has been fuUy admitted in subsequent discus- sions, that at most 51.4 parts of fat could possibly be derived from 100 parts of proteid substance; and more recently a much lower figure has been adopted. After these general remarks, we may now turn to the con- Results. ' sideration of the results of the different experiments. i In experiment 1, two pigs of the same fitter, of almost ex- [ actly equal weight, and, as far as could be judged, of similar \ character, were selected. One was killed at dnce, and the VOL. VIL T 290 THE EOTHAMSTED EXPERIMENTS. a ? a " ^ -a CO o la n' o CO CO IP t- li CD ^1 O -* CO Mixture, equal parts bran and lentil-meal ad Uh. M Bean-meal, lentil-meal, bran, and maize-meal, each ad Mb, rrt Bean-meal, lentil-meal, and bran, each 1 part, barley-meal 3 parts. 1 CO ^O ^i- CO r-H tavi •#r-l t-»0 CO t- CO rH 01 ■* ■* o« i-l (M 1-* ^ . «!-*3 ^ rt ftp u a> r^ .»-g ^3 .b T a a > 0, ID S.2 •35 .SrH Ut o o« 00 atto CO oeq 00 64 00 m ad CO o 1 OO o oeo t- rHiO to TlioO oit^ -rtt MrH d lO 1 o© ■* OS a. o iHO 1-4 fti «D 00 t* lO 1 o>o U3 fHcn (N CO rH (M -4;^: fO com CO »£S 1 OiiS U) (N Tt* 00 CO-* Oi ^o mi>: iO 1 1 OCQ )>■ OCO *- tot- Ol o^ w « h' I &: Bl > ot* CO «0(N •* o 3 Oi I-H OO iric4 t- !zi I-H l-t + i OrH O) a»^ lO M iH a> l^O + o 1^ OJ rC\ r^ cotn w* t-I N s OS CO ■* + CM . fl ■ ^3 tS 1 u ^ .l ^ .s ■ ■ • .^ ►d 1 . ^t s ? >. ^ I'a °T, 13 i 03 , Sid s ■rt .Sfi ss ss oa' • ■ 3)-g n S OO s <, 3 FEEDING OF ANIMALS. 291 amount of total dry or solid matter, of nitrogenous substance, of fat, and of mineral matter, determined in it. Tlie other was then fed for a period of ten weeks, on a mixture consist- ing of — bean-meal, lentil-meal, and bran, each 1 part, and barley-meal 3 parts, given ad libitum. It was then weighed, killed, and its composition determined as in the case of the other animal. In fact, the object of the experiment was, to determine the composition of a " store " and of a "fat " pig, and to estimate the composition of its increase whilst fatten- ing ; and the data thus provided have formed the basis of the estimate of the fat in the increase, not only in the case of experiment 1, to which they directly apply, but in that of each of the other eight experiments, the results relating to which are recorded in the table. On this point it may be observed that, taking into consideration the weight and con- dition of the animals at the commencement, the character of the foods, the length of the fattening period, the proportion of increase upon the original live-weight, and the final con- dition of the animals, it may perhaps be concluded, that the tendency of error in the calculations would be to give the proportion of fat in the increase somewhat too high in experi- ments 2 and 3, and somewhat too low in experiments 6, 7, 8, and 9. In experiments 4 and 5, however, the animals were the fattest in the series-; and it will be seen further on, that the high estimates of fat in the increase in their case are pro- bably not too high — indeed, in experiment 5 even somewhat too low. It might be supposed that, at any rate in the case of experiment 1, the results would be admirably adapted for our present purpose. But that experiment was made in 1850, that is nearly forty-five years ago, and before we had acquired sufficient evidence against the view then prevailing — namely, that the increase of the fattening animal was largely depend- ent on the richness of the food in nitrogenous constituents ; and everybody having experience in the fattening of pigs will admit that, in this case, the food was much more highly nitrogenous than is recognised as most favourable for the fattening of the animal. In fact, it is seen that the propor- tion of the crude non-nitrogenous to 1 of crude nitrogenous substance in the food, was only 3.6 instead of about 6, as in barley-meal. There was, therefore, an excess of nitrogenous Hxcess of substance consumed. ^bstcTT" Eeferring to the middle division of the table, the calculated in the food. results show that, for 100 increase in live-weight, 100 of nitrogenous substance was consumed in the food. Of this, it is estimated that only 7.8 parts were stored up in the increase, leaving 92.2 parts available for the possible formation of fat. 292 THE EOTHAMSTED EXPERIMENTS. Results liable to correction. It is next seen, that the 100 of increase was estimated to contain 63.1 parts of fat, whilst the food supplied only 15.6 parts, leaving therefore, at least, 47.5 parts to be produced within the body. The figures show that this would require 36.6 parts of carbon, whilst 44.0 parts are estimated to have been available from the nitrogenous substance of the food; leaving, therefore, according to the mode of calculation adopted, 7.4 parts more carbon available than were required for the formation of the produced fat. Or, as shown in the bottom division of the table, for 100 carbon in the estimated newly formed fat, 120.2 parts were available from the nitro- graaous substance consumed in the food. Here, then, the calculations aifford no evidence that fat must have been produced from carbohydrates. But, as already explained, the mode of estimate adopted assumes the whole of the ready-formed fat in the food to have been stored up, and the whole of the carbon of the nitrogenous substance, beyond that in the animal increase, and in the urea formed, to have been utilised for fat formation. Neither of these assumptions is, however, admissible ; and it will be seen further on, when due correction is made in regard to these points, that, even in this experiment, with so abnormally high a proportion of nitrogenous substance in the food, it is pretty certain that some of the produced fat must have had its source in the carbohydrates. In experiment 2, the food consisted of bean-meal, lentil- meal, bran, and maize-meal, each given separately, and ad libitum ; and in experiment 3, of an equal mixture of bean- meal and lentil-meal, also given ad libitum. It is seen that, in both cases, the proportion of crude non-nitrogenous to 1 of crude nitrogenous substance in the food was even lower than in experiment 1 ; being, in experiment 2, 3.3, and in experiment mt/rogen- 3, only 2.0, against 3.6 in experiment 1. Here again, as might be expected, with so high a proportion of nitrogenous substance in the food, the calculations show that there was more than sufficient carbon available from the nitrogenous substance of the food for the formation of all the fat that was estimated to be produced. Experiments 4 and 5 show a very different result. In experiment 4, the food consisted of maize-meal alone, and in experiment 5 of barley-meal alone, in each case given ad libitum. In America especially, maize-meal is largely used for the fattening of pigs, almost, if not quite alone ; and in our own country barley-meal is (undoubtedly recognised as the most appropriate fattening food of the animal. It is seen that, in experiment 4, with maize-meal, the proportion of OILS sub- stance ogam in excess. Appropri- atefood for pigs. crude non-nitrogenous to 1 of nitrogenous substance in the FEEDING OF iNIMALS. 293 food was 6.6, and in experiment 5, with barley-meal, it was 6.0 ; or, in both cases, nearly that which is recognised as appropriate in the fattening food of the animal, but rather low in nitrogenous substance. Accordingly, the calculations show much less nitrogenous substance consumed for the production of 100 increase in live-weight, and much less left available for fat formation, after deducting the amount estimated to be stored up in the increase. Then, as to the fat, the animals were undoubtedly much fatter than the analysed "fat" pig. Deducting the amounts of fat supplied in the food from that in the increase, there remained, in the one case 52.7, and in the other 58.8 parts, formed within the body, requiring in the first case 40.6, and in the second 45.3 of carbon ; whilst the amounts of carbon estimated to be available from the nitrogenous sub- stance of the food were only 24.7 and 27.4 parts; leaving, in the one case 15.9, and in the other 17.9 parts, to be provided from other constituents of the food. Or, if the calculations are made for 100 carbon in the, estimated newly- formed fat, the figures show that, in one case 39.2, and in the other 39.5 per cent, of the total carbon of the produced fat must have been derived from other constituents of the food. In other words, even on this mode of calculation, nearly 40 iOper cent per cent of the newly-formed fat must have had its source ^^* derived in the carbohydrates. "We shall see further on, that even a hydrates. considerably larger proportion still must in reality have been so derived. The peculiarity of the experiments 6, 7, 8, and 9 was, that the food contained less ready-formed fat than in any of the other cases, and that a large proportion of the non-nitrogenous substance supplied was in the form either of pure starch, pure sugar, or both. In experiments 6, 7, and 8, a fixed quantity of lentil-meal and bran, averaging 3 lb. 3 oz. of lentil-meal, and 9 oz. of bran, was given per head per day ; and, in addi- tion, in experiment 6 sugar ad libitum, in experiment 7 starch ad libitum, and in experiment 8 sugar and starch, each sepa- rately, ad libititm. Lastly, in experiment 9, lentil-meal, bran, sugar, and starch, were each given separately, and ad libitum. It will be seen that the proportion of crude non-nitrogenous to 1 of crude nitrogenous substance was 4.1 in experiments 6 and 7, 4.7 in experiment 8, and only 3.9 in experiment 9 ; that is, the food contained a higher proportion of non-nitro- genous Substance than in experiments 1, 2, and 3, but con- siderably lower than in experiments 4 and 5. Accordingly the final result of the calculations is intermediate between that for the other two series. 294 THE EOTHAMSTED EXPEEIMENTS. Fat again shown to he derived from carbo- result. Vait criti- cises the Rotham- sted results. To go a little into detail, it is seen that, /or 100 increase in live-weight, the amount of nitrogenous substance estimated to be available for fat-formation was, in this series, intermediate between that in the other two. With much less fatty matter supplied in the food, the amount or fat estimated to be newly-formed was about the same as in the other cases. The amount of carbon estimated to be available for faWormation from the nitrogenous substance of the food was, in each case, notably less than the amount required for the production of the newly-formed fat. The indication is, therefore, that in each case a considerable proportion of the produced fat must have had its source in other than the nitrogenous constituents of the food. The bottom division of the table shows that, reckoned for 100 carbon in the estimated newly-formed fat, in the first case 18.9, in the second 18.8, in the third 25.2, and in the fourth 14.1 per cent, or, on the average, about 20 per cent of the whole must have been derived from other sources — in fact from the carbohydrates. Nor can there be any doubt that the figures under-estimate the proportion of the produced fat which could not have had its source in the albuminoids of the food. The general result of the whole series of experiments is, then, that when the food of the fattening animal contains an abnormally high amount and proportion of nitrogenous sub- stance, enough of it will probably be available for the possible formation of all the fat produced in the body ; but that, when the amount and proportion of such substances in the food are only normal, or low, there will remain a large proportion of the produced fat which could not have had its source in the proteids, and must have been derived from the carbo- hydrates. Eeferring to our results and conclusions as given above. Professor Voit, in a paper which he published in 1869,^ admits that in the experiments in which there was only a medium albuminoid supply in the food, there was, as the figures stand, a considerable deficiency for the formation of the fat produced, and a still greater deficiency when the relation of the nitro- genous to the non-nitrogenous constituents was lower still; and hence it would appear that in these instances a con- siderable amount of fat had been derived from the carbo- hydrates. Still, he says, he cannot allow himself to consider that a transformation of carbohydrates into fat is proved thereby. He says he has not been able to get a clear view of the experiments from the figures recorded, and suggests ^ Zeitschrift filr Biologie, Band 5. FEEDING OF ANIMALS. 295 several possible sources of error. He proposed that new ex- periments with geese and with pigs should be made ; and, in a subsequent conversation one of us had with him, he ex- pressed his willingness to undertake a conclusive experiment with pigs. Weiske and Wildt ^ did undertake an investigation with Bxperi. pigs to determine the point. But one animal was fed on food ^^^^^ so rich in nitrogen that it suffered in health, and the experi- wiidt. ment had to be discontinued ; and the other on food so poor, that it fattened extremely slowly, and hence, at the conclu- sion, calculation showed that there was enough of the con- sumed nitrogenous matter available for fat formation to cover the whole of the fat which had been produced. Professor Emil von Wolff, in his work entitled Die Wolff's rationelle Fiitterung der landioirthschaftlichen Nutzthiere, auf '»«'"''• Grundlage der neueren Thier - physiologischen Forschv/ngen, published in 1874, assumed that albumin was probably the exclusive source of the fat of the fattening herbivora of the farm. But he made the reservation, that the amounts of increase produced in relation to constituents consumed, which common observation showed may be obtained with pigs, and still more the results recorded of some direct experiments with those animals (presumably our own), are almost incom- prehensible without assuming the direct concurrence of the carbohydrates in the formation of the fat. Nevertheless, he considered that such evidence was inconclusive, and that experiments with pigs should be made in a respiration appara- tus to settle the question. After the inconclusive results of Weiske and Wildt, and Me-caicu- the publication of Professor Wolff's views, as above quoted, ^^q^^. we carefully reviewed and re-calculated many of the results of sted experi- our feeding experiments, including some with oxen and with '"*"**• sheep as well as those with pigs, in order to satisfy ourselves whether any doubt could be entertained of the .views we had previously advocated. The result of this examination, so far as the ruminants Sourm of were concerned, was to show that, owing to the comparatively ■{^^^^ small amount of increase obtained with them from a given amount of constituents consumed, the quantity of nitrogenous substance passed through the system for the production of a given amount of increase was, in most cases, so large as to admit of the assumption that the whole of the fat formed might have had its source in transformed nitrogenous matter. As will be seen further on, however, some of the experiments Sonne of with sheep showed that, at any rate part of the fat stored up f^^ 1 Zeitschrift filr Biologie, Band 10, 296 THE EOTHAMSTED EXPERIMENTS. mtist have had some other source than the fatty matter and the proteids of the food. Views as to The reconsideration of the results with pigs fully confirmed '^mnZS *^® '^^'^'^ *^^*' ^'^ many cases, much more fat had been pro- duced than could possibly have been derived from trans- formed albumin of the food. We concluded, therefore, that we were not called upon to institute new experiments ; and decided instead, again to direct attention to the results which had already been published. Paper read Accordingly, we gave a paper on the subject in the Section tumi"Z' fo^ Agriculture and Agricultural Chemistry, at the meeting of 1876. the Naturforscher Versammlung, held at Hamburg, in 1876, at which there were present a number of the chief agricultural chemists of Germany. The results given in Tables 69 and 70 were discussed, and it was pointed out that, even according to the mode of calculation adopted, which would imply about 62 parts of fat to be producible from 100 parts of nitrogenous substance, the experiments 4 and 5, in which the proportion of the non-nitrogenous to the nitrogenous constituents in the food was the most appropriate for fattening, showed that about 40 per cent of the produced fat could not have had its source in the nitrogenous substance consumed ; whilst if, according to Henneberg and Voit, it were assiimSd that 100 parts of albumin can at most yield 51.4 of fat, the results would be much more striking still. They Would, of course, be still more so if, as has more recently been estimated, only 42 instead of 51.4 parts of fat can be derived from 100 of albumin. It was next considered what amount of error in the esti- mates would have to be admitted to turn the scale, and to show that the whole of the produced fat might have been derived from the albuminoids of the food. After going into considerable detail on the point, it was Concluded that any such range of error was simply impossible. A test ex- Further, it was maintained that, in the case of pigs fatteti- ing rapidly on their most ' appropriate fattening food, the amount of fat stored up in proportion to the amount of fat and nitrogenous substance consumed was so large that the question of whether or not the carbohydrates contribute to fat-formation might be conclusively settled by a properly conducted feeding experiment with those animals, without any analysis of the fseces or the urine, or any determination 6f the products of respiration. It was stated that it was only necessary to select two animals of a breed of good fattening quality, and as nearly alike as possible in Character and in weight; a convenient size and weight being — say about 90 lb. per head. Each should then be fed with ground barley of good quality, giving it, by degrees, until both weighed about FEEDING OF ANIMALS. 297 100 lb. Thea slaughter one, and determine its total amount of nitrogenous substance and of fat. Continue to feed the other with barley meal (and water) exclusively, as much as it will consume, until it reaches a weight of about 200- lb. ; then slaughter and analyse it as the first. The quantity and composition of the food must, of course, also be determined. Such an animal Would probably consume about 500 lb. of barley, and increase in live-weight from 100 to 200 lb., in from eight to ten weeks — more or less, according to the quality of the animal, the quality of the food, and other conditions. It was desirable that the animals selected should have been feeding on fairly good food previously, so that the transition to full fattening food should not be too sudden. It was also, of course, desirable, that the experiments should be made in duplicate if possible. In the discussion which followed, Professor Henneberg, who Professor was, we believe, the first to have a Pettenkofer respiration fj.^' apparatus constructed for experimenting with the larger opinion. animals of the farm, and had perhaps, at that time, conducted more experiments on feeding than any other agricultural chemist in Germany, said he did not doubt the formation of fat from carbohydrates ia the case of pigs. He added, that probably sooner or later the carbohydrates would be restored to their former position so far as fat-formation in other animals was concerned, for already some experiments had shown that such formation was quite close upon the limits of the amount possibly derivable from the fat albuminoid matters of the food. Professor Emil von Wolff also spoke in the same Wolff's sense so far as pigs were concerned. opimion. Since that time, experiments have been made on the sub- ject in Germany with various animals; but, even in those with pigs, the conditions above indicated as desirable with a view to obtaining decisive results the most easily, were not followed. Experiments were made with cows by Voit at Munich,^ by Experi- Wolff at Hohenheira,2 and by G. Ktihn at Mockern. ^ In those ^««** »™ at Munich and at Hohenheim, the amount of fat in the food, with cows. and that possibly derivable from the albumin consumed, Very nearly corresponded with the amount of fat in the milk. In the experiments at Mockern, however, a small excess of milk- fat was produced. None of these experiments, therefore, afforded evidence of the formation of fat from the carbo- hydrates. •^ ZeUschrift fij/r Biologie, 1869, p. 113. ^ Die Versuchsstatvmen, Hohenlieim,, Berlin, 18?0, p. 50 ; also M. Fleiacher in VirefUnv's ArcMnfii/r PaMdeg, Anat., Band 51, 1870. \ Versuchsstationen, 1869, Band 12, p. 451. 298 THE EOTHAMSTED EXPERIMENTS. In experiments made by Kern and Wattenberg, at Gptting- en ^ with sheep of various ages, in ten cases the fat stored up vHthshZp. fell short by 24 to 64 per cent of that which could have been derived from the fatty matter and nitrogenous substance con- sumed. In one experiment, however, one animal was killed and the initial composition determined, and the other was fed for ten weeks, and the composition and digestibility of the food were determined. The results showed that 29.4 per cent of the fat stored up must have been derived from other sources than the fat and the albumin of the food ; and, even making all allowance for possible error, it was concluded that fat must have been derived from the carbohydrates consiimed. In other experiments at Gottingen, by T. Pfeiffer and Lehmann,^ a similar result was obtained with a sheep fed with a considerable quantity of sugar. Wolff's ex- In an experiment made by Wolff at Hohenheim,* a young ^a»^s P^S ^^^ ^^^ ^°^ -"-^^ ^^y^ "viith. barley and maize-meal, with the addition of pure starch. The constituents digested were deter- mined. Eeferring to the results, Wolff says that, having re- gard simply to the amounts of constituents consumed, and of increase produced, it is scarcely possible to suppose that the quantity of fat which must have been stored up could have been formed without the co-operation of the carbohydrates. He points out that fat equal to only 29 per cent of the increase in live-weight could have been produced from the fat and the albumin of the food; and in this calculation he takes the whole of the albumin as available, without reckoning any to have been stored up. He adds that, according to the percent- age of fat in increase in the Eothamsted experiment No. 1, Rotham- there must have been 60 per cent or more. According to our wiwTe^ own calculation of Wolff's results, it seems probable that periment. about 60 per Cent of the total fat in the increase must have been derived from carbohydrates. It is particularly to be observed that, in the case of this experiment, Wolff concluded that the formation of fat from the carbohydrates might be considered established, not only without any respiration apparatus, but even without any direct determination of fat in the animal. Various ex- Wolff quotes the results of experiments with pigs at SS^ Moscow, by Tschirwinsky, in 1880-81 and in 1881-82.* It Rotham- was estimated that in the one case 61.6 per cent, and in the other 76.9 per cent of the fat of the increase must have had its source in the carbohydrates of the food. ^ Jowrn. fur Lomdw. Jahrg. 26, p. 549, ^ Journ.. fur Landw. 1885, Band 33, p * Die raUonelle Fiitterung der lamdwi _ _, 1888, p. 48. *. Fermchsstationen, 1883, Band 29, p. 317, 2 Journ.. fur Landw. \i9,5, Ban(f 33, p. 337 ; also 1886, Band 34, p. oo. ' Die rationelle Fiitterung der lamdwirthschaftlicJien Nutzthiere, 5*« Aufl., FEEDINS OF ANIMALS. 299 In an experiment made with a pig at Vienna by Meissl and Strohmer,^ it was estimated that 82.2 per cent of the stored-up fat must have been derived from the carbohydrates consumed. At Proskau, Weiske and B. Schulze^ made experiments with geese ; and they concluded that in one case 13 per cent, and in the other 17.6 per cent of the stored-up fat must have been derived from carbohydrates. At Peterhof, near Eiga, Chaniewski^ experimented with geese ; and from the results concluded that in one case 71.1 per cent, in another 78.6 per cent, and in a third 86.7 per cent of the stored-up fat must have been derived from carbo- hydrates. Wolff also quoted recent experiments by A. von Planta and Erlenmeyer, at Munich, with bees,* in which it was proved that wax had been formed from sugar. Lastly, in 1880-81, Soxhlet made experiments with three Recmt ex- pigs, at the Agricultural Experiment Station at Munich.^ 1^™^^ The animals were five to six months old ; they were fed for a preliminary period of 321 days, with equal but limited amounts of barley-meal. No. 1 was then killed and analysed ; No. 2 was fed for 75 days, and No. 3 for 82 days, with 4.4 lb. steamed rice per head per day for most of the time, but only three-fourths as much afterwards. Meat extract was also given for 50 days. Finally, Nos. 2 and 3 were killed and analysed. Calculation showed that the increase of No. 2 contained 14.19 per cent of nitrogenous substance, and 25.80 per cent of fat ; and that of No. 3, 7.25 per cent of nitrogen- ous substance, and 57.23 per cent of fat. That is, the increase of No. 3 contained only half as much nitrogenous substance, and more than twice as much fat, as that of No. 2 ; and even the higher proportion of fat (57.23) is low compared with that which would be obtained with animals of good breed, and rapidly fattened on appropriate food given ad libitum ; whilst the composition of the increase of No. 2, both as to nitrogenous substance and fat, can hardly be called that of fattening increase at all. Still, calculation showed that, of the total fat in the increase of No. 2, 79.38, and in that of No. 3, 81.84 per cent, must have been derived from the carbohydrates of the food. Notwithstanding the extraordinary difference in the com- position of the increase of Soxhlet's pigs No. 2 and No. 3, ^ Ber. Acad. Wissens., Wein, 1883, Band 88, p. iii. 2 and ' E. Wolff, Die raiionelle Futtervng der landwirthschaftlichen Nwtz- thiere, St" Aufl., 1888, p. 50. * Bienemeitung 11. A. Schmidt, 1878, p. 181. 5 Zeits. d. kmdiv. Vermis in Bayern, 1881, pp. 423-436. 300 THE EOTHAMSTED EXPERIMENTS. after having been fed alike, he says that only our experiment No, 1 is admissible for calculation, because it is only in that case that the initial and final composition was determined in parallel animals. He, in fact, accepts our least conclusive result, obtained with food abnormally rich in nitrogenous substance, and repudiates our most conclusive experiments with appropriate fattening food. Accordingly he maintains that we had only shown the probability of the formation of fat from the carbohydrates, and that his own results as above were the first to prove it. The discussion of the results of the nine experiments recorded in Table 70 must have sufficed to show that in some of them a very large proportion of the fat of the increase must have been produced from the carbohydrates. The mode of calculation adopted showed, however, a maximum amount of the fat of the increase to have been possibly derivable from fatty matter in the food, a maximum amount of the nitrogen- ous substance of the food to be available for fat-formation, and a maximum amount producible from a given amount of nitrogenous substance ; and hence a minimum amount neces- sarily derived from carbohydrates. But, as the results so calculated, and discussed with due reservation on these points, are those upon which we have for so many years maintained that the formation of fat from the carbohydrates has been proved, and as it is those results, and the conclusions drawn from them, that have instigated so much subsequent investigation leading to the confirmation of our views, it seemed desirable prominently to direct attention to the evi- dence as so brought out. We have, however, as already said, long ago re-calculated many of our feeding expesriments, making allowance, as far as practicable, for the probable amount of indigestible and neces- sarily effete matters of the foods. We have also, as referred to at pp. 280-283, arranged tables founded on our direct analy- tical results on the ten animals, showing the probable average percentSige composition of the different descriptions of animal, each at eight gradationary points from the store to the very fat condition, and have applied the factors thus obtained, not only for the calculation of the composition of the increase in a number of cases of ordinary practice, and of direct experiment, but also for the re-calculation of some of the results to which Table 70 relates. Accordingly, in the Tabiell next Table (71) are given the results obtained' in experi- ment No. 1, which were inconclusive according to the orig- inal mode of calculation, and also those obtained in experi- ments 4 and 5., which, even as originally calculated, could FEEDING OF AKIMALS. 301 «1 ,-iM S'3 I— I CO ^ CO CD ^ to 03 00 ■* <» t>. C4 (N 00s ?o U3 t-r-l K3 CO 10 00 I-H »o ■* t^ CT 10 C^ CO o-* CO CO l>.»CJ c^ H M 10 10 t^(N »Q t^ 06 C4 01^ CO 00 t^ «> j-t U3 CO I-t K-I ^ 00 (N rH r-cM 00 o<>l Cfl 0» - S ™ 2 o C3 :3 (j! o l-tlJ J; fr4 rH <1> S » 302 THE EOTHAMSTED EXPEKIMENTS. Basis of re-calcu- lation. calcula- tions. MesiUts from rich food. leave no doubt of very considerable formation of fat from the carbohydrates. All these re-calculations are in the first place based on the assumption, since generally adopted by others, that 100 nitro- genous substance can at the most yield 51.4 of fat, instead of nearly 62, which would be the figure according to the original plan of calculation adopted in the construction of Table 70. Then, each experiment is now calculated three ways : — first, on the assumption that the whole of the fatty matter and nitrogenous substance of the food were digested ; secondly, supposing that only 90 per cent, and thirdly that only 80 per cent was digestible and available. Lastly, in the case of experiments 4 and 5, after very carefully considering the weights and character of the animals, and the duration of the jfattening period, the initial and final composition have been taken, not as in Table 70, the same as in experiment 1, but the initial at a composition three-eighths in advance from the store to the fat condition, and the final composition at a quarter in advance of fatness, compared with the fat pig of experiment 1. It is worthy of remark, that this carefully re -considered independent mode of estimate gives almost precisely the same percentage of nitrogenous substance, and precisely the same of fat, in the increase in experiment 4, as in the former estimate — namely, now 5.4 instead of 5.3 per cent of nitrogenous substance, and in both cases 79 per cent of fat, the animals being all very fat. Again, the new mode of calculation gives for experiment 5, 6.4 per cent of nitro- genous substance, and 72.3 per cent of fat in the increase, in- stead of 6.5 and 71.2 per cent as formerly adopted. Let us first refer to the results of experiment 1, in which parallel animals were analysed, but in which, as has been pointed out, the food was much more highly nitrogenous than is appropriate in the fattening food of the pig. Those given in column 1, in which it is assumed that the whole, both of the nitrogenous substance and of the fat of the food, was digestible and available, show that, when we now reckon only 51.4 instead of about 62 parts of fat to be derivable from 100 nitrogenous substance, even this experiment indicates that the fat in the food, and that derivable from the nitrogenous substance consumed, were scarcely sufficient to cover the whole of the fat of the increase. Obviously, too, if it be as- sumed, according to the more recent estimate, that only about 42 parts of fat can be derived from 100 of albuminoid sub- stance, there would then, even in this experiment, with such abnormally high nitrogenous food, be a considerable forma- tion of fat from carbohydrates. Turning to the results in the second column, which are FEEDING OF ANIMALS. 303 calculated on the assumption that only 90 per cent of the nitrogenous substance, and 90 per cent of the fatty matter, of the food would be digested, it is seen that — for 100 increase in live-weight 6.8 parts, for 100 total fat in the increase 10.8 parts, or for 100 newly-formed fat 13.9 parts, must have been derived from carbohydrates. Lastly, in regard to experiment 1, reckoning only 80 per cent of the nitrogenous substance and fat of the food to have been digested and available, the result would be that 13.5 of the 63.1 parts of fat in 100 of increase must have had some other source than fat and nitrogenous substance of the food ; or reckoned for 100 total fat in the increase, 21.4 parts, or for 100 newly formed fat 26.7 parts, must have been derived from carbohydrates. In regard to the alternative assumptions that only 90 or Portum of only 80 per cent of the nitrogenous and fatty matters of the ^S*?^?^ food were digested, it may be stated that in Wolffs tables, matters published in Mentzd und v. Lengerke's landvArthschaftlicher digested. XaleTider for 1890, he reckons 88 per cent of the nitrogenous substance of beans, 89.9 per cent of that of lentils, 77.9 per cent of that of bran, 79.2 per cent of that of maize, and 77 per cent of that of barley, to be on the average digested ; and of the fatty matter of these foods, he reckons 87.5 per cent of that of beans, 84.6 per cent of that of lentils, 70.6 per cent of that of bran, 85.1 per cent of that of maize, but the whole, or 100 per cent, of that of barley to be digestible. So far, there- fore, as experiment 1 is concerned, according to Wolff's factors the truth would lie somewhere between the results supposing 90 and those supposing 80 per cent digested. Even in this experiment, then (Xo. 1), there is clear evi- Clear evi- dence of the formation of fat from the carbohydrates, when ^^^^, deduction is made for indigestible nitrogenous and fatty mat- drates ters consumed, and when it is reckoned that only 51.4 parts ■^^*'^ of fat may be produced from 100 albuminoid substance. Obviously, if only 42 parts of fat, as assumed by some, can be formed from 100 albumin the evidence is clearer stUL Turning now to experiment 4, in which the food was maize- meal alone, given ad libitum, and the relation of non-nitro- genous to 1 of nitrogenous substance was much higher than in experiment 1, and much more appropriate for the rapid fattening of the pig, the results are much more decisive. They stm more were indeed quite conclusive as originally calculated, without <^«<^'«- the emendations now adopted. The results, even as given in the first of the three columns, in the calculation of which it is assumed that the whole of the nitrogenous substance and fat of the food were digested and available, show that — for 100 increase in live-weight 26.2 304 THE pOTHAMSTED EXPERIMENTS. Percentages parts of fat, for 100 total fat in increase 33.2, and for 100 tJ^lohy.°^ newly-formed fat 49.7 parte, must have been derived from Arates. carbohydrates. Eeckoning, as in the second column, that 90 per cent of the nitrogenous substance and fatty matter consumed were digest- ible and available, the calculations show that — for 100 increase in live-weight 31.7 parts of fat, for 100 total fat in increase 40.1 parts, and for 100 newly-formed fat 57.3 parts, would be derived from carbohydrates. Or, reckoning as in the third column, that only 80 per cent of the nitrogenous substance and fat of the food were digested and available, the results show that — for 100 increase in live-weight 37.3 parts of fat, for 100 total fat in the increase 47.2 parts, and for 100 newly- formed fat 64.3 parts, or nearly two-thirds, of the total pro- duced fat, would have its source in the carbohydrates. It may be observed that, in the case of this experiment with maize, the results given in the third column would very nearly accord with those which would be obtained if Wolff's average percentages of digestible had been adopted. Results Let us now refer to the results of experiment 5, in which Meaibum^ the food was barley-meal alone, given ad libitum, and the inoid ratio, albuminoid ratio was nearly that recognised as most suitable for the rapid fattening of the pig. The first of the three columns, calculated on the assumption that the whole of the nitrogenous substance and fat consumed were digested, shows that under such conditions there would be — for 100 increase in live-weight 30.3 parts of fat, for 100 total fat in increase 41.9 parts, and for 100 newly-formed fat 50.6 parts, or about half, that must have been derived from other constituents than the fatty matter and nitrogenous sub- stance of the food. The results in the second column, calculated on the as- sumption that 90 per cent of the fatty matter and nitro- genous substance were digested, show that — in lOO increase in live-weight 34.8 parts of fat, in 100 of total fat in increase 48.1. parts, and of 100 newly-formed fat 57.0 parts, must have been formed from carbohydrates. Lastly, the results in the third column; reckoning pnly 80 per cent of the nitrogenous substance and fat to be digested, show that on this supposition — of 100 increase in live-sveight 39.4 parts of fat, of 100 total fat in increase 54.5 parts, or of 100 newly-formed fat 63.1, or again nearly two-thirds, must have been derived from carbohydrates. Evidence So much for the evidence of results relating to pigs, in their ZUddSu- bearing on the question of the sources of their fat, when fed we. on their appropriate fattening food. It is cumulative and decisive that, at any rate, a large proportion of the stored-up FEEDING OF AOTMALS. 305 fat must have its source in other constituents than the 'fat and nitrogenous substance of the food — in other words, in the carbohydrates. The Eooperiments at Bothavisted with Sheep. It has been pointed out that, compared with pigs, there is Expm.- with ruminants a much smaller amount of increase obtained, T^t^*' -,. ., .,. . . ■; she^ less in proportion both to their weight withm a given time, and amOuswe. to a given amount of food passed through the body; that there is also a much larger amount of necessarily effete matter in their food; and that, therefore, the result of cal- culations of feeding experiments with them in regard to the question of the sources in the food of the fat stored up in the body are less conclusive. It wUl, nevertheless, be of interest to adduce some direct experimental evidence on the point. Some time after the discussion at Hamburg in 1876, two Rotham- sets of experiments made at Eothamsted with sheep, in which !^^^^^ the concentrated foods were barley or malt, and in which, s!w^. therefore, the amount and proportion of nitrogenous sub- stance consumed was low, were selected for calculation. The first series comprised five pens, with four or five sheep in each. The experiments had been made in the spring of 1849, and extended over a final fattening period of ten weeks. In each pen barley or malt was given in fixed quantity per head per day, and in each mangels were given in addition, ad libitum. The second series also comprised five pens, but with twelve sheep in each. The experiments were made in the winter of 1863-64, and they extended over a final fattening period of twenty weeks. The animals were at an earlier stage of progress at the commencement, and not quite so mature at the conclusion, as those of the other series. In each pen barley or malt was given in fixed quantity per head, in each clover-chaff also in fixed quantity, and in each roots were given ad libitum — Swedish turnips during the first sixteen weeks, and a mixture of one-fourth swedes and three-fourths mangels during the last four weeks of the twenty. The results of these two series of experiments with sheep, calculated to show their bearing on the question of the sources of the fat stored up by the animals, are given in Table 72. It will be seen that the form of the table is, so far as the Taile 72 facts will allow, the same as has been adopted in the case «^i'^«»«<*- of the various experiments with pigs. A general, descrip- tion of the food of each series is given over the columns VOL. vn. V 306 THE EOTHAMSTED EXPERIMENTS. relating to the series, and at the head of each separate column is given a description of the limited food supplied to each pen. The results are calculated for 100 increase in live-weight. Referring to the upper division of the table, there are first shown — the amounts of nitrogenous substance (digestible) in the fixed food, the amounts in the increase, and the difference = the amounts available for fat-formation. Next are given — the amounts of fat in the increase, in the total food (digest- ible), and the difference = the newly-formed fat; the amounts derivable from the available nitrogenous substance in the fixed food, and the difference = the amount required to be produced from other sources. Then, in the lower division of the table are given, for each pen, the amounts of fat derivable from the nitrogenous substance of the roots, on the alternative assumptions that 50, 60, 70, 80, 90 per cent, or the whole, of that which they contain will be digestible and available for fat-formation. Pereemtage It should be further explained, that 80 per cent of the ^^f^"" nitrogenous substance of barley or of malt is reckoned as stance digestible and available for the purposes of the system. digested. -Wolffs estimates were— in 1874, 80 per cent ; in 1888, 77.3 per cent; and in 1890, 77 per cent. In malt-dust 80 per cent is assumed to be digestible, against Wolffs estimate of 80 per cent in 1874, and 82 per cent in 1888 and 1890. In clover-chaff two-thirds, or 66.7 per cent, of the nitrogenous substance is reckoned as digestible, against a range in Wolffs Tables, according to quality, from 51.4 to 69.9 per cent. In the cases of Swedish turnips and mangels, Wolff assumes the whole of the nitrogenous substance to be digestible and available, drawing no distinction in this respect between the amounts existing as albuminoids, as amides, or in other forms. To this point we shall have to refer in more detail presently. Percentage Then as to the fat of the foods : the percentage of it reck- o/fMy oned as digestible is that given in Wolff's tables of 1874. digested. In the case of barley he then reckoned only 68 per cent of the total to be digestible ; but more recently he has supposed the whole of it to be so. For clover-chaff his figures are the same at all three periods, as they are also for mangels. Resvlts. Let US now turn to the calculated results as given in the table, and first to those relating to the first series of five pens, in which the fixed food was either barley or malt, and the ad libitum food consisted of mangels only. As already said, the period of experiment comprised only the last ten weeks of fattening. Hence it commenced at a somewhat advanced stage of progress, and the animals were, at the conclusion, FEEDING OF AKT MA T.S. 307 i II It '3 a Si it II ■g X % fcO Barley (8), malt (45, and clover- chaff. ■* Malt and clover- chaff. Barley and clover- chaff. Ol ■2.. iH Barley and clover- chaff. 1 § a O & ! g 1 »o Malt and malt- dust, extra quantity. '^ Malt and malt- dust steeped. CO ^ CT ■a •-• 1 to lO Id 61.1 69.0 13.8 (NCO id CD la 2. ■si 11 .2 S " 1 • • 1 • i .2 .3 " i - 1 ■■§ s « — i -S3 -S 03 1 n 2 o El p^ W E-i o EH h- ( Pi « o o o I— I P3 n OO CO !>• C^ CO CO Cfl O CI *^ CD Tt< O OS t^lO CO i-C O OO CO ■* W o I— ( OS *^ CO -rjt (N CO O CO CO OS CO CD 0> 04 CO OS (N ■^ OO - 308 THE EOTHAMSTED EXPERIMENTS. a/oaildble. Fat avail- able. True al- buminoid nitrogen in mangels. Amines and ni- trates in mangels. probably fully as fat as, if not fatter than, the sheep Which had been analysed as "fat." Taking into account the weight and condition of the animals at the beginning and at the end, and the percentages of carcass and of inside fat in the live- weight, it is calculated that the increase over this short finish- ing period would contain 74 per cent of fat, and only 6.5 per cent of nitrogenous substance. On these assumptions the figures show that, after deduct- ing the estimated amount of nitrogenous substance in 100 of increase from the amount supplied in the fixed food, there remained in the different cases — 18.5, 16.8, 13.4, 18.5, and 21.4 parts, of nitrogenous substance available from the fixed foods for the formation of fat. Next as to the fat : — deducting the amount of the digestible fat 'supplied in the total food from the fat in the increase, there remain in the respective cases 63.7, 65.2, 64.4, 63.7, and 63.8 parts, which must have been newly-formed. There is next shown the amount of this which may have been de- rived from the available nitrogenous substance of the fixed food ; and it is seen that there remain — 54.2, 56.6, 57.5, 54.2, and 52.8 parts, out of the total of 74 in the 100 of increase, that must have been derived from other sources — ^in fact, either from the nitrogenous substance of the roots, or from the carbohydrates of the fixed food and the roots. The next question is, whether the nitrogenous substance of the roots could have yielded the amounts of fat indicated to have been produced from other spurces than the fat of the total food, and that derivable from the available nitrogenous substance of the fixed foods. Comparing the figures in the bottom line of the lower division of the table with those in the bottom line of the upper division, it is seen that, even on the impossible assumption that the whole of the nitrogen of the mangels existed in compounds of the same fat-forming value as the albuminoids, in neither of the five cases would the amount so available completely supply the amount required. The amount of true albuminoid nitrogen varies very much in different descriptions of roots, and in the same description according to season, maturity, &c. Thus,: at Eothamsted we have found it in mangels as low as 20.5 per cent of the total nitrogen under unfavourable conditions of growth and ripen- ing, and as high as 44.2 under favourable conditions. We generally assume in calculation that 40 per cent of the nitrogen of mangels will, on the average, exist as albumin- oids; and Wolff's average figure, as given in 1888, is 36.1 per cent. The amount existing as amides will probably in most crises va,ry from 40 to 50 per cent or more, whilst there FEEDING OF ANIMALS. 309 is frequently a considerable quantity as nitrates, the more the less ripe the roots ; and we have sometimes found the amount to be more than 10 per cent of the total nitrogen of the roots. It is clear, therefore, that even supposing as little as 50 per Percentage cent of the nitrogen of the roots to be available for, and cap- j^^™S able of, fat-formation, as assumed in the top line of the lower available division of the table, that amount would generally include ■^"^Z"'"^'^" other than albuminoid compounds. Nevertheless, Wolff in his tables assumes the whole of the nitrogen of roots to be digestible and available for the purposes of the system, since it has been shown that amides are transformed in the body and yield urea ; leaving, therefore, by-products of transfor- mation available for expenditure in respiration, and so pro- tecting the true albuminoids, or the carbohydrates. There is, however, so far as we are aware, no direct experi- Amides mental evidence yet at command, indicating that the by- %^'^„ products of the transformation of amides may directly con- tribute to the formation of fat. Eesults of independent ex- perimenters have, however, shown that the heat of combus- tion of asparagine for example, is only about, or little more than, half that of albumin ; and supposing that the amides do directly contribute to the formation of fat, it may safely be concluded that a given quantity of amide would yield very much less fat than an equal quantity of albuminoid. As bearing upon this point, it is to be borne in mind that, on the average, the amide bodies most frequently occurring in food-stuffs have a higher percentage of nitrogen than the albuminoids. Wolff estimates that whilst the nitrogen of food Wolff's should be X 6.25 to represent albuminoids, 5.5 would, on <^'"^*«- the average, be a more appropriate factor for calculating the amount of amide from that of the nitrogen. Further, he admits that so far as the nitrogen in potatoes, roots, and other food-stuffs exists as amides, the nutritive value of the food is reduced ; nevertheless, as has been said, in his tables he assumes the whole of the nitrogenous substance of roots to be digestible, and of equal value with the albuminoids. Then, again, as generally more or less of the nitrogen in Nitrates roots will exist as nitrates, it will so far not only have no '^^°^ food value, but it may be positively injurious. It may be added that, other things being equal, the higher the percen- tage of nitrogen in roots, the lower as a rule will be the pro- portion of it as albuminoids, and the higher that as amides, and as nitrates, &c. Further, in direct experiments at Eoth- amsted with sheep feeding on roots alone, it was found that whilst the animals even gained in weight on ripe roots, low ■''*^? ""^ in nitrogen, they actually lost on roots that were less ripe, rmu. 310 THE BOTHAMSTED EXPERIMENTS, J mount of fat produc- ible from imcertain. A la/rge proportion of increase from carbo- Nitrogen m clomr- hay. high in nitrogen, and doubtless containing a larger proportion of their nitrogen as non-albuminoid compounds. From these various considerations it is obvious that by no means the whole of the nitrogen of the mangels can be estimated as having existed in compounds which could, in their transformation, yield the amount of fat possibly deriv- able from true albuminoids. However, with the great varia- tion in the proportion of albuminoids and amides in roots, and the absence of exact knowledge as to the probable value, if any, direct or indirect, of amides for fat-formation, it is impossible to form any certain estimate as to which of the percentages given alternatively in the lower division of the table most probably represents the amount of fat producible from the nitrogenous substance of the mangels given ad libitum in each of the five pens of the first series of experi- ments with sheep. It is, however, quite safe to conclude that very much less than the whole would be so available ; and if we were to assume that of the nitrogenous constituents of the roots only the albuminoids would be available for fat- formation, the figures given in the top line of the lower division of the table, according to which it is reckoned that only 50 per cent of the total nitrogenous compounds of the roots would be capable of fat-formation, would in each case represent less than half the amount required. It is quite clear that, at any rate a large proportion of the increase estimated to be necessarily derived from other sources than the fat of the total food, and the nitrogenous substance of the fixed food, must have been derived from other sources than the nitrogenous substance of the roots ; in other words, it must have had its source in the carbohydrates of the fixed food or of the roots. Let us now examine the evidence of the results of the second series of experiments, on somewhat similar lines. As in Series 1, a fixed quantity of barley or malt was given in each pen, but now a fixed quantity of clover-chaff also. This introduction of clover-chafF into the fixed food brings us again face to face with the difficulty as to the estimation of the food-value of the amides. As already said, the calculation of the amounts of the nitrogenous substance in the clover- chaff which will be available are made on the assumption that 66.7 per cent of the total nitrogen will be digestible, and so available; and this figure agrees fairly with Wolff's estimates. But this amount includes amides as well as albuminoids. In Wolff's most recent tables he estimates that the proportion of the nitrogen of clover-hay existing in non-albuminoid compounds may range from 13.9 to 29.9 per cent of the whole, and probably be on the average about 19 FEEDING OF AinMALS. 311 per cent What proportion, however, of the two-thirds of the total nitrogenous substance of clover -hay, which is estimated to be digestible, will probably be non-albuminoid, there is no evidence to show. Under these circumstances we have, in the calculations, assumed the whole of the digestible nitrogenous substance of clover-hay to have the food-value of albuminoids. The figures will, therefore, doubtless overstate the amount of the nitrogenous substance consumed in the fixed foods, which is really available for nitrogenous increase and for fat-formation. Taking the figures as they stand, it is seen that, after deducting the amount of nitrogenous substance estimated to be stored up in 100 of increase from the amount suppUed in the fixed food, there remain in the several experiments 44.9, 43.6, 48.3, 48.4, and 51.1 parts, possibly available for fat- formation. Then deducting the amount of digestible fat in the total food from the fat estimated to be stored up in the increase, there remain — 55.9, 56.1, 56.0, 55.7, and 55.2 parts, which must have been newly - formed. Deducting from these amounts those producible from the available nitrogenous substance of the fixed foods, there remain — 32.8, 33.7, 31.2, 30.8, and 28.9 parts, to be formed from other sources. Com- paring with these amounts those derivable from the nitro- genous substance of the roots, assuming, as shown in the bottom line of the table, that the whole of it would have the same value for fat-formation as true albuminoids, it is seen that in four out of the five cases the fat so assumed to be formed would be less than that required. In these experiments the roots consisted chiefly of Swedish parogen turnips, and in only small proportion of mangels. The evidence ™ ««'«^««- at command leads to the conclusion that, in Swedish turnips a larger proportion of the total nitrogen exists as albuminoids, and a less proportion as nitrates, than in the more succulent mangels. We have found the proportion of the total nitrogen of Swedish turnips existing as albuminoids as low as 32.9, and as high as 55.8 ; and for the purposes of calculation we assume that, on the average, 45 per cent will be in that form. As large or a larger amount will, however, exist as amides than in mangels. It is evident, therefore, that even if we assume 50 per cent of the total nitrogenous substance of the roots consumed in this second series of experiments to have been of value for fat-formation, some amide wiU be included. But, even on the assumption that 50 per cent had the value of albuminoids for fat-formation, less than half the amount of fat required would be derivable from the nitrogenous substance of the 312 THE EOTHAMSTED EXPEEIMENTS. Conclu- sions with drates re- instated. roots. Assuming, however, that the amides of the roots would, as such, have a certain, though not an equal, value with the albuminoids for fat-formation ; or that, as protectors of other constituents, they may contribute indirectly to such formation, there would still remain a considerable amount of the produced fat to be derived from other sources — that is, from carbohydrates. Upon the whole, then, although the evidence of fat-forma- tion from the carbohydrates of the food is admittedly less direct in the case of sheep than in that of pigs, yet, when the foregoing results are carefully considered with due regard to the facts which have been discussed, no doubt can be enter- tained that there was a considerable formation of fat from carbohydrates in both of the series of experiments with sheep. And when it is borne in mind that neither of these series of experiments was arranged for the purpose of elucidating this particular question, it must b6 admitted that the results are more definite and conclusive than might have been anticipated. Nor can there be any doubt that if experiments were made with oxen under suitable conditions, they would yield equally conclusive evidence on the point. Indeed, as anticipated by Henneberg in the observations he made at Hamburg in 1876, we may consider that the carbohydrates are re-instated in their position in the formation of the fat of ruminants as well as in that of pigs. Views of German Summary on the Sources of the Fat of the Animals of the Farm. It was in 1865 — that is, nearly thirty years ago — that Voit first called in question the then very generally accepted opinions on the subject ; and as his evidence, derived from experiments with the omnivorous dog, accumulated, he more and more urged that his conclusions were equally applicable to herbivora. His views on the point came to be very generally adopted by agricultural chemists in Germany, and in 1874 Professor EmU von Wolff adopted them, but with some reservation so far as pigs are concerned, in his text- book, entitled. Die rationelle Futterung der landwirthschaft- lichen Nittzthiere ; auf Gruvdlage der neueren thierphysio- logischen Forsckwrigen. It has been already stated that, in the discussion at Ham- burg in 1876, Wolff more clearly admitted that pigs might behave exceptionally in the matter ; whilst Henneberg as- sumed that ruminants also would prove to be exceptions to the application of Voit's views. '■ Since' that date, a number of experiments have beeinjmade FEEDING OF AXDIALS. 313 in Germany and elsewhere, both with pigs and with rumi- nants, to elucidate the point ; and when the conditions of the experiments were suited to the object, the results contributed to the re-estabUshment of the conclusion that the carbo- hydrates play a very direct and important part in the fat- formation of the animals of the farm. Further, in the edition of Wolffs work published in 1888, Wolff a'nA he almost unreservedly admits the t6U of the carbohydrates J^^iT^^ in the formation of at least a great part of the fat not only of opinions. pigs but of ruminants. Indeed, some years previously, Voit himself had made substantial concessions on the point.^ It happens, however, that about 1880 Dr Armsby, now the Armsby's Director of the Agricultural Experiment Station at the Penn- catt^°^ sylvauia State College, published a work which has since Feeding, passed through several editions, entitled Manual of Cattle- Feeing ; a Treatise on the Laws of Animal Nvirition and the ChemistT~y of Feeding-Stuffs in their application to the Feeding of Farm-Animals, which was a very good digest, chiefly of the work done in Germany, on the subject. So far as the question of the sources of fat is concerned, it gives numerous tabular illustrations from Voit's work ; and it follows almost exclusively the views of Voit and of Wolff at that time. He, however, quotes results obtained both with pigs and with other animals, which, he admitted, indi- cate, according to the figures, the formation of fat from the carbohydrates. But he considered that the data at command were not sufficient to solve the problem; and, with Wolff, assumed that the question could not be satisfactorily settled without experiments in a respiration apparatus. He also considered that estimates founded on the composition of the increase of fattening animals as determined at Eotham- sted are uncertain. He, nevertheless, concluded that the carbohydrates may serve as a source of fat to swine, and under some circumstances to other animals also. It happens that Dr Armsby's book, founded to a great Prevailing extent on Wolffs earlier editions, was the only work of the ^^^ kind in the English language ; and hence, many of the rising young generation of agricultural chemists, both in this country and "Aemiste. in America, adopted the view that the albuminoids are the main, if not the exclusive, source of the fat of our farm stock, and of the butter of cows' milk. Under these circumstances it seemed desirable to consider in some detail, both the experimental evidence bearing upon the question, and the discussions which have taken place in regard to it, during the last quarter of a century or more. ^ Hermann's Sandbuch d. Fhysiologie, Band 6, Theil 1, von C. v. Toit, Leipzig, 1881. 314 THE EOTHAMSTED EXPERIMENTS. change of Points proved ir, Rotham- Conolu- Hans. It must be admitted that the importance of the carbo- hydrates as a direct source of much, if not of the whole, of the fat stored up in the animals which the farmer feeds has been clearly re-established. We have reason to believe that Dr Armsby himself adopts the charge of view ; though it will probably be some time before the truth is thoroughly recognised by the younger agricultural chemists. It was maintained by Voit and others, that to establish the formation of fat from the carbohydrates, it must be experi- mentally shown that the fat deposited was in excess of that supplied by the food, ■plus that which could be derived from transformed albumin. But it is obvious that the mere fact that the food contained enough nitrogenous substance for the formation of all the fat that had been produced, would of itself be no proof that that substance had been its source. It has been seen, however, that Volt's requirement was amply fulfilled in the Eothamsted experiments, both with pigs and with sheep ; and hence it must be admitted to be proved, that at any rate some of the stored-up fat must have had another source, which could only be the carbo- hydrates. In winding up the discussion, perhaps we cannot do better than reiterate the conclusions given in our paper on the sub- ject in 1866, namely : — 1. That certainly a large proportion of the fat of the her- bivora fattened for human food must be derived from other substances than fatty matter in the food. 2. That when fattening animals are fed upon their most appropriate food, much of their stored-up fat must be pro- duced from the carbohydrates it supplies. 3. That nitrogenous substance may also serve as a source of fat, more especially when it is in excess, and the supply of available non-nitrogenous constituents is relatively defective. Food and Milk Production. Milk production, and the dairy industry, are of such great and growing importance, and their various branches involve so many points of interest, that much time and space would be required to adequately discuss them. But when consider- ing what are the animal products of value derived from the consumption of food on the farm, it would obviously be in- appropriate not to refer, however briefly, to the question of milk production in some of its aspects. Attention must, however, be confined almost exclusively to the great difference in the demands made on the food — on the one hand for the production of meat, that is of animal, in- FEEDING OF AXDIALS. 315 crease, and on the other for the production of milk. But, as not only do cows of different breeds yield different quantities of milk, and milk of characteristically different composition, but individual animals of the same breed have very different milk-yielding capacity ; and whatever the capacity of a cow may be, she has a maximum yield at one period of her lacta- tion, which is followed by a gradual decline. Hence, in com- paring the amounts of constituents stored up in the fattening increase of an ox, with the amounts of the same constituents removed in the milk of a cow, we must assume a wide range of difference in the yield of milk. Accordingly, Table 73 shows — the amounts of nitrogenous TcAU 73 substance, of fat, of non-nitrogenous substance not fat, of ' ' " TABLE 73. — Comfabibon op the C!oNSTrnjE2«TS of Food caeried OFF r^ Milk, asd rs the Fattening Ixceease of Oxen. [lGaUon=10.S31b.] Nitro- genous sul)- stance. Fat, Non-nitro- genons substance not fat (sugar). Mineral matter. Total solid matter. IX MILK PER WEEK. If— lb. lb. lb. lb. lb. 4 qnarts per head per day 2.64 2.53 3.33 0.54 9.04 6 „ 3.96 3.80 4.99 0.81 13.56 8 ,. 6.28 5.06 6.66 1.08 18.08 10 „ 6.60 6.33 8.32 1.35 22.60 12 „ 7.92 7.59 9.99 1.62 27.12 14 ., 9.24 8.86 11.65 1.89 31.64 16 .. ., .. 10.56 10.12 13.32 2.16 36.16 18 „ 11.88 11.39 14.98 2.43 40.68 20 „ 13.20 12.65 16.65 2.70 45.20 IN INCREASE IN LIVE-WEIGHT PEK W EEK.— OJLEN. If 10 lb. increase If 15 lb. increase 0.75 1.13 6.35 9.53 ... 0.15 0.22 7.25 10.88 mineral matter, and of total solid matter, carried off in the weekly yield of milk of a cow, on the alternative assumptions of a produce of— ^ 6, 8, 10, 12, 14, 16, 18, or 20 quarts per head per day; and, for comparison, there is given at the bottom of the table, the amounts of nitrogenous substance, of &t, of mineral matter, and of total solid matter, in the weekly increase in live-weight of a fattening ox, of an average weight of 1000 lb. — ^first, on the assumption of a weekly increase of 10 lb., and, secondly, of 15 lb. The estimates of the amounts of constituents in the milk Percentage are based on the assumption that it will contain 12.5 per ^^^"" cent of total solids, consisting of 3.65 albuminoids, 3.50 mUk. 316 THE EOTHAMSTED EXPERIMENTS. butter-fat, 4.60 sugar, and 0.75 of mineral matter. The estimates of the constituents in the fattening increase of oxen are founded on the determinations at Eothamsted of such increase as already described. Varying Eeferring to the very wide range of yield of milk per head yields of ^gj. jja,y which the figures in the table assume, it may be '"* ' remarked that it is by no means impossible that the same animal might yield the largest amount — namely, 20 quarts, or 5 gallons, per day — near the beginning, and only 4 quarts, or 1 gallon, or even less, towards the end of her period of lactation. At the same time, an entire herd of, say. Short- horns or Ayrshires, of fairly average quality, well fed, and including animals at various periods of lactation, should not yield an average of less than 8 quarts, or 2 gallons, and would seldom exceed 10 quarts, or 2| gallons, per head per day, the year round. Basis of For the sake of illustration, then, let us assume an average comjpan- y^gj^ ^f ^^Y^ ^f ^0 quarts, equal 2| gallons, or between 25 and 26 lb. per head per day ; and let us compare the amount of constituents in the weekly yield at this rate with that in the weekly increase of the fattening ox at the higher rate assumed in the table — namely, 15 lb. per 1000 live-weight, or 1.5 per cent per week. Substances Thus, whilst of the nitrogenous substance of the food the carried off amount stored up in the fattening increase of an ox will be required Only 1.13 lb., the amount carried off as such in the milk would for fatten- foe 6.6 lb., or nearly six times as much. Of mineral matter, ^^^' again, whilst the fattening increase would only require ; about 0.22 lb., the milk would carry off 1.35 lb., or, again, about six times as much. Of fat, however, whilst the fattening increase would contain 9.53 lb., the milk would contain only 6.33 lb., or only about two-thirds as much. On the other hand, whilst the fattening increase contains no other non-nitrogenous sub- stance than fat, the milk would carry off 8.32 lb. in the form of milk-sugar. It may be observed that this amount of milk- sugar reckoned as fat would correspond approximately to the difference between the fat in the milk and that in the fatten- ing increase. Greater From the foregoing comparison, it is evident that the drain Sl"^""^ upon the food is very much greater for the production of milk mak than than for that of meat. This is especially the case in the im- ^ro^ti n PO'^t^^t ^^^^ 0^ nitrogenous substance ; and if, as is frequently assumed, the butter-fat of the milk is, at- any rate largely derived from the nitrogenous substance of the food, so far as it is so, at least about two parts of such substance would be required to produce one of fat. On such an assumption, therefore, the drain upon the nitrogenous substance of the FEEDING OF AMilALS. 317 food would be very much greater than that indicated in the table as existing as nitrogenous substance in the milk. To this point further reference will be made presently. We will next call attention to the amounts of food, and of Table ji certain of its constituents, consumed for the production of a s^^'"*^- given amount of milk. This point is illustrated in Table 74, which shows the constituents consumed per 1000 lb. live- weight per day, in the case of the Eothamsted herd, then of 30 cows, in the spring of 1884. TABLE 74 — Coustituexts coxscieed peb 1000 lb. Lite-Weight PER Day, fob Susiexaxce axd pob Milk PEODrrciioN. The EOTHAilSTZD HeBD OF 30 CoWS, SPBETG 1884. Total diy substance. Digestible. Nitro- genous substance. X'»n-mtro- genons substance (as starch). Total nit. and non-nit. snbstanee. 3.1 lb. Cotton-cake 2.7 lb. Bran .... 2.8 lb. Hay-chaflf . 5.6 lb. Oat-stniw-chaff . 62.8 lb. Mangels . lb. 2.76 2.33 2.34 4.64 7.85 lb. 1.07 0.33 0.15 0.08 1.01 lb. 1.50 1.09 1.18 2.21 5.73 lb. 2.57 1.42 1.33 2.29 6.74 Total. Required for sustenance . 19.92 2.641 0.57 11.711 7.40 14.35 7.97 Available for milk . In 23.3 lb. milk . ... 2.07 0.85 4.31 3.02 6.38 3.87 Excess in food . ... 1.22 1.29 2.51 PEE 1000 lb. LIVE- WEIGHT. Wolff. lb. 24 lb. 2.5 lb. 12.52 lb. 15.4 2 Albuminoid ratio 1 — 4.4. ^ Exclosive of 0.4 fat ; albuminoid ratio 1 — 5.4. On the left hand are shown the actual amounts of the different foods consumed per 1000 lb. live-weight per day ; and in the respective columns are recorded — first the amounts of total dry substance which the foods contained, and then the amounts of digestible nitrogenous, digestible non-nitrogenous (reckoned as starch), and digestible total organic sub- stance, which the different foods would supply ; these being calculated according to our own estimates of the percentage composition of the foods, and to Wolff's estimates of the pro- portion of the several constituents which would be digestible. 318 THE EOTHAMSTED EXPERIMENTS. Food con- sumedper 1000 lb. tionof ■matter for sustenance and milk- production. Wolff's estimate. Is milk-fat derived from al- or carbohy- drates, or botht The first column shows, that the amount of total dry sub- stance of food actually consumed by the herd, per 1000 lb. live-weight, per day, was scarcely 20 lb., whilst "Wolff's ^ estimated requirement, as stated at the foot of the table, is 24 lb. But his ration would doubtless consist in larger pro- portion of hay and straw-chaff, containing a larger proportion of indigestible and effete woody-fibre. The figures show, in- deed, that the Eothamsted ration supplied, though nearly the same, even a somewhat less amount of total digestible constituents than Wolff's. Of digestible nitrogenous substance, the food supplied 2.64 lb. per day, whilst the amount estimated to be required for sustenance merely is 0.57 lb. ; leaving, therefore, 2.07 lb. available for milk-production. The 23.3 lb. of milk yielded per 1000 lb. live-weight per day would, however, contain only 0.85 lb.; and there would thus remain an apparent excess of 1.22 lb. of digestible nitrogenous substance in the food supplied. But, against the amount of 2.64 lb. actually consumed, Wolff's estimate of the amount required for sus- tenance and for milk-production is 2.5 lb., or but little less than the amount actually consumed at Rothamsted. On the assumption that the expenditure of nitrogenous substance in the production of mUk is only in the formation of the nitrogenous substances of the milk, there would appear to have been a considerable excess given in the food. But Wolffs estimate assumes no excess of supply, and that the whole is utilised ; the fact being that he supposes the butter-fat of the milk to have been derived largely, if not wholly, from the albuminoids of the food. lb has been shown that although it is possible that some of the fat of a fattening animal may be produced from the albuminoids of the food, certainly the greater part of it, if not the whole, is derived from the carbohydrates. But the physiological conditions of the production of milk are so different from those for the production of fattening increase, that it is not admissible to judge of the sources of the fat of the one from what may be established in regard to the other. It has been assumed, however, by those who maintain that the fat of the fattening animal was formed from albuminoids, that the fat of milk must be formed in the same way. Dis- allowing the legitimacy of such a deduction, there do, never- theless, seem to be reasons for supposing that the fat of milk may, at any rate in large proportion, be derived from albuminoids. Thus, as compared with fattening increase, which may in ^ Landw. FuUerungslehre, 5te Aufl., 1888, pi 249. FEEDING OF AN TMAT.S . 319 a sense be said to be little more than an accumulation of MOk-pro- reserve material from excess of food, mUk is a special pro- ^^. duct of a special gland, for a special normal exigency of the pendent animal Further, whilst common experience shows that the ^^,^^ herbivorous animal becomes the more fat, the more, within upm nUro- certain limits, its food is rich in' carbohydrates, it points to f^"™ *^' the conclusion that both the yield of milk, and its richness in butter, are more connected with a liberal supply of the nitrogenous constituents in the food. Obviously, so far as this is the case, it may be only that thereby more active change in the system, and therefore greater activity of the special function, is maintained. The evidence at command is, at any rate, not inconsistent with the supposition that a good deal of the fat of milk may have its source in the break- ing up of albuminoids, but direct evidence on the point is still wanting; and, supposing such breaking up to take place in the gland, the question arises — what becomes of the bye- products ? Assuming, however, that such change does take place, the amount of nitrogenous substance supplied to the Eothamsted cows would be less in excess of the direct re- qidrement for nulk-production thau the figures in the table would indicate — if, indeed, in excess at all. The figures in the column relating to the estimated amount NoTi-nitTo- of digestible non-nitrogenous substance reckoned as starch, ^^^^f„ show that the quantity actually consumed was 11.71 lb., sustenance whilst the amount estimated by Wolff to be required was "''"^^^ 12.5 lb., besides 0.4 lb. of fat. The figures further show that, deducting 7.4 lb. for sustenance from the quantity actually consumed, there would remain 4.31 lb. available for milk-pro- duction, whilst only about 3.02 would be required supposing that both the fat of the milk and the sugar had been derived from the carbohydrates of the food ; and, according to this calculation, there would still be an excess in the daily food of 1.29 lb. It is to be borne in mind, however, that estimates of the Variations requirement for mere sustenance are mainly founded on the ^"^^!^I^ results of experiments, in which the animals are allowed only /or susten- such a limited amount of food as will maintain them without ''"^ either loss or gain when at rest. But physiological considera- tions point to the conclusion that the expenditure, independ- ently of loss or gain, will be the greater the more liberal the ration ; and hence it is probable that the real excess, if any, over that required for sustenance and mUk-production, would be less than that indicated in the table, which is calculated on the assumption of a fixed requirement for sustenance for a given live- weight of the animal Supposing that there really was any material excess of 320 THE EOTHAMSTED EXPEEIMBNTS. tion. Excess of either the nitrogenous or the non- nitrogenous constituents ajilSf ^ supplied over the requirement for sustenance and milk-pro- destina- duction, the question arises — Whether, or to what extent, it conduced to increase in live - weight of the animals, or whether it was in part or wholly voided and so wasted? It would obviously be of interest to trace the connection between variation in the quantity and composition of the food, and the quantity and composition of the milk yielded. But when the influence on the result, of breed, of varying character of individual animals, of period of lactation, and of other circumstances, are borne in mind, it will be seen that to treat the subject at all adequately would involve a great deal of detailed illustration and consideration, and occupy very much more space than could appropriately be devoted to it in this place. We must, indeed, limit further reference to the subject of milk-production to one more illustration, showing the influence of period of the year, with its char- acteristic changes of food, on the quantity and composition of the milk. The first column of the second division of Table 75 shows the average yield of milk per head per day of the Eotham- TABLE 75. — Percentage Composition oi" Milk each month op THE YEAR ; ALSO AVERAGE YIELD OF MiLK, AND OP CONSTI- TUENTS, PER HEAD PER DAY, EACH MONTH, ACCORDING TO EOTHAMSTED DaIEY ReCOEDS. Average composition of milk Eothamsted Dairy. pnr>Vi Tnn'n'l'.li 1RR4- (Dr Vieth— 14,235 analyses). Average Estimated quantity of constituents in milk ' per head per day each month. Speoiflo gravity. Per cent. yield of milk per head per day, 6 years. Butter- fat, Solids not fat. Total solids. Butter- fat. Solids not fat Total solids. Per Per Per cent. cent. cent. lb. lb. 11). lb.- January 1.0325 3.55 9.34 12.89 20.311 0.72 1.90 2.62 February 1.0325 3.53 9.24 12.77 22.81 0.80 2.11 2.91. Marcli . 1.0323 3.50 9.22 12.72 24.19 0.85 2.23 3.08 April . 1.0323 3.43 9.22 12.65 26.50 0.91 2.44 3.35 May 1.0324 3.34 9.30 12.64 31.31 1.05 2.91 3.96 June 1.0323 3.31 9.19 12.50 30.81 1.02 2.83 3.85 July . 1.0319 3.47 9.13 12.60 28.00 0.97 2.56 3.53 August . 1.0318 3.87 9.08 12.95 25.00 0.97 2.27 3.24 September . 1.0321 4.11 9.17 13.28 22.94 0.94 2.11 3.05 October 1.0324 4.26 9.27 13.53 21.00 0.89 1.95 2.84 November . 1.0324 4.36 9.29 13.65 19.19 0.84 1.78 2.62 ■December 1.0326 4.10 9.29 13.39 19.31 0.79 1.79 2.58 Mean . 1.0323 3.74 9.22 12.96 24.28 0.90 2.24 3.14 — '. — L_/ 1 Average over 5 years only, as the records did not commence until February 1884i FEEDING OF ANIMALS. 321 sted herd, averaging about 42 cows, almost exclusively Period of Shorthorns, in each month of the year, over six years, 1884- SI?"*^ 1889 inclusive ; and the succeeding columns show the quality of amounts of butter-fat, of solids not fat, and of total solids, in "**^*- the average yield per head per day in each month of the year, calculated, not according to direct analytical determina- tions made at Eothamsted, but according to the results of more than 14,000 analyses made under the superintendence of Dr Vieth, in the laboratory of the Aylesbury Dairy Com- pany, in 1884 ; ^ the samples analysed representing the milk from a great many different farms in each month. It should be stated that the Eothamsted cows had cake Food throughout the year ; at first 4 lb. per head per day, but after- '^^^°^^- wards graduated according to the yield of milk, on the basis of 4 lb. for a yield of 28 lb. of mUk, the result being that then the amount given averaged more per head per day during the grazing period, but less earlier and later in the year. Bran, hay, and straw-chaff, and roots (generally man- gels), were also given when the animals were not turned out to grass. The general plan was, therefore, to give cake alone in addition, when the cows were turned out to grass, but some other dry food, and roots, when entirely in the shed during the winter and early spring months. Eeferring to the column showing the average yield of milk Greater per head per day each month over the six years, it will be ^^i{ seen that during the six months — January, February, Sep- summer tember, October, November, and December — the average ^'^ '"^ yield was sometimes below 20 lb., and on the average, only about 21 lb. of milk per head per day ; whilst over the other six months it averaged 27.63 lb., and over May and June more than 31 lb., per head per day. That is to say, the quan- tity of milk yielded was considerably greater during the grazing period than when the animals had more dry food, and roots instead of grass. Next referring to the particulars of composition, according Variations to Dr Vieth's results, which may weU be considered as typical ^^'^/^jf j for the different periods of the year, it is seen that the specific at different gravity of the milk was only average, or lower than average, ««<"'»«■ during the grazing period, but rather higher in the earlier and later months of the year. The percentage of total solids was rather lower than the average at the beginning of the year, lowest during the chief grazing months, but considerably higher in the later months of the year, when the animals were kept in the shed, and received more dry food. The percent- age of butter-fat foUows very closely that of the total solids, 1 The, Analyst, April 1885, vol. x. p. 67. VOL. VII. X 322 THE KOTHAMSTED EXPEKIMENTS. Variations in quanti- ties of dif- ferent con- stituents per head per da/jj. Yidd of rthUk in sumTner greater in quantity but poorer in quality than in winter. Fwrther in- being the lowest during the best grazing months, but con- siderably higher than the average during the last four or five months of the year, when more dry food was given. The percentage of solids not fat was considerably the lowest during the later months of the grazing period, but average, or higher than average, during the earlier and later months of the year. It may be observed that, according to the average percent- ages given in the table, a gallon of milk wiU contain more of both total solids and of butter-fat in the later months of the year; that is, when there is less grass and more dry food given. Turning now to the last three columns of the table, it is seen that although, as has been shown, the percentage of the several constituents in the milk is lower during the grazing months, the actual amounts contained in the quantity of milk yielded per head, are distinctly greater during those months. Thus, the amount of butter-fat yielded per head per day is above the average of the year from April to September inclu- sive; the amounts of solids not fat are over average from April to August inclusive ; and the amounts of total solids yielded are average or over average from April to August inclusive. Prom the foregoing results, it cannot be doubted that the quantity of milk yielded per head is very much the greater during the grazing months of the year; but that the percent- age composition of the milk is lower during that period of higher yield, and considerably higher during the months of more exclusively dry-food feeding. Nevertheless, owing to the much greater quantity of milk yielded during the grazing months, the actual quantity of constituents yielded per cow is greater during those months than during the months of higher percentage composition, but lower yield of milk per head. It may be added, that a careful consideration of the number of newly calved cows brought into the herd each month shows that the results as above stated were perfectly distinct, independently of any influence of the period of lactation of the different individuals of the herd. The few results which have been brought forward in rela- tion to Milk-production are admittedly quite insufficient ade- quately to illustrate the influence of variation in the quantity and composition of the food, on the quantity and composition of the milk yielded. Indeed, owing to the intrinsic difficulties of experimenting on such a subject, involving, as has been pointed out, so many elements of variation beside those which it is sought to investigate, any results obtained have to be interpreted with much care and reservation. Nevertheless,. FEEDING OF ANIMALS. 323 exercising such care and reservation in regard to the num- erous results of ourselves and others which are at command, it may be taken as clearly indicated that, within certain limits, high feeding, and especially high nitrogenous feeding, High feed- does increase both the yield and the richness of the milk. !^j|^ But it is evident that, when high feeding is pushed beyond a mak. comparatively limited range, the tendency is to increase the weight of the animal — that is, to favour the development of the individual, rather than to enhance the activity of the func- tions connected with the reproductive system. This is, of course, a disadvantage when the object is to maintain the milk-yielding condition of the animal ; but when a cow is to be fattened off it will be otherwise. It has been stated that, early in the period of six years in Food cO- which the Eothamsted results that have been quoted were ^^^^^^ obtained, the amount of oil-cake given was graduated accord- according ing to the yield of milk of each individual cow ; as it seemed ^^^ °^ unreasonable that an animal yielding, say, only 4 quarts per head per day, should receive, beside the home foods, as much cake as one yielding several times as much. The obvious supposition is, that any excess of food beyond that required for sustenance and milk-production would tend to increase the weight of the animal, which, according to the circum- stances, may or may not be desirable. But there remains the important question — Whether the period of lactation is lengthened, or the yield of the higher yielding cows is main- tained the longer, by an increased amount of food ; or whether, on the other hand, the period of lactation, or the yield of milk, is reduced by the limitation of the supply of food? The point is, at any rate, deserving of careful experiment and observation. It may be observed that direct experiments at Eothamsted confirm the view, arrived at by common experience, that roots, and especially mangels, have a favourable effect on the flow of milk . Further, the Eothamsted experiments have shown influence that a higher percentage of butter-fat, of other solids, and of "IJ^^^ total solids, was obtained with mangels than with silage as yidd of the succulent food. The yield of milk was, however, in a '^^• much greater degree increased by grazing than by any other change in the food ; and with us, at any rate, the influence of roots comes next in order to that of grass, though far behind it, in this respect. But, with grazing, as has been shown, the percentage composition of the milk is considerably reduced; though, owing to the greatly increased quantity yielded, the amount of constituents removed in the milk whilst grazing may, nevertheless, be greater per head per day than under any other conditions. 324 THE EOTHAMSTED EXPERIMENTS. Lastly, it has been clearly illustrated how very much greater is the demand upon the food, especially for nitrogen- ous and for mineral constituents, in the production of milk than in that of fattening increase. Gonstitvr ents of crops re- tamed, on fa/nns. a source of manv/re. Food and Manuee. At the commencement of this Section on the Feeding of Animals, it was shown, by reference to a special example, how large was the proportion of the constituents of the crops grown in a rotation which was retained on the farm for further use — in fact, for consumption by animals, or for litter. It was shown that, in the case selected for illustration, there would be so retained on the farm for such further use, more than two-thirds of the total vegetable substance grown, more than half of the nitrogen of the crops, and about six-sevenths of the total mineral matter ; whilst, of the individual mineral constituents of the crops, less than half of the phosphoric acid, but about four-fifths of the potash, would be retained. Of course, in the very varied practice of Agriculture at the present day, there will sometimes be larger, and sometimes smaller, proportions of the various constituents of the crops at once sold off, or retained on the farm ; but the example given may be taken as essentially typical, and as so far con- veying a very useful impression on the subject. But, besides the constituents of the home-grown rotation crops retained upon the farm for food and litter, there will be more or less produce from grass land, whilst modern practices frequently involve the purchase of a considerable quantity of imported food-stuffs. Eesults relating to the feeding of animals for the produc- tion of meat, and of milk, have been considered ; and we have now to discuss the subject of feeding as a source of manure. Numerous Eothamsted experiments have shown how small is the proportion of the various constituents con- sumed in food by fattening, or even by growing animals, which is stored up in their increase, and which will therefore be lost to the manure. In the production of milk, however, it has been seen that the loss to the manure is very much greater. Of the mineral matters of the food, we know that there need be no loss to the manure beyond that carried off in the animal increase or in milk. Of the non-nitrogenous organic substance of the food, a very large proportion is lost by the respiration of the animals, and a not inconsiderable quantity contributes to the animal increase or milk ; and what remains for manure is of no material .value as a direct supply of con- FEEDING OF ANIMALS. 325 stituents, and of comparatively little by the action of its pro- ducts of decomposition within the soil. Indeed, the most whatpw- important point to consider is — what proportion of the nitro- ^^^Z^'^ gen of the food remains for manure ? As has been shown, mfood n- and as will be further illustrated presently, only a compara- ^aMM/o/ tively small proportion is carried off in animal increase ; but a much larger amount is lost to the manure in the production of milk. But the further questions arise — Is there any, so to speak, vital exhalation of nitrogen, or of any compounds of it, by the animal ? Or, may we estimate that the whole of that consumed which is not carried off in the animal in- crease, or in milk, will be found in the solid and liquid de- jections, and so remain for manure ? Or, on the other hand, is there any assimilation by the animal, of the free nitrogen of the atmosphere? The further practical question still remains — Is there any material loss of nitrogen after the solid and liquid excretions leave the body, and before their utilisation within the soil for the production of future crops ? First, then, is there any vital exhalation by the animal of Exlmla- nitrogen or of any of its compounds? _ ^^o^ion Obviously, this is a question which could not be experi- of nitrogen mentally investigated before definite knowledge was attained ^z^o"*™"^- in regard to the composition of the atmosphere. But after such knowledge had been acquired, rather more than a century ago, the subject of the mutual relations of the at- mosphere, and of vegetable and animal growth, came to be studied ; and, among other points, it was sought to determine whether, on the one hand, the free nitrogen was assimilated by animals ? or, on the other, whether it was exhaled, at the expense of the nitrogenous substance of the food, of the blood, or of the more fixed substance of the body ? Commencing towards the end of the last century, numer- Various ous investigations have been undertaken from various points J^''*^"' of view bearing upon the subject; and among the investi- gators or writers may be named — ^Lavoisier, Laplace, Sdguin, Dalton, Sir H. Davy, Pfaff, Provengal and Humboldt, Allen and Pepys, Despretz and Dulong, Brunner and Valentin, Marchand, von Erlach, Baumert, Regnault and Reiset, Ber- thoUet, Milne-Edwards, and C. G-. Lehmann ; besides others more recently. It is impossible shortly, and at the same time adequately, either to describe or to criticise the numerous and, upon the whole, discordant results, that have been obtained in regard to the question of the assimilation or exhalation of free nitro- gen by animals. It is noticeable that the earlier investi- gators, Lavoisier, Laplace, and S^guin, concluded that the amount of nitrogen expired was neither more nor less than 326 THE EOTHAMSTED EXPEEIMENTS. tionsnot condimve. Loss of I ireath- ing cmd that inspired ; and in this view they are in the main sup- ported by the conclusions, though not entirely by the results, of Allen and Pepys, of Brunner and Valentin, and von Erlach. In favour of the view that free nitrogen is ab- sorbed and assimilated, may be cited the opinions of Sir Humphrey Davy and of Pfaff, so far as certain warm-blooded animals are concerned ; and of Provenqal and Humboldt, and of Baumert, in regard to fish. On the other hand, that there is evolution of free nitrogen has been concluded, by Sir H. Davy, BerthoUet, Dulong and Despretz, Magnus, Marchand, Grassi, Eegnault and Eeiset, and C. G. Lehmann. In regard to evolution, the most extensive and elaborate experiments are those of Eegnault and Eeiset. But the amounts which their results indicated would imply the loss, in that way, of an incredibly large proportion of the total nitrogen consumed in the food ; whilst Liebig estimated that the evolution which Dulong assumed was so great that, in the case of one of the experimental animals, the whole of the nitrogen of the body would be lost in seven days ; and that, at the rate assumed by Despretz, the nitrogen of one pound of flesh would go off in thirty-one hours. Then, the results indicating absorption are the most pro- nounced in the experiments with fish. The question arises, therefore, whether in their case the result may not be ex- plained i)y supposing that oxygen has been absorbed from the air within the body, especially in the swimming bladder, and nitrogen stored up in its place, under the conditions of limited supply of oxygen from external sources to which the animals have generally been subjected during experiment. Upon the whole it must be concluded that, from a variety of causes, connected sometimes with the conditions under which the animals were placed under experiment, sometimes with the circumstances under which the samples assumed to represent the inspired and expired air, respectively, were taken for analysis, and sometimes with the methods of analysis themselves, the results of the experiments on respiration which have been referred to, have not been sufficiently free from doubt to be accepted as establishing so important a con- clusion as either the assimilation of free nitrogen by animals, or the evolution of it from its compounds within the body. The next point to consider is — whether there is any loss of ammonia, or of other compounds of nitrogen, in the breath, or by the skin. Louis Thompson, Thiry, Grouven, and others, have found some emanation of ammonia ; but Lessen, and others, consider it doubtful whether the ammonia in the air itself might not account for the results. FEEDING OF AiOMALS. 327 Various experiments have been made to determine the loss of nitrogen in sweat. In the sweat of man ammonia and urea have been found. In the sweat of a horse Grandeau and Leclerc ^ found ammonia, urea, and albumin. Professor r. Smith, of Aldershot,^ has also examined the sweat of horses. Besides various inorganic salts, he found ammonia, and 3.381 per cent of albumin. He reckons that a pint of sweat will thus contain 0.676 ounce of albumin, and that this amount would be equivalent to the nitrogen in 5f ounces of oats. He further thinks it probable that the reduction of sweating by clipping would, with hard work, be equivalent to 1 lb. of corn per day. It seems safe to conclude that the loss of combined nitrogen Loss im- by gaseous emanations from the lungs and skin is, for all ''^«™*^- practical purposes, quantitatively immaterial The sweat would seem to be a more important source of loss in animals submitted to much muscular exercise. But, even in their case, it does not seem to be large ; whilst in that of the ani- mals of the farm fed for the production of meat or milk, it would presumably be much less material. We now come to the consideration of evidence of quite Anumnts another kind as to the loss to the manure of the nitrogen of i^/oo™S the food, beyond the amount stored up in increase, or removed manwre. in milk : namely, that afforded by the results of experiments made to determine the relation of the amount of nitrogen voided in the solid and liquid excretions, to that consumed in the food. Most of these have been made with the animals of the farm ; indeed, most of them have had for their object the direct determination of the amount of the nitrogen of the food consumed which is recovered in the manure in practical feeding. The chief results may be very briefly summarised as follows : — Boussingault made experiments^ with a cow, with a Bomsin- horse, and with turtle-doves (probably between 1830 and ^^^^^ 1840). In the experiment with a cow, the animal was fed on the same food for about a month, and the results relate to the three concluding days of that period. Boussingault observes that the animal did not suffer any material change in weight. Besides the nitrogen removed in the milk, there was an amount not recovered in the excrements which represented a loss of 13.4 per cent of the total nitrogen of the food. In the experiment with a horse, the animal had received 1 Armalesde la Science agrononUque, 5™ annie, 1888, tomeii. pp. 311-314. 2 Jmmfial ^Physiology, 1890, vol. xi. p. 497. 3 Agronomic, Chimie agricole et Fhysiologie, 2"'* ed., 1874, vol. v. p. 144. 328 THE KOTHAMSTBD EXPERIMENTS. Nitrogen not ac- counted for. ments at Rothaim- sted; how conducted. Food used. consumed and voided. Nitrogen not ac- covmted for. the same ration for three months, and did not either gain or lose in weight appreciably. There was here again an amount unaccounted for, representing a loss of 17.2 per cent of the nitrogen of the food. In the two experiments with turtle-doves, one over five and the other over seven days, each of the birds rather lost weight. Their food was millet ; and in the one case there was a loss of 35.9, and in the other of 34.1, per cent of the nitrogen in the food. Boussingault thought that there was undoubtedly a loss of nitrogen, as the amouijt unre- covered was far too great to be accounted for by errors of analysis. Experiments were made on the subject at Eothamsted in 1854 with pigs. Individual male animals were experimented upon, for periods of three and of ten days. Each animal was kept in a frame, preventing it from turning round, and having a zinc bottom sloping slightly from each side towards the centre, where there was an outlet for the urine to run into a bottle beneath. They were watched night and day, and the voidings carefully collected as soon as passed, which could easily be done, as the animals never passed either faeces or urine without getting up, and in so doing rang a bell, and thus attracted the notice of the attendant. The constituents determined were — in the food and faeces, dry matter, ash, and nitrogen; and in the urine, dry matter, ash, nitrogen, and urea. In preparing samples of faeces or of urine for nitrogen determinations, a mixture was made of a proportional part of the voiding of each twenty-four hours, and oxalic acid added. In the case of the faeces, portions of the acid mixture were taken for the determination of dry matter; and nitrogen determinations were made in the partially dried substance, and calculated up to the fully dried condition. In the case of the urine, portions of the acid mixture were fully dried, and other portions partially dried, and then mixed with about half the weight of fully dried oak-dust, in which the nitrogen was determined. Over a preliminary period, and also over each period of exact experiment, one animal received the highly nitrogen- ous lentil-meal, and the other the low - in - nitrogen barley- meal. In each case, the one receiving lentil-meal consumed more than twice as much nitrogen in food, and voided more than twice as much in the solid and liquid excrements. Notwithstanding the great attention paid to the collection, the sampling, and the preparation of the samples of the excre- ments for nitrogen determinations, as above described, there was, in each case, a considerable amount of the nitrogen of the food unaccounted for in that estimated in the increase FEEDING OF ANIMALS. 329 and in that found in the excrements. There was, too, a much greater loss indicated by the results of the direct nitrogen determinations in the urine dried with an excess of oxalic acid, than when the nitrogen was calculated from the amount of urea found daily in the fresh urine. As, however, nitrogen determinations (by soda-lime and platinum salt) were made by two analysts, whose results agreed very fairly, it may be concluded that the loss was connected with the methods of collection, sampling, and preparation for analysis, rather than with those of the analysis ; and it is probable that the same remark applies to the results obtained with the fteces. In illustration of the range of loss of nitrogen indicated, it may be stated that when the nitrogen in the urine was reckoned from the amount of urea, the loss ranged in the four experi- ments between 20 and 30 per cent of that in the food, and when by direct nitrogen determinations in urine as well as in faeces, from under to over 40 per cent. However, in the case of each food, whether the nitrogen in the urine was deter- mined, or calculated from the urea, there was considerably less loss indicated over the ten-day than over the shorter three- day period; again connecting the error with the collection, sampling, and preparation, rather than with the analysis. In view of these unsatisfactory results, and of the evidence Further ex- that much at any rate of the loss was probably due to experi- ^'^"^^ mental difficulties and errors, the subject was taken up again in 1862. The pigs were kept in frames as before, and the voidings were collected in the same way; but they were sampled morning and evening, instead of only once in the twenty-four hours, as in 1854. Advantage was also taken of the previous experience in regard to various other points of manipulation. Lastly, the direct nitrogen determinations were made by soda-lime as before, but with titration instead of platinum salt. Two animals were experimented upon,' each for a period Food used, of ten days, and after an interval of a few weeks for five days more. The food of one consisted of three parts bean-meal and one part bran, and of the other of three parts barley-meal and one part bran. In the case of the pig having the highly nitrogenous bean- mtrogen meal and bran, the nitrogen balance for the ten days showed f^^^t a gain of 4.04 per cent when direct nitrogen determinations accmmud were made in the urine, and of only 2.32 per cent when the f"^- nitrogen in the urine was calculated from the amount of urea. On the other hand, over the five-day period there was a loss indicated of 3.35 per cent with the direct nitrogen determina- tions in the urine, and of only 1.61 per cent when the nitrogen was calculated from urea. In the latter case, therefore, the 330 THE ROTHAMSTED EXPERIMENTS. amount of nitrogen accounted for was again less with direct determination than by calculation from urea. In the case of the pig having the low-in-nitrogen barley- meal and bran, there was, over the ten-day period, a loss indi- cated of 7.16 per cent of nitrogen with direct determination, and of only 4.90 per cent when the nitrogen was calculated from the urea. In this case, therefore, there was again less loss of nitrogen by calculation from urea than by direct determination. Lastly, over the five-day period there was, with the barley-meal and bran, a gain of nitrogen indicated of 7.76 per cent with direct determination of nitrogen in the urine, and of 11.02 per cent when calculated from the urea. In both cases, therefore, there was more nitrogen accounted for by calculation from urea than by direct determination. These results obtained in 1862 show, therefore, with the beans and bran, a slight gain over the ten days, and a slight loss over the five days. On the other hand, with the barley and bran there was a comparatively small loss over the ten days, and a somewhat greater gain over the five days. When the fact that there was a much greater variation in the amounts of the daily voidings than in those of the food daily consumed, and also the uncertainty in the estimation of the proper increase of the animals over short periods and of the nitrogen in it, are taken into account, these results must No reed be admitted to afford no evidence of any real loss to the ^itrfqen ^^^^^^'s of *^6 nitrogen of the food beyond that in the increase and in the excrements. The next results to consider were obtained at Eothamsted in 1861 with sheep. There were four pens with five sheep in each. Besides the determination of the total dry matter, ash, and nitrogen, in the food and in the excrements, one special object was to determine what proportion of the cellulose of the food was digested, and whether more or less was so utilised according *to the character of the foods givea with it. Accordingly, foods containing a comparatively large amount of cellulose were selected, as under : — Food used. Pen 1. Meadow hay-chaff alone ad libitum. II 2. 1 lb. of ground beans per head per day, and meadow hay-chaff ad libitum. M 3. 1 lb. of ground barley per head per day, and meadow hay-chaff ad libitum. II 4. About 6^ oz. of ground beans, and about 3J oz. of linseed-oil, per head per day, and meadow hay-chaff ad libitum. In Pen 4 the object was to give an amount of beans con- taining the same quantity of nitrogen as the barley of Pen 3, and then to make up the deficiency of starch in the smaller Tnents with FEEDIKG OF ANIMALS. 331 quantity of beans compared with that in the barley by oil, in the proportion of 1 part of oil for 2J parts of starch. With a view to the careful collection, sampling, and analysis, of the excrements, the sheep were kept under cover, on rafters, through which (but with some loss) the solid and liquid ex- creta passed on to a sheet-zinc flooring, at such an incline that the liquid drained off at once into carboys containing oxalic acid ; and the solid matter was removed two or three times daily, and also mixed with oxalic acid. After a preliminary period of eight weeks the exact feeding experiment was continued for thirty-two weeks more — from Janiiary 25 to September 6. Commencing on March 26, and ending on August 9, samples of the excrements were taken at intervals, in each case for several consecutive days — namely, 4, 5, 5, 7, 7, 7, 7, 7, 7, and 7 days ; and the results here given are the means of the. seven 7-day periods. The amounts oi Nitrogen nitrogen so indicated to be not recovered in either the increase "^,^ or in the excreted matters were, in the four pens, respectively for. 12.5, 25.4, 15.2, and 17.7 per cent of the nitrogen supplied in the food. It is to be observed that the estimated loss is the greatest with the most, and the least with the least, nitrogen in the food. The question arises — Whether the greater esti- mated loss is connected with an under-estimate of the nitro- gen in the increase of the animals feeding on the more highly nitrogenous food, or with an actually greater loss from de- composition in the case of the more highly nitrogenous excrements. In 1858, Henneberg^ made experiments with two oxen, Henne- each separately. The animals were kept on sustenance food ^'^^^^ only. After a preliminary period of several weeks, there were three periods of more exact experiment — the first from February 27 to March 27, the second from March 28 to May 21, and the third from May 22 to July 15 ; and during three days towards the end of each of these periods the excrements were collected and analysed. Ox No. 1 gained 6 lb. during the three days of the first period, 1 lb. during those of the second, and 11 lb. during those of the third. The percentage NUrogm of the nitrogen of the food which was not recovered in the ^^"^^ excrements was, for the respective three-day periods, 5.7, 28.8, for. and 15.1, or an average of 16.5. Ox No. 2 neither gained nor lost during the first three-day period, lost 3 lb. during the second, and 8 lb. during the third ; and the analyses of the excrements showed a gain of nitrogen compared with that in the food of 9.6 per cent over the first three days, a loss of 24.7 per cent over the second three, and again of 6.3 per cent 1 BHtrdge zv/r Begrwndung einer rationellen Futtenmg der Wiederkduer, Heft 1, 1860. 332 THE EOTHAMSTED EXPERIMENTS. No litter ments at Wobwn, Park. Nitrogen not ac- cotmted for. over the third. That is to say, Ox No. 1, with more or less gain over each of the three-day periods — which raay perhaps he interpreted as retention in the alimentary canal or bladder rather than increase in the substance of the body — showed a considerable deficit of nitrogen in the excrements compared with that in the food. Ox No. 2, on the other hand, with loss of weight — which probably only represented more com- plete evacuation in relation to the food consumed — indicated more of tendency to excess of nitrogen in the excrements compared with that in the food. In experiments in 1860-61, also with two bullocks, Henneberg found — this time over six- day instead of three-day periods — deficits of nitrogen in the excrements corresponding to the following percentages of the amounts supplied in the food — 35, 37, 21, 12, 10, 0. It may be observed that the percentage of loss was, upon the whole, the greater with the larger amounts of nitrogen in the food. Later results of Henneberg will be referred to further on. In none of the foregoing experiments, either by Bous- singault, at Eothamsted, or by Henneberg, was any litter used, the excrements beiug collected and analysed by them- selves. In 1851, we made experiments with oxen, at Woburn Park Farm, by the permission of the Duke of Bedford. In the experiment, the results of which are given below, there were five Herefords, each in a separate box, and the experimental period extended over thirty-five days. Liberal fattening food was given, consisting of a cooked mixture of equal parts of ground oil-cake, barley, and beans, besides clover-hay-chaff, and swedes. The litter consisted of wheat-straw; and an absorbent, composed of 2 parts sawdust and 1 part sulphuric acid, was used ; a small quantity being daily sprinkled over the manure in the boxes just before spreading the fresh litter. At the end of the experiment the whole of the dung was got out, put into a large shed, turned over by men, pulled to pieces by boys, and thoroughly mixed; and in that state it was weighed, and several separate 100 lb. samples were taken, each being put into a clean cask, in which state the samples were sent to Eothamsted for analysis. In the preparation for analysis, the whole of the 100 lb. sample was coarsely ground, then divided into por- tions, one or more of which was finely ground for analysis, and in the sample so prepared the nitrogen was determined by the soda-lime method. It was so determined separately in samples from two of the 100 lb. casks. Deducting the amount of nitrogen in the increase (reckoning it to contain 1.27 per cent), there was a deficiency of nitrogen in the dung, compared with that in the food and litter — according to one FEEDING OF ANIMALS. 333 100-lb. sample, of 8.03, and to the other or duplicate one, of 10.55 per cent. Such, then, were the results of the earlier experiments Rmewof made by various investigators, to determine whether or not '^J^V^^ <^ there was any loss of nitrogen in the feeding of animals nitrogen. beyond that stored up in their increase. It will be observed that, with the exception of the turtle-doves experimented upon by Boussingault, all the other results were obtained with the animals of the farm ; and in all cases, excepting those of the experiments at Eothamsted with pigs and with sheep, and at Woburn with oxen, the animals were assumed to be fed on only sustenance rations, and no allowance was made in the calculations for any increase or loss in their weight. It has been seen that in every case, excepting in the experiment with Henneberg's Ox No. 2, and in the experi- ments at Eothamsted with pigs in 1862, the figures indicate a notable, and in some a very considerable, loss of nitrogen ; which, assuming it to be not explained by storing up of nitrogen in the animal, or deficient evacuation, might be supposed to point to a probable loss by respiration, or perspira- tion, or both. From a study in much detail of the direct experiments on respiration and perspiration which have been referred to, we ourselves have been disposed to conclude that there was no material exhalation of either free nitrogen or of its com- pounds. Further, notwithstanding our own early results with pigs, those with sheep, and those at Woburn with oxen, all indicated more or less, and sometimes a considerable loss, the observations made during the conduct of the investi- Loss of gations so fully impressed us with the liability to error, ^*j&^ especially on the side of loss, that we have always considered it doubtful whether there was in reality any material loss at all. In the first place, there is the uncertainty in the estima- tion of the changes in the weight of the body — whether to attribute them to increase or loss of its fixed substance, or to excess or deficiency in the evacuations in relation to the food consumed within the period of experiment ; and there are, besides, great difiiculties to be overcome, both in the com- plete collection, the proper sampling, and the preparation, without change, of the excreted matters ; and there are also special difficulties in the adaptation of analytical methods to secure exact representative results. Indeed, most of the results so far quoted, whether of ourselves or others, must be looked upon as little more than pioneer; though, taken as such, the experience gained has proved to be of essential value in directing attention to the difficulties and sources of error incident to such work, and to the improve- 334 THE EOTHAMSTED EXPEEIMENTS. Further merits in Germany. Nitrogen of the food, entirely re- appewring in excre- ments. ment in methods of collection, sampling, preparation, and analysis. Por ourselves, being satisfied that much if not the whole of the losses that had been indicated was to be explained by the methods of experimenting, and being very fully occupied with other subjects, we decided, after our experiments with pigs in 1862, not to devote the very great amount of time and labour that would be involved in the repetition of the investigation with still further precautions. In Germany, however, Henneberg and his colleagues (G. Kiihn, H. Schultze, and B. Schultz), at Weende, as well as others, continued to work on the subject with the animals of the farm. Henneberg ^ pointed out that the experiments of Bischoff and Voit with dogs in 1859,^ of Eanke with man in 1860-61,3 of Voit with pigeons in 1860-62,* and of Petten- kofer and Voit with man,® showed almost complete re-appear- ance of the nitrogen of the food in the solid and liquid excre- tions ; and, if this were the case with carnivora and omnivora, there seemed no reason why it should not be so with herbi- vora. He further pointed out how small an actual loss or gain in the determined amount of nitrogen in the faeces or the urine might make a great difference in the balance ; and he admitted that more attention than had hitherto been given to certain points must in future be devoted — as, for instance, to the rinsing and washing of the stalls, and to the determin- ation of the dry matter in the food, faeces, and urine, more frequently and uniformly throughout the experimental period. In the Weende experiments of 1865, and subsequently, more attention was paid to such points, and the periods of exact experiment were longer. There was, accordingly, great improvement in the results. Thus, in a series of eight experi- ments with oxen, in five with only sustenance or maintenance rations, the result was that, in three of them the percentage deficit of nitrogen in the excrements compared with that in the food was 0.4, 2.7, and 2.2. ; whilst in the other two there was a gain representing 0.8 and 3.7 per cent. In the three other experiments, fattening food containing about twice as much nitrogen was given, and in these the deficits in the excrements were 12.1, 12.0, and 17.7 per cent of the nitrogen in the food. Henneberg concluded that, with only susten- ance rations, the whole of the nitrogen of the food of oxen reappeared in the excrements, and that it was no longer > Neue Beitrage, Gottingen, i. 373-375, 1872. ^ Die Gesetze der Emahrung des Fleischfressers, Leipzig, 1860. ' Archw fii/r (mat., phys. umd wissensehaftUche Mediein, Leipzig, 1862, p. 311. * Annalen, II. Suppl. p. 238, 1862. 5 Zeits.f. Biol., II. p. 459. FEEDING OF ANIMALS. 335 necessary to infer from the results obtained with other animals what would take place with ruminants. Henneherg also quotes results ^ obtained with cows by Expert- Voit at Munich, by G. Kiihn and Fleischer at Mockern, and ^^ ^''* by Fleischer at Hohenheim. Voit's results, obtained in 1867, showed a deficit of nitrogen in the milk, fseces, and urine, representing 1.2 per cent of that in the food. In eight ex- periments made at Mockern in 1867-68 with cows, six showed respectively losses corresponding to 2.9, 11.1, 3.8, 5.6, Losses and 16.4, and 7.0 per cent of the nitrogen in the food; and the ^^^f£_ other two showed gains corresponding to 1.2 and 4.8 per cent. In the case of the larger losses more nitrogen was consumed in the food, and the animals gained in weight, and presumably stored-up nitrogen. At Hohenheim, in 1870, experiments were made by Fleischer with two cows, one of which showed a loss of 0.3, and the other a gain of 0.6 per cent of nitrogen compared with that in the food. Experiments were also made with sheep by Maercker and E. Schulze, at Weende,^ which confirmed the conclusions drawn from those with oxen and cows as above, as also did others made by Stohmann with goats '^ at the Halle experi- mental station. We will conclude the citation of experimental evidence on Trials the point, by reference to some of the results obtained by '^^'^'^°ff^- Voit from 1859 to 1863 with dogs.* In none of these cases was the period of exact experiment less than 6 days, whilst in some it was 12, 14, 20, 23, 49, and even 58 days. In eight out of the eleven cases there was an excess of nitrogen in the excrements compared with that in the food, representing the following percentages of gain on that in the food, 1.0, 0.7, 0.4, Gains and 0.4, 0.6, 0.3, 0.1, and 0.1 ; whUst the deficits represented 1.4 ^/**»^^{ and 0.3 per cent, and one experiment showed neither gain nor loss. Since the publication of the various results above quoted. Practically there has been little doubt entertained that, not only in the '^^i}°^g°£ case of carnivora and omnivora, but also in that of herbivora, and even of ruminants, practically the whole of the nitrogen of the food which does not contribute to animal increase or to milk, reappears in the excrements. In our estimates of the value of the manure from the con- Mawwndi sumption of different foods by animals on the farm, so far as J^^^ "■^ the nitrogen was concerned, we many years ago deducted 1 mue Beitrage, Heft I. p. 383, 1872. ^ Jcm/rn. f. Landw., 1870 and 1871; Armsby, Manual of Oattle-f ceding, 3rd ed„ ,1877, pp. 99, 100. 3 Zeits. f. Biol., 1870, p. 204 ; Armsby, loc. cit., pp. 100, 101. •' Bisehoff and Voit, Die Oesetze der Emahrung des Fleischfresaers, 1860 ; and Wolff's Die Emahrung d. landw. Nutzihiere, 1876. 336 THE EOTHAMSTED EXPERIMENTS. Valucdion ofunex- 10 per cent from the amounts consumed in oilcakes and leguminous seeds, which contain high percentages of nitro- gen, and 15 per cent from the amounts in the foods which contain lower percentages. These deductions were reckoned to include the amounts of nitrogen actually stored up in the increase of live-weight, and also some little loss if any, but not to cover the larger losses that may take place in the manure after it is voided by the animals. More recently, however, we have estimated the amount actually stored up in the animal, and have assumed the whole of the remainder to be voided in the solid and liquid excretions. For details on the point, we must refer to our most recent paper bearing upon the subject, entitled On the Valuation of Unexhausted Manures} The calculations relate to the use of food for the production of fattening increase. It is assumed that, on the average, such increase will contain 8 per cent of nitrogenous substance, corresponding to 1.27 per cent of nitrogen in the increase. According to the calculations it Percentages of nitrogen assimilated and voided ty aniinals. Fattening animals. GrovAng animals. Cows. results that, of the total nitrogen consumed in foods rich in that substance, such as oilcakes and leguminous seeds, there will generally be less than 5 per cent retained in the fatten- ing increase in live-weight. In the case of the cereal grains, on the other hand, which are much less rich in nitrogen, a much larger proportion of the total amount consumed will be retained in the increase — generally, perhaps, about 10 per cent of it. Of the nitrogen in gramineous straws a still higher proportion will probably be devoted to increase; whilst roots will, on the average, lose by feeding, perhaps, only about 5 or 6 per cent of their nitrogen. Thus, when fattening increase only is produced, the pro- portion of the nitrogen of the food which will be retained by the animal, and so lost to the manure, is very small in the case of the richer foods, but more in that of the poorer ones ; though, even with them, it will seldom exceed 10 per cent, except possibly in the case of straws. It may be assumed, however, that when foods are consumed by store animals, whose increase is largely growth, about twice as much of the nitrogen of the food is retained, and so lost to the manure. And when, as is more and more the case with early maturity, the increase comprises a larger proportion of growth than in mere fattening, the amount of the nitrogen of the food which will be lost to the manure will be more than with fattening only, but less than with merely store animals. When, how- ever, food is consumed for the production of milk, a very much greater proportion of its nitrogen wiU be lost to the manure. ^ Joum. Roy. Ag. Soc. Eng., vol. xxi., SS., Part II., 1885. feeding of animals. 337 Food and the Exekcisb of Foece. We now come to the last branch of our subject — namely, The Feeding of Animals for the Hxercise of Force. With the very limited space still left at our disposal, we will commence our historical sketch with a brief account of the views of Liebig as first put forward in 1842 in his work On Organic LieUg's Chemistry in its applications to Physiology and Pathology. There is, indeed, a special appropriateness in so doing, since there can be no doubt that the course of subsequent inquiry and discussion has been materially influenced by the opinions he then enunciated. The following quotations from the above-mentioned work will suf&ce to indicate his more specific views in regard to the connection between food requirements and the exercise of force : — As an immediate effect of the manifestation of mechanical force, we see that a part of the muscular substance loses its vital properties, its character of life ; that this portion separates from the living part, and loses its capacity of growth and its power of resistance. We find that this change of properties is accompanied by the entrance of a foreign body (oxygen) into the composition of the muscular fibre (just as the acid lost its chemical character by combining with zinc) ; and all ex- perience proves, that this conversion of living muscular fibre into com- pounds destitute of vitality is accelerated or retarded according to the amount of force employed to produce motion. Nay, it may safely be affirmed that they are mutually proportional ; that a rapid transfor- mation of muscular fibre, or, as it may be called, a rapid change of matter, determines a greater amount of mechanical force ; and con- versely, that a greater amount of mechanical motion (of mechanical force expended in motion) determines a more rapid change of matter. —Pp. 220, 221. And again : — The amount of azotised food necessary to restore the equilibrium between waste and supply is directly proportiona,l to the amount of tissues metamorphosed. The amount of living matter, which in the body loses the condition of life, is, in equal temperatures, directly proportional to the mechanical effects produced in a given time. The amount of tissue metamorphosed in a given time may be measured by the quantity of nitrogen in the urine. The sum of the mechanical effects produced in two individuals, in the same temperature, is proportional to the amount of nitrogen in their urine ; whether the mechanical force has been employed in voluntary or involuntary motions, whether it has been consumed by the limbs or by the heart and other viscera. — Ibid., p. 245. Such, in fact, were the views in regard to the special exigencies of the system in the exercise of force, which be- came at once identified with Liebig's name, and continued to VOL. VIL Y 338 THE EOTHAMSTED EXPEKIMENTS. Rofham- sted re- he SO identified for many years. Thus, Professor Frankland, in Lis lecture at the Eoyal Institution in 1866 ^ on the ex- periments of Fick and Wislicenus,^ refers to these views of Liebig as having, up to that time, been pretty generally adopted by text-book writers. The results of our own feeding experiments, which were commenced some years after the appearance of Liebig's work, being apparently inconsistent with the then current views on some important points, we were led at once to tiirn attention to the subject of human dietaries ; and also to a consideration of the management of the animal body undergoing somewhat excessive labour, as for instance, the hunting-horse, the racer, the cab-horse, the fox-hound, and also pugilists and runners. The conclusions to which we were led by this study were briefly summarised in a paper published in the Beport of the British Association for the Advancement of Science, for 1852, as follows : — Conclu- sions of 1852. Hespir- atory Tna- terictl and muscula/r force. . . . that in the cases, at least of ordinary exercise of force, the exi- gencies of the respiratory system keep pace more nearly with the de- mand for nitrogenous constituents of food than is usually supposed ; and further :- — A somewhat concentrated supply of nitrogen does, however, in some cases, seem to be required when the system is overtaxed ; as for in- stance, when day by day more labour is demanded of the animal body than it is competent without deterioration to keep up ; and perhaps ako, in the human body, when under excitement or excessive mental exercise. It must be remembered, however, that it is in butcher's meat, to which is attributed such high flesh-forming capacity, that we have also, in the fat which it contains, a large proportion of respiratory material of the most concentrated kind. It is found, too, that of the dry substance of the egg, 40 per cent is pure fat. A consideration of the habits of those of the labouring classes who are under- rather than over-fed, will show that they first have recourse to fat meat, such as pork, rather than to those which are leaner and more nitrogenous ; thus perhaps indicating, that the first instinctive call is for an increase of the respiratory constituents of food. It cannot be doubted, however, that the higher classes do consume a larger pro- portion of the leaner meats ; though it is probable, as we have said, that even with these as well as pork, more fat, possessing a higher re- spiratory capacity than any other constituent of food, is taken into the system than is generally imagined. Fat and butter, indeed, may be said to have about twice and a half the respiratory capacity of starch, sugar, &c. It should be remembered, too, that the classes which con- sume most of the leaner meats, are also those which consume the most butter, sugar, and in many cases, alcoholic drinks also. It is further worthy of remark, that wherever labour is expended in the manufacture of staple articles of food, it has generally for its object the concentration of the mom-nitrogenous, or more peculiarly respiratory constituents. Sugar, butter, and alcoholic drinks are notable instances. ' Journ. B. Inst., 1866. 2 Phil. Mag., 1866, 4th series, vol. 31, pp. 485-503. FEEDING OF ANIMALS. 339 of this. Cheese, which at first sight might appear an exception, is in reality not so ; for those cheeses which bring the highest price are always those which contain the most butter ; whilst butter itseK is always dearer than cheese. In conclusion, it must by no means be understood that we would in any way depreciate the value of even a somewhat liberal amount of nitrogen in food. We believe, however, that on the current views too high a relative importance is attached to it ; and that it would conduce to further progress in this most important field of inquiry if the pre- vailing opinions on the subject were somewhat modified. It is to be borne in mind, that at the time these opinions were put forward, now more than forty years ago, the views expressed were directly contrary to all recognised authority on the subject ; and that it is since that date that so much evidence has been accumulated, as to the amounts of urea, and of carbonic acid, given off under varied conditions as to food and exercise. Still, from the facts already at command. Food con- it was concluded that the increased demand for food resulting ^^^J^^^ from the exercise of muscular power was specially character- Tyy uhawr. ised by the requirement for an enhanced amount of the non- nitrogenous constituents. Confirmatory evidence was, however, not long wanting. Further Thus, in 1854, we selected two pigfe as nearly as possible of *"■'''*• equal weight and character; to one was given, ad libitum, lentil-meal (containing about 4 per cent of nitrogen), and to the other, also ad libitum, barley-meal (containing less than 2 per cent). Each animal was kept in a frame, with arrange- ments for collecting the faeces and urine separately, as already described. After they had been kept for a certain time on their respective foods, one comparative experiment was con- ducted for three days, and later on another for ten days. The weights of the animals were taken at the beginning and at the end of each experiment ; and, besides other particulars, the amounts of nitrogen consumed in the food, and of urea voided, were determined. The results are summarised in the follow- ing table : — TABLE 76. — Experiments at Eothamsted with Pigs. June to August 1854. Quantities per head per day. Periods. Foods. Nitrogen in food. Urea voided. Urea= nitrogen. Days. 3 3 10 10 No. 1. LentU-meal . No. 2. Barley-meal . No. 1. Lentil-meal . No. 2. Barley-meal . grams. 123.0 58.9 120.6 51.2 grams. 134.0 61.5 141.0 52.1 grams. 62.6 28.7 65.8 24.3 340 THE EOTHAMSTED EXPERIMENTS. Lidiig's mew not confirmed. The result was, then, that with exactly equal conditions as to exercise, both animals being in fact at rest, the amount of urea passed by the one feeding on the highly nitrogenous lentil-meal was, in each case, more than twice as great as that voided by the one fed on the barley-meal, supplying less than half the amount of nitrogen. It was clear, therefore, that the rule laid down by Liebig, and so long generally adopted by others, did not hold good, namely, that — " The sum of the mechanical effects produced in two individuals in the same temperature is proportional to the amount of nitrogen in their urine; whether the mechanical force has been employed in voluntary or involuntary motions, whether it has been consumed by the limbs or by the heart and other viscera " — unless, indeed, as has been assumed by some experimenters, that there is, with an increase of nitro- genous substance in the food, an increased amount of mecha- nical force employed in the " involuntary motions " suf&cient to account for the increased amount of urea voided. It was at any rate obvious that, if the amount of urea voided by one animal at rest could be more than twice as great as that voided by a similar animal also at rest, and under otherwise equal coiiditions, provided only that the food of the one contained more than twice as much nitrogen as that of the other, the amount of urea passed could not be any measure of the amount of muscular power exerted. The subject was taken up again at Eothamsted in 1862, and accordant results were obtained as follows : — Later trials. TABLE 77. — Expbriments at Eothamsted with Pigs. August-September 1862. Quantities per head per day. Periods. Poods. Nitrogen in food. Urea ■ voided. Urea= nitrogen. Days. 10 10 5 5 No. 1. Barley and bran No. 2. Beans and bran No. 1. Barley and bran No. 2. Beans and bran grams. 41.6 66.0 46.2 82.5 grams. 43.6 89.6 52.3 116.6 grams. 20.4 41.8 24.4 54.4 DrB. Smith'i trials. Not long after the publication of our views in 1852, and the experiments with pigs in 1854, with the results of which he was acquainted, the late Dr Edward Smith instituted ex- periments to determine the amounts of carbonic acid exhaled in respiration under various conditions as tp muscular exer- cise. His results were published in a paper presented to the FEEDING OF AKIMALS. 341 Eoyal Society on December 16, 1858.^ He records the Muscular quantities of carbonic acid exhaled in grains per minute, and ^''^^. these we have calculated into grams per hour, and so give tuim of them below :— ""^^ During light sleep Lying down, scarcely awake Sitting quietly . Walking two miles per hour "Walking three miles per hour On treadwheel, ascending 28.65 feet per minute acid. Carbonic acid, grams per hoar. 19.2 23.0 38.1 70.4 100.4 189.2 There was, therefore, very greatly increased exhalation of carbonic acid with increased muscular exercise. Dr E. Smith also conducted experiments on the amounts Labour of urea eliminated under different conditions, both as to food ?"?„^!^„ J . rrii • • - T • T *^ ^J urea. and exercise. Ihe investigation was commenced in January 1860, and continued up to March 1862, a period of two years and two months. These results were also published in a paper in the Philosophical Transactions of the Eoyal Society.^ The general result was, that there was great variation in the amount of urea passed when there was concurrent variation in the amount of nitrogenous substance in the food ; but, on the other hand, comparatively little variation in the amount of urea voided, with great variation in the amount of labour performed. Thus, then, Dr Smith's results, both those showing the Confirming amounts of carbonic acid exhaled, and those relating to the ^^^^ amount-s of urea voided, fully confirmed the view that with muscular exertion there was marked increase in the demand for the non-nitrogenous, and but little if any in that for the nitrogenous, constituents of food. Experiments made by Eischoff and Voit in 1858 and VoU'sex- 1859 ® with a dog, either submitted to hunger, or fed from -P^""**"'*- time to time on foods containing very different amounts of nitrogenous substance, showed very variable amounts of urea voided, although the animal was kept under equal conditions as to exercise. Still, on the publication of their results in 1860, the authors assumed, that although there had been no Voifs greater exercise of force manifested in the form of external '"*^*- work, yet when the amount of nitrogenous substance in the food was greater, and the amount of urea voided correspond- ingly greater, there must have been a corresponding increase in the force exercised in the conduct of the actions within the 1 PhU. Trans., 1859, vol. 149, pp. 681-742. 2 rhU. Tram., 1861, toU 151, pp. 747-834. ' Die Gesetze der Ernahrung des Fleischfresaeri, 1860. 342 THE EOTHAMSTED EXPERIMENTS. with Voit. Further trials by Voit. body, in connection with the disposition of the increased amount of nitrogenous substance consumed ; so that, after all, the amount of urea eliminated was a measure of the exercise of force, though not in the voluntary exercise of muscular power. One of us being in Germany in the summer of 1860, and visiting Munich, had some conversation with Professor Voit on the subject of their results and conclusions. Eeferring to our own results obtained in 1854 with pigs, it was pointed out that they were entirely consistent with those which he and Professor BischofP had obtained with a dog, but that we had drawn very different conclusions from them. He con- veyed the impression, however, that he considered we were entirely, in error. Later in the same year, however, Voit published ^ the results of further experiments with a dog. In these, he submitted the animal to alternate rest and labour, sometimes fasting, sometimes with a moderate, and sometimes with a liberal supply of nitrogenous substance in food. The labour con- sisted of working in a kind of treadwheel. He found that the amount of urea eliminated was not in proportion to the exercise of force, but was very nearly proportional to the amount of nitrogenous substance consumed. He considered that by such a result the views which he and others had maintained as to the connection between the exercise of force, the degradation of nitrogenous substance within the body, and the elimination of urea, were completely overturned. In 1862 Pettenkofer and Voit published a paper ^ giving the results of experiments with a dog made in 1861 and 1862, in which the food consumed, the amount of urea voided, and the quantity of carbonic acid given off by the lungs and skin, were determined — the latter in Pettenkofer's respiration apparatus. These experiments were more or less preliminary, and during their conduct the animal was not submitted to any labour. Subsequently, Pettenkofer and Voit made experiments in pT/* ^f "^^^^^ ^^sy determined both the nitrogen in the urine, and --■ the carbonic acid evolved, not only in rest but in work; sometimes fasting, and sometimes with food. Their results were published in 1866 in the Zeitschrift fiir Biologie. Table 78 gives average results for twenty-four hours, in experiments made with a man, with the aid of Pettenkofer's respiration apparatus. Thus, not only was there no increased transformation of ' XJntersuchvmgen ilber den Einfluss des Kochsalzes, Kaffees imd der Muskel- iewegungen cmfden StoffwecAsel, 186Q. " Ann. Ohem. Pharm,, II. Supplement-band, I. Heft, p. 52. Former mews over- twrned. Bxperi- and Voit. FEEDING OF AisIMALS. 343 nitrogenous substance by the exercise of force, but there was OmfimUng a very greatly increased exhalation of carbonic acid. It is ^^^^ evident, therefore, that in the exercise of force, the exigency of the system is specially characterised by an increased demand in the food for, so to speak, respiratory material The results of Pettenkofer and Voit are indeed of great importance ; but in Germany they are even looked upon as being the first to establish the correct view on the subject. TABLE 78. I Nitrogen in mine. Carbonic add exhaled. EN" HUXGER. la rest . In work . grams. 12.39 12.26 grams. 716 1187 WITH MODERATE DIET. In rest . In work . 17.01 17.33 928 1209 Abundant further confirmation of the now generally accepted view is available, and it will be of interest to give some illustrations. In 1866 results were published ^ as to the amount of Jt^stOis in nitrogen excreted before, during, and after ascending the ^^^ Faulhorn, by Professor Pick and Wislicenus, in August 1865. The experimenters took an ordinary meal at mid-day on the 29th, but then only starch, fat, and sugar until after the ascent, which commenced early the next morning. Table 79 is a summary of the results so far as they relate to the point under consideration. The record of the actual quantities is sufficient to show that much less nitrogen was excreted by both experimenters during, and after, than before the ascent. But the calculated amounts of nitrogen excreted per hour during each of the periods, as given in the last column of the table, bring the main results more clearly to view. It is seen that, on the average, only about two-thirds as much nitrogen was excreted per hour during and after the ascent, as prior to it, when there would be more or less residue in the system from the last albuminous meal. The above results of Pick and Wislicenus were brought FranMand forward by Professor Frankland, in a lecture which he gave ^^ , at the Eoyal Institution in 1866 — On the Source of Muscular muscular ■gawer. 1 T%a. Mag., 1866, 4th Series, vol. 31, pp. 4S5-503. 344 THE EOTHAMSTED EXPEKIMENTS. Poiver. He subsequently himself made numerous calori- metrical determinations of the energy evolved by the combus- tion of muscle, urea, and various foods, or constituents of foods, the results of which were published in a paper — On the Origin of Muscular Power} Stated in a few words, his main conclusion was, that the transformation of muscular tissue alone cannot account for. more than a small fraction of the muscular power developed by animals. TABLE 79. Total nitrogen. Nitrogen excreted per hour (average). FICK. grams. grams. grams. grams. Night before ascent 12.4820 5.8249 6.9153 0.63 During ascent 7.0330 3.2681 3.3130 0.41 After ascent .... 5.1718 2.4151 2.4293 0.40 Night after ascent ... 4.8167 0.45 WISLICENUS. grams. grams. grams. grams. Night before ascent 11.7614 5.4887 6.6841 0.61 During ascent 6.6973 3.1254 3.1336 0.39 After ascent .... 5.1020 2.3809 2.4165 0.40 Night after ascent 5.3462 0.51 Kellner's experi- ments. Dr Oskar Kellner, who was one of Professor Emil von Wolff's associates in numerous investigations with animals at Hohenheim, made experiments there with a horse ^ from June 15 t6 August 10, 1878. The daily food of the animal con- sisted of 5 kilog. meadow-hay, 6 kilog. oats, and 1.5 kilog. wheat-straw-chaff. The horse was made to go different dis- tances, and to draw different weights, the draught being mea- sured by a horse-dynamometer. Table 80 gives a summary of some of the conditions and results of the experiments. In reference to these results, which certainly do show an excretion of increased excretion of nitrogen with increased work during the with^^- second, third, and fourth periods, as compared with the first creased and fifth, Kellner considers that they are inconsistent with the conclusions of Pettenkofer and Voit, and others, which connect muscular action more exclusively with the oxidation of non-nitrogenous matters, and that those views require to be modified. At the same time, admitting that the transfor- 1 Phil. Mag., 1866, 4th Series, vol. 32, pp. 182-199. ^ Lwndwirthschaftliche Jahrbilcher, vol. viii., part v., 1879, pp. 701-712. Increased FEEDDsG OF ASIUALS. 345 mation of oiganic substance is to be considered the source of muscular power, he considers that, in the first line, comes the oxidation of non-nitrogenous matters, carbohydrates and fat; in the second, the transformation of circulation-albumen ; and lastly, that of the organised albumin, which is only attacked if other matters are not available in sufficient quantity. Further, he considers it is evident that the increased albumin transformation was not sufficient to cover the requirements of the increased work, and that this increased transformation, and the loss of body-weight, show the insufficiency of the food, and of the available fat of the body. TABLE 80. Experiments. Kumbep of days. Live-weight. Per day. Work done, kilo- gram-metFes. Urine voided. Xitrogen in mine. 1 2 3 6 10 14 Mog. 534.1 529.5 522.5 kg.-m. 475,000 950,000 1,425,000 cc. 6730 6473 8106 grams. 99.0 109.3 116.8 4 12 508.8 950,000 8686 110.2 5 14 518.0 475,000 9548 98.3 The table, in fact, does show that, with increased work done, there was decline in body- weight ; and, assuming with Kellner that there was a deficiency of food and of body fat, it seems probable that the increased elimination of nitrogen in the urine is the necessary coincident of real dilapidation of the system. It is obvious that, so far as this is the case, the resiilts are not discordant with our own early view on the subject, since fully established by others. These results of Kellner's are, indeed, a confirmation of the view we put forward in 1852, that " a somewhat concentrated supply of nitrogen does, however, in some cases, seem to be required when the system is overtaxed ; as for instance when, day by day, more labour is demanded of the animal body than it is competent without deterioration to keep up." In 1885 Grandeau and Leclerc published the results of an experiment with a horse ^ of which the following is a summary : — yitrogen in nrine for 100 in food. Best ... .... 62.4 per cent Walking ... . 67.7 n Trotting .... . 64.9 n T. < Walking ... . 60.9 „ Drawing I ^„^| ... . gg.g „ Probable explana- tion of KeUner's results. Grandeau and Le- clerc's ex- periments. ^ Annates de la Science Agronomique, 1885, 2"!« annie, tome i. p. 326. 346 THE KOTHAMSTED EXPERIMENTS. Zuntz and LehmaiwCs merits. The results show, over the first three experiments, some, but not great, variation in the amount of nitrogen eliminated with exercise; hut the amounts are less in the fourth and fifth experiments, and almost identical with walking and trotting. Upon the whole, there is no evidence of direct con- nection between the amount of exercise of force and that of nitrogen eliminated in the urine. The next results give very definite evidence as to the con- nection between the amount of carbonic acid exhaled,' and that of the force exercised. The experiments were made with a horse, by Zuntz and Lehmann, in 1887 and 1888,^ and the average results were as follows : — Rest . "Work After work Carbonic acid exhaled per hour (average). With Mask. With Tracheal-oanula. 3.327 cubic feet. 2.861 cubic feet. 19.643 .1 17.291 ii 4.662 „ 3.899 ti Exhalation Thus, then, there were about six times as much carbonic fj^'^at^^" acid exhaled per hour during work as in rest ; and when the work and work had ceased, there was very great reduction in the amount rest. ' ' ' F. Smith's results. of carbonic acid given off. The following results by Professor F. Smith, of Aldershot, were published by him in the Journal of Physiology^ in 1890 :— TABLE 81. OO2 expired per hour. Pony (work, trotting). Horse (work, galloping). Horse (work, galloping). Rest . . . . Work . After work cubic feet. 0.7648 2.3954 0.4631 cubic feet. 20.6265 1.3133 cubic feet. 12.4353 1.1693 As in the experiments of Zuntz and Lehmann, quoted above, the great increase in the amount of carbonic acid exhaled during work, and the great reduction in the amount after the cessation of the work, are here again clearly illustrated. Table 82 summarises numerous results, by Professor P. Smith, with horses at different paces {loc. cit., p. 77). These strictly gradationary results, with one slight excep- tion, illustrate more clearly still the greater exhalation of carbonic acid the greater the exercise of force. ^ LandWk Jahrlucher, vol. xviii. , 1889, p. 1. ^ Vol. xi., No. 1. FEEDDTG OF TABLE A^^iIALS. 82. 347 CO2 expired per hour. Series A. Series B. Rest . Waiving Trotting Cantering Galloping • cnbic feet 1.0282 1.0972 2.9482 4.9159 14.9725 cabic feet. 1.2346 1.0586 4.8309 5.0080 Turning from the foregoing evidence of direct experiment, indicating the characteristic food requirements for the exer- cise of force, it will be of interest to give a few examples of the rations adopted as the joint result of direct experiment Adopted and large experience. ToiUms. At p. 345 the results of some experiments by Grandeau and Leclerc with a horse were given, showing no direct connection between the amount of force exercised and that of nitrogen eliminated in the urine. Their experiments were made at the establishment of the Petites Voitures Company in Paris ; and the following table gives the standard daily ration of the Mationsjoi horses at the time, the experimentally determined mainten- ^^ *" ance ration, and that finally adopted for work : — TABLE 83. Ration. Beans. Oats. Uaize. Uaize- cake. Hay. Straw. Total food. Total dry substance. Previous . lb. 1.54 lb. 7.23 lb. 5.34 lb. 1.06 lb. 3.84 lb. 2.09 lb. 21.10 lb. 18.14 Maintenance, No. 1 . Maintenance, No. 2 . 0.93 0.84 4.34 3.91 3.20 2.88 0.63 0.57 2.30 2.07 1.24 1.12 12.64 11.39 10.87 9.79 For work . 1.39 6.51 481 0.95 3.46 1.87 18.99 16.33 It seems that the system of the establishment was to work the horses alternate days; and to give less hay, straw, and maize, but more oats and beans, though less total food, on the days of work. The figures in the top line, representing the "Previous" ration, are, in each case, the means of the two days' ration. The " Maintenance Eation, No. 1," was fixed at three-fifths of the "Previous" ration; but, as the animals gained in weight, the "Maintenance Eation, No. 2," which 348 THE EOTHAMSTED EXPERIMENTS. was one-tenth less than No. 1, was subsequently adopted. Even then the horses rather gained in weight. Finally, as it was considered that the standard or " Previous " ration was too high, the ration for work, as given in the bottom line of the table, which is one-and-a-half time as much as the " Maintenance Eation, No. 1," and about one-tenth less than the "Previous" ration, was adopted. It is, however, said that under the new regime the horses were somewhat under- fed, but whether the reduced ration is still maintained we are not aware. It will be observed that the proportion of the highly nitrogenous leguminous corn (beans) was very small compared with that of the gramineous grains. Still, it will be seen presently that the proportion was very consider- ably higher than in the case of the omnibus horses of Paris. The following table gives the average daily ration of the horses of the General Omnibus Company of Paris for each of the six years— 1879, 1880, 1884, 1885, 1886, and 1887. The average number of horses was about 13,000, and their average weight was from 1200 to 1240 lb., whilst, so far as the evi- dence goes, those of the Petites Voifures Company weighed little more than two-thirds as much ; and certainly the former are much heavier than as a rule are the omnibus or tramway horses of our own country. The figures are calculated from the results given in the annual reports of M. E. Lavalard,^ the general secretary of the company, the quantities being converted from kilograms into their equivalent in English pounds : — TABLE 84. Beans. Oats. . Maize. Hay. straw. Bran, ■ &c. Total food. Total dry substance. lb. lb. lb. lb. ' lb. - lb. lb. lb. 1879 1.36 10.04 6.85 9.14 10.45 ... 37.84 32.17 1880 1.41 8.84 8.25 7.80 11.10" 37.40 31.83 1884 1.44 8.67 8.53 8.44 8.71 0.91 36.70 31.29 1885 0.89 6.21 11.30 8.50 8.36 0.84 36.10 30.84 1886 0.10 5.51 12.96 8.64 7.32' 0.54 35.07 30.03 1887 0.01 8.08 10.77 8.65 8.21 ... 35.72 30.52 It will be seen that the actual amount of dry substalice supplied per head per day is nearly twice as much as in the case of the Petites Voitures horses ; that is, much more in proportion to a given live-weight. It will be further seen that the proportion of beans to cereal grains is much less than in the case of the Petites Voitures horses, and was reduced to a very small quantity in the later years. In fact, the corn "• Eapports sur Us operations dii, service de la, Cavalerie et des Fourrages. FEEDING OF ANIMALS. 349 given consisted almost exclusively of oats and maize, that of the oats heing reduced, but that of the maize in a greater degree increased, in the later years, coincidently with the reduction in the amount of heans. On the occasion of a visit of one of us to M. Lavalard in 1887, it was suggested to him that the supply of the highly nitrogenous leguminous seeds might be mainly, if not exclusively, reserved for old or overworked horses ; and he subsequently informed us that he had found their use in such cases advantageous. In his annual report for 1886, published in 1887, M. Lavalard gives, on the authority of Dr Fleming, Principal Veterinary Surgeon of the army, a list of the average daily rations of horses of tramway companies in the United King- Rations dom, which are quoted in the following table from Dr Plem- ^^ammv^ ing's book.^ We have also calculated the quantity of dry iwrses. substance in the total food according to the supposed average composition of each. There can be little doubt that the average weight of tram- way horses in the United Kingdom is much less than that of the omnibus horses of Paris, and it will be seen that the quantity of total food, or total dry matter of food, given per head per day is also considerably less ; though it is much greater than in the case of the smaller Feiites Voitwes horses of Paris. TABLE 85. Beans or peas. Oats. Uaize. Hay. straw. Bran. Total food. Total dry substance. lb. lb. lb. lb. lb. lb. lb. lb. North Metropolitan 2 3 13 7 3 ... 28 24.09 London . 3 3 7 12 1 ... 26 22.20 London Street 1 3 12 11 1 28 24.09 South London 1 4 7 10 7 6 11 '3 ... 29 32 24.76 Birmmgham . 12 27.30 Liverpool 4 12 14 ... i 31 26.58 Manchester . 15 15 30 25.55 Glasgow . ... 6 11 8i i oi 27 23.24 Edinburgh 4 8 4 14 I 32 (25.56) Dublin . ... 3 14 12 ... "oi 29i 25.41 1 Also 2 lb. of " Marsblam " — (Mashlun — mixed com ?), The details show that, at any rate at that date, the tram- way horses in the United Kingdom received much more of the highly nitrogenous leguminous corn, beans or peas, than the Paris horses; and, according to the figures, this was ^ The Praaical Rorse^^eeper, by C. Fleming, LL.D., p. 88. 350 THE BOTHAMSTED EXPEEIMENTS. Remew of Gonstitu- ents of labow rations. especially the case ia Birmingham, Liverpool, and Edinburgh. Oats and maize, nevertheless, contributed most of the corn ; the maize generally predominating, whilst at the present time it will doubtless do so in a greater degree. Reviewing the whole of the results which have been adduced illustrating the characteristic food requirements for the exercise of force, it may in the first place be observed that the evidence is cumulative and decisive that, with normal feeding, and with only moderate exercise, there is practically no increased demand for the nitrogenous con- stituents of food ; vyhilst there is, on the other hand, an increased demand for the more specially respiratory con- stituents, largely in proportion to the amount of force exercised. If, however, the labour is abnormally heavy — that is, if it be pushed to the point of dilapidation, as indicated by loss of weight — there will, in that case, be an increased elimination of nitrogen in the urine, resulting from the degradation of nitrogenous substance, and accordingly an increased demand for the nitrogenous constituents of food. Lastly, it is of interest to observe, that where the subject has been the most carefully investigated, the rations adopted for horses include scarcely any of the more highly nitrogenous foods, such as leguminous seeds ; but, in addition to hay and straw-chaff, consist almost exclusively of the comparatively low-in-nitrogen cereal grains, and would, therefore, be char- acterised by containing a comparatively large amount of digestible non-nitrogenous constituents in proportion to the digestible nitrogenous substance of the food. It has, however, been found that in the case of old or overworked animals, it is advantageous to supply a somewhat larger amount of the highly nitrogenous leguminous seeds. In fact, as we put it in 1852 — " a somewhat concentrated supply of nitrogen does, however, in some cases, seem to be required when the system is overtaxed ; as for instance, when day by day more labour is demanded of the animal body than it is competent without deterioration to keep up." Summary on the Feeding of Animals. In introducing the subject of the feeding of the animals of the farm, attention was first called to the amount of the con- stituents of the crops grown in an ordinary four-course rota- tion, which would, if the grain only were at once sold, be retained upon the farm for further use — in fact, for the pro- duction of meat, milk, and manure, and for the exercise of force. There will, as a rule, be a greater or less amount of grass in admixture with the arable land of the farm ; and. FEEDING OF ANIMALS. 351 according to its amount and other circumstances, there will, of course, be more or less stock-food available in addition to that produced on the arable land. So far as manure is con- cerned, in some cases the grass-land, and in others the arable, will be the gainer by the admixture of the two, accordingly as the one or the other receives back more or less than the amount derived from the consumption of its own produce. Then, again, the influence of the growing modern practice of selling more than the grain, and of importing cattle food and manure from external sources, has to be taken into account. Nevertheless, the illustration derived from a consideration of the proportion of the constituents of the crops grown under a particular system of rotation, which will probably be available for feeding purposes, is not without interest and utility. The facts and arguments which have been adduced may be Relative very briefly summarised as follows. It has been shown that o/^^^^. the amount of food consumed, both for a given live-weight of ous and animal within a given time, and for the production of a given »"»'-»*<'- amount oi increase, is, as our current rood-stuns go, measui- stuuenu. able more by the amounts they contain of digestible and available non-nitrogenous constituents, than by the amounts of the digestible and available nitrogenous constituents they supply. That this should be the case, so far as the consumption for a given live-weight within a given time is concerned, seems consistent enough when the prominence of the respiratory function in the maintenance of the body, and the large re- quirement for non-nitrogenous constituents of food to meet the expenditure by respiration, are borne in mind. But, at first sight, it seems less intelligible that the quantities con- sumed to produce a given amount of increase in live-weight, should also be much more dependent on the supplies of the non-nitrogenous, than on those of the nitrogenous constituents of food. It has been shown, however, that store animals may contain PrmmiUm as much, or even more, of the non-nitrogenous substance — ^^-^^ """^ fat — ^than of nitrogenous substance ; whilst the bodies of fat- matter in tened animals may contain two, three, four, or more times as j^«^f *™ much dry fat as dry nitrogenous matter. Obviously, there- '^"^^ fore, the proportion of fat to nitrogenous substance in the increase in live- weight of the fattening animal, must be much higher than in the entire bodies of the animals. Then, it has been fui'ther shown that the non-nitrogenous Source substance of the increase — the fat — is at any rate in great "ff'^^- part, if not entirely, derived from the non-nitrogenous con- stituents of the food. 352 THE ROTHAMSTED EXPERIMENTS. Propor- Of the nitrogenous compounds of food, on the other hand, ni^oCTi °^^y ^ small proportion of the whole consumed is finally retained Stored up in the increase of the animal. In other words, a and voided, ygry large amount of nitrogen passes through the body beyond that which is finally retained in the increase, and so remains for manure. It is, therefore, only what should be expected, that the amount of food consumed to produce a given amount of in- crease in live-weight, as well as that required for the susten- ance of a given live-weight for a given time, should, provided the food be not abnormally deficient in nitrogenous substance, be characteristically dependent on its supplies of digestible and available non-nitrogenous constituents. Force and Again, it has been shown that, in the exercise of force, there ■''"' ■ is a greatly increased expenditure of the non-nitrogenous con- stituents of food, but little, if any, of the nitrogenous. Food for Thus, then, for maintenance, for increase, and for the exer- mcfiT- <^ise of force, the exigencies of the system are characterised crease, and. more by the demand for the digestible non-nitrogenous or force. more specially respiratory and fat-forming constituents, than by that for the nitrogenous or more specially flesh- forming ones. Composi- In our paper — On the Composition of Oxen, Sheep, and Pigs, ^^n'she '^'"'^ "-^ their Increase whilst Fattening — published in 1860,^ 'we concluded that — if fattening oxen were liberally fed upon good food, composed of a moderate proportion of cake or corn, some hay or straw chaff, with roots or other succulent food ; if sheep were fattened under somewhat simi- lar conditions, but with a less proportion of hay or straw ; and if pigs were liberally fed chiefly on cereal grain — the increase would, with as much as 5 or 6 parts of total non- nitrogenous to 1 of nitrogenous compounds in the dry sub- stance of such fattening food, probably be very fat. Further, that in the earlier stages of growth and feeding, a lower pro- portion of total non-nitrogenous constituents, that is, a higher proportion of the nitrogenous compounds, is desirable ; in- deed, that it is frequently the most profitable, having regard both to the rapidity of fattening and to the value of the man- ure, for the farmer to employ, even up to the end of the feed- ing process, a somewhat higher proportion of nitrogenous constituents in his stock-foods, than is necessary to yield the maximum proportion of increase in dive-weight for a given amount of dry substance of food consumed. But that, when the mixed fattening food contains less than about 5 parts of non-nitrogenous to 1 of nitrogenous compounds, the propor- ' Jour. Hoy. Ag. Soe. Eng., 1st Series, vol. xxi., 1860, p. 433. FEEDING OF ANIMALS. 353 tion of increase in live-weight for a given amount of dry sub- stance of the food will not increase with the increased pro- portion of nitrogenous compounds consumed ; whilst, so far as these are in excess, the proportion of carcass in the live- weight will probably be somewhat less, and the carcasses themselves will be somewhat more bony and fleshy, and less fat. We at the same time pointed out, however, that the com- Estinuu- parative values of food-stuffs, even as stich, could not be uncon- ^^^^^^ ditionally determined by the percentage of the total nitro- genous and the total non-nitrogenous constituents; that it was necessary — to examine more closely into the nature and condition of the proximate compounds of food-stuffs ; to distinguish those which are digestible and assimilable from those which are not so ; to determine the relative values of the comparable or mutually replacable portions ; and, finally, to fix our standards of comparative value with more of refer- ence to direct experimental evidence on the point, and to existing knowledge of the composition of the animal bodies, than had hitherto been usual or even possible. Since then, an immense amount of labour has been ex- pended in the determination of the digestibility of the indi- vidual constituents of various food-stuffs ; and the results so far obtained form a valuable contribution to our information on the subject. There is, however, wide variation in the composition of different samples of nominally the same description of food. Then, the determinations of the amounts of the various constituents remaining undigested have gener- ally been made with animals fed on limited supplies of food, for maintenance only ; and the experiments have frequently been made with the individual foods given separately. Great Necessity care and reservation are, therefore, necessary in the applica- ■^^L^^^T tion of the results to actual practice. Thus, in the liberal tsUmaies feeding of animals for the production of increase, it is gener- "•^/""f ally economical to give, within limits, an excess of food, if a maximum result is to be obtained for a given live-weight of animal within a given time ; and, in the case of animals liber- ally fed for the exercise of force, there will also generally be an excess of food given. It is obvious that, under the conditions of actual practice here assumed, greater proportions of the various constituents consumed will remain undigested than would be indicated by the figures representing indiges- tibility obtained under the usual conditions of experimenting on the point above referred to. Then there is the important consideration, that conclusive evidence is still wanting as to the exact role in the system of some prominent constituents of food-stuffs. For example, there is yet much uncertainty VOL. vn. z 354 THE EOTHAMSTED EXPERIMENTS. Uncer- in regard to the position of the various amides, which enter J^iZlf '^o largely into the composition of feeding roots and hays — in food con- fact of all succulent and unripened products. Indeed, in the stituents.. calculation of " nutritive ratios," the amides have sometimes been classed with the albuminoids, and sometimes in large proportion with the non-nitrogenous constituents. We have, from time to time, had the results of our numerous feeding experiments, with both sheep and pigs, calculated according to the published tables of digestibility. But the so-cal- culated " ratios " varied so considerably for different rations within the range of good practice, that it would be mis- leading to attempt to give anything like a summary of the results, and general conclusions therefrom, without full discussion. Relative In conclusion, as our current fattening food-stuffs go, as- vaiue of suming, of course, that they are not abnormally low in the aid nm- nitrogenous constituents, they are, as foods, more valuable in nitrogenous proportion to their richness in digestible and available non- ents. nitrogenous than to that of their nitrogenous constituents. As, however, the manure of the animals of the farm is valu- able largely in proportion to the nitrogen it contains, there is, so far, an advantage in giving a food somewhat rich in nitro- gen, provided it is in other respects a good one, and, weight for weight, not much more costly. Missing Page Missing Page INDEX. Addresses to Sir John B. Lawes and Sir J. H. Gilbert .... 8 Albuminoid-ratio in cereals and roots 66 Ammonium salts — and nitrate of soda compared 29, 111, 172 exbanstion of . . . . 177 for leguminous crops . Ill, 114 for sugar-beet .... 33 for swedes ... 26, 28 for -wheat 172 for white turnips . . 21 Analysis of soil .... 95 Analytical work, amount of 6 Armsby on food and feeding . 313 Barley— and clover after beans . . 124 carbohydrates in . . . 106 carbon in 106 compared with wheat . 68, 99 composition of, as influenced by manures and seasons . . 85 dry matter in rotation and con- tinuous crops . . . 218 effects of artificial manures on . 73 experiments with ... 68 general results with artificial manures .... 80 habit of its growth . . 69, 100 Influence of seasons . 71, 76, 81, 85 in rotations .... 205 most effective manures for 99, 100 nitrogen for . . 72, 99, 100 II in rotation and con- tinuous crops . . . 225 phosphoric acid in rotation and continuous crops . . _ . 231 potash in rotation and continu- ous crops .... 237 produce of, without manure and with dung . . ■ • ,S^ soils for . . • . -100 summary of results with . . 99 weight of, as influenced by season 87 yield of nitrogen m . . . lUl Beans — carbohydrates in . . . 106 carbon In 106 continuous cropping with . 114 dry matter in rotation and con- tinuous crops . . . 218 effects of nitrogenous manures on Ill experiments with . . 102, 112 failure of, when grown at short intervals . . . 112 influence of fallow on . . 116 in rotations .... 208 liability to parasitic attacks . 112 lime in rotation and continuous crops 244 nitrogen in rotation and con- tinuous crops . . . 225 phosphoric acid in rotation and continuous crops . , . 231 potash in rotation and contin- uous crops .... 239 yield of nitrogen in , . 101, 114 Beginning of Eothamsted experiments 3 Bequest by Sir John B. lawes for continuing experiments . . 7 Boussingault's experiments — as to fixation of free nitrogen . 137 on stock-feeding . 267, 284, 327 Carbohydrates — and milk production . . 318 and the formation of fat, 110, 284, 289, 293, 303, 310, 312 in various crops . . . 106 of plants and supply of nitrogen 46 Carbon in — food and in increase in weight' ofpigs 289 soils and yield of wheat . . 189 various crops . . . 106, 175 Carbonic acid, exhalation of, and muscular exercise . . . 341 Cattle- compared with sheep and pigs in feeding .... 270 researches in feeding . . 255 356 INDEX. Cereals — albuminoid-ratio in . . .66 dry matter in . . . 218, 224 Chlorophyll-formation in wheat . 175 Clover — and barley after beans . . 124 continuous cropping of . . 118 dry matter in rotation and con- tinuous crops . . . 218 effects of nitrogenous manure on 117 enriching soil in nitrogen 104, 125 experiments with . . , 102 following beans . . . 123 grown at short intervals . . 116 in garden soil .... 113 in rotations .... 208 lime in rotation and continuous crops 244 nitrogen in . . . 101, 128 11 in rotation and con- tinuous crops . . . 225 phosphoric acid in rotation and continuous crops . . . 231 potash in rotation and continu- ous crops .... 239 soil-source of nitrogen for . 123 variations in crop of . , 117 Clover sickness 103, 117, 118, 121, 163 Composition of barley as influenced by — manures and seasons . . 85 maturation .... 87 Composition of — live stock in process of fatten- ing ... . . 275, 352 white turnip's . . . .22 Composition of the roots and leaves of — mangels 53 sugar-beet .... 34 swedes and white turnips 23, 27, 30 Constituents of crops removed from and consumed on land . 255, 324 Consumption of roots on land 203, 208, 211, 214, 228 Contents, summary of . . .11 Continuous cropping — and mechanical condition of soil 234 of barley . . . ... 69 of beans .... 114 of clover 118 of swedes, effects of . . .29 of wheat . ■ . . . .166 Culture of root crops ... 20 Dry matter — in roots 66 in rotation and continuous crops 218 Dyer's analyses of Eothamsted soil S6 Fallow crops — ^ dry matter in . influence on beans . 218, 224 . . 116 in rotations . 203, 208 Farmyard manure for — barley . 70 mangels . . 50 sugar-beet 32,39 wheat . . . 170, 188 white turnips . . 20 Fat in food, and in increase of weight in pigs 288 Fat- sources of 110, 284, 288, 301, 307, 310, 312, 314, 351 M conclusions regarding 314 M views of other scien- tists . 285, 294, 312 Fattening qualities of animals. See Feeding. Feeding — and food-value of feeding-stuffs 257 and improvement in live stock . 256 and manure .... 324 constituents of crops retained on farms . . . 324 experiments at Woburn . . 332 ,f with cows . . 335 I, with dogs . . 335 n with sheep . . 330 nitrogen, amount of, in food and manure . . 327 II consumed and voided 328, 352 II exhalation and ab- ' sorption of, by ani- mals . . . 325 II not accounted for . 328 review of results as to loss of nitrogen .... 333 value of unexhausted manures . 336 Feeding and milk-production . . 314 constituents of food carried off in milk and in the fattening increase of oxen . . . 315 constituents of milk . . 315, 320 II as affected by food 323 II as affected by season 321 demands upon food . . . 316 dependence on nitrogenous sub- stances ... . . . 319 destination of excess food-supply 320 source of mUk .... 318 substances consumed for susten- ance and milk-production 317, 319 yields of milk . . . 316, 320 II as affected by food . . 323 II as affected by season . 321 Feeding and the exercise of force . 337 Continental experiments . . 344 exhalation of carbonic acid . 341 experiments with pigs . . 339 food constituents demanded for force .... . 339, 352 Fraukland's views . . . 343 Liebig's views . . . 337 rations for horses . . 347, 350 review of results . . 350 voiding of urea .... 341 Voit's views .... 341 Feeding and the sources of fat 110, 284, , 288, 301, 307, 310, 312, 314, 351 conclusions regarding . . 314 views of Armsby . . . . 313 of Henneberg . . . 297 ofSoxhlet . . .299 of Voit . . 285, 294, 312 of Weiske and "Wildt . 295 of Wolff . . 295,297, 312 INDEX. 357 Feeding — Boussingault's investigations 257, 284, 327 composition, changes in, in pro- cess of fattening . . 279, 351 composition, difference of, in growing and fattening increase 281 composition of animals at differ- ent stages of maturity . . 280 composition of oxen, sheep, and pigs in process of . . 275, 352 composition of carcasses . . 277 II of increase in weight 281 1, of offal ... 277 II of the entire animal 278 ti relation of the min- eral matter and the nitrogen- ous constituents of the body . 275 Continental researches . 257, 262 economy in, and increased pro- duction 256 equivalent rations . . . 259 Feeding experiments with cattle . 255, 275 Feeding experiments with pigs 269, 288, 339 as to -carhon in food, and in increase in weight . . 289 as to fat in food and in increase in weight . . . .288 hy Weiske and Wildt . . 295 conclusions . 271, 274, 289, 294 diagrams explained . . . 271 increase in weight . . . 270 nitrogenous and non-nitrogen- ous substances as substitutes for each other . . .274 II relative influence in increasing weight . 283 M relative proportions of, in composition of animals . 283, 354 II relative value of 274, 351, 354 II results from . . 272 pigs compared with cattle and sheep 270 Feeding experiments with sheep 265, 305, 330 fat-formers the regulating factors 268 fat in food and in increase in weight 307 food demands for maintenance and increase in weight . 267, 352 general results . . . 268, 312 Wolff's tables re-calculated . 269 Feeding — fat in animals and foods . . 284 fattening qualities of animals . 286 Liebig's views . . 260, 284, 337 nitrogen in foods . 258, 308, 327 produce retained lor . 256, 324 relation to rotation farming . 255 Rothamsted experiments . . 261 II points embraced in 263 summary of researches on . 350 Thaer's experiments . . . 257 Fertility, accumulated . . 192 Field experiments — list of,- Table I. . . 14, 15 not reported on . . .18 plan of 14 Foods— and manure .... 324 and the exercise of force . . 337 carbon in 289 classification of . . . 260 digestible and indigestible nitro- genous and non -nitrogenous substances in . , 259, 306 equivalent rations . . . 260 fat in 284 feeding value of . 63, 257, 353 for maintenance, increase, and force 352 for milk-production . . . 314 function of different ingredients 353 Liebigon . . . . . 261 necessity for mixing ... 67 nitrates and value of . . 309 nitrogen in . . 258, 308, 327 nitrogenous and non -nitrogenous substances as substi- tutes for each other 247 II relative influence of, in increasing weight 283 II relative feeding value of . . 274, 351, 354 nutritive ratio of . . . 273 Force, feeding for . . . . 337 Frankland on the source of muscular force 343 Garden soil — experiments with clover in . 118 • nitrogen in ... . 121 Gilbert, Sir J. Henry — joining Sir J. B. Lawes . . 5 honours to .... 9 portrait of . . facing 19 Gramineous crops .... 68 Hellriegel's researches . . . 138 History of.Eothamsted Experiment Station 1 Horses, rations for . . . 347, 350 Improvement in live stock Introductory statement . 256 13 Jubilee Boulder . . .8 illustration of . . . facing 8 Jubilee of Rothamsted Experiments 7 Laboratory at Rothamsted . . 6 illustration of . . . . 4 Lawes, Sir John B. — bequest by, for continuing the experiments .... 7 descent of family ... 2 education and early tastes . 2 honours to .... 7 portrait of . . . facing 1 Leaf and root, composition of, in — mangels 53 swedes . . . . 27, 30 white turnips .... 23 358 INDEX. Leguminous crops — a substitute for fallow . .212 and nitriiication . . . 131 and nitrogen in soil 104, 125, 129, 133, 211 continuous cropping with . 114 effects of nitrogenous manures on Ill experiments with . . 100, 126 failure of, when grown at short intervals .... 112 in rotations .... 208 liability to parasitic attacks . 112 limefor 249 nitrate of soda for . . . 114 nitric acid, a source of . . 134 nitrogen for . . 101, 111, 134 sources of nitrogen for 123, 125, 129, 136, 137 yield of nitrogen in . . . 101 Liebig — ' analyses of Eothamsted soil by 94 views as to rotation-cropping . 197 „ stock -feeding 260, 284, 337 Lime, function of, in — plant growth . . . 245, 248 rotation and continuous crops . 244 Live stock — feeding of .... 255 improvement of . . . 256 Lucerne, experiments with . 127 Mangel-wurzel — albuminoid nitrogen in amides and nitrates in and turnips compared carbohydrates in carbon in . effects of large and small sup plies of nitrogen II manures on ir season experiments with large yields of . leaf and root 308 308 49 106 106 52 49 51 48 51 61,53 nitrogen in, available for fat formation . . 309 II supplied and recov- ered in crop . . 56 sugar-production in . . .59 Manor of Eothamsted ... 2 Manor-House of Eothamsted . . 2 illustration . . . facing 10 Milk-production, experiments in . 314 Mineral matter in rotation and con- tinuous crops .... 231 Mineral plant-food in soil . . 95 Nitrate of soda — and ammonium salts 29, 111, 172 and rape-cake .... 57 for leguminous crops . . 114 for wheat 172 Nitrification — after leguminous crops . . 131 and nitrogen in soil . . .130 in raw clays .... 135 Nitrogen — accumulation of, in soil 72, 73, 126, 129, 131, 134 and carbon assimilation and chlorophyll formation in wheat 175 and quality of grain ... 87 and root crops .... 19 and sugar-production in beet 37, 45 and the carbohydrates in plants 46, 106 and the formation of fat . . 110 application of, to wheat . 174 Nitrogen, assimilation of — alternative explanations of . 159 and root-nodules 139, 144, 148, 151 assistance of lower organisms 155, 160 Atwater's experiment . . 137 Berthelot's views . . 144 Boussingault's experiments . 137 by higher chlorophyllous plants uncertain .... 156 from the air 51, 100, 126, 129, 131, 134, 137 fuller information required . 152 fully established . . . 153 Hellriegel's experiments . . 138 how to be explained . . 153, 159 negative results at Eothamsted 140 new doctrine regarding . . 137 II its practical im- portance . . 161 Nobbe's experiments . . 150 practical importance of . 161, 164 recent experiments at Eotham- sted 144 summary of results . 165 Nitrogen — condition of, in roots . . 62 II as influenced by manures and seasons . . 63 consumed and voided by ani- mals .... 328, 352 dependence of wheat on . . 174 do mangels draw it from the atmosphere? ... 51 evolution of free . . . 179 exhalation and absorption of, by animals .... 325 exhaustion of, in soil 42, 77, 122 fixation of free. See Assimila- tion of. for barley for leguminous crops for wheat . in arable land . in crop-residue . in foods . in garden soil . in grass lands . in residue of wheat-crop . in root-nodules in rotation and continuous crops in soil after clover . 104, 125, 129 II various leguminous crops 126, 133, 211 II wheat . . . 179 in soil, and nitrification . . 131 72, 99, 100 101, 111, 134 99, 100, 168 . 192 . 179 258, 308, 327 . 121 192 179 159 225 IXDEX. 359 Nitrogen — continued. in soils, and yield of wheat . 189 in the root of sugar-beet . . 43 in wheat soU .... 170 loss of, in drainage after wheat 169, 172, 178 ti in soil after barley . 73 II in stock feeding . . 333 missing-link regarding sources of, for crops .... 137 new doctrine regarding sources of, for crops .... 137 rapidly and slowly available . 80 removed in clover and lost in soil 122 soil enriched in, by clover 104, 125, 129 soil-source of, for clover . . 123 supplied in manure and to sugar-beet regained in crop . 48 supplied to and recovered in mangels ... 56 the main sources of . . 136 the power of roots to draw, from the subsoil . . . 135 the soil the source of, for roots and cereals .... 52 what becomes of surplus . . 177 yield of, from barley and clover after beans . . 125 II in different crops, 101, 125, 128, 169 Nitrogenous and non-nitrogenous substances in foods — as substitutes for each other . 274 digestibmty of . . . 259, 306 relative feeding-value of, 274, 351, 354 relative influence of, in increas- ing weight .... 283 Nitrogenous manures — and sugar-production iu mangels 59 effects of, in increasing produce of various crops . . . 106 effects of, on leguminous crops 111 II on wheat . . 172, 176 for barley . . 78, 99, 100 for crops poor in nitrogen 46, 109 for mangels .... 50 for sugar-beet . . . 32 for swedes ... 25, 29 for wheat. ... 99, 100 for white turnips ... 22 influence on non-nitrogenous constituents of crops . . 172 Vines' views on the influence of 109 Nitrogenous substances and milk- production 319 Norfolk white turnips, experiments with 20 Phosphoric acid — accumulation of, in soU . . 95 and quality of grain ... 88 in barley as influenced by man- ure and season . . 86, 88 iu root and leaf of sugar-beet 38, 43 in rotation and continuous crops 231 in wheat . . . 183 Pigs, compared with cattle and sheep iu fattening 270 experiments with . 269, 288, 339 Potash — ■ accumulation of, in soil . . 95 and quality of barley . . 88 and sugar-production in mangels 59 II iu beet 37, 44 for barley ... 74, 78 for mangels . . 50 for sugar-beet . 33 for swedes . . 28 for wheat 173 in barley as influenced by season and manure . . . 86, 89 in rotation and continuous crops 237 in sugar-beet .... 42 in wheat .... 183 residue of, in wheat soil . 184 Potatoes — carbohydrates in . 106 carbon in . . . . 106 Prefatory Note by the Editor . . 1 Produce of — barley 69 crops grown in rotation and contmuously. . . . 218 crops in different countries . 197 leguminous crops . . . 101 sugar-beet . . 33, 39 sugar from sugar-beet . . 43 swedes . . 25 wheat . . 166 white turnips . 20 Rape-cake— accumiHation from . . 28 and nitrate of soda ... 57 and sugar-production in man- gels . . . 59 for barley . . . 79 Rations — equivalent .... 259 for horses .... 347, 350 Reports on Rothamsted experi- ments .... Reverting turnips . Root-crops — abnormal root-development in advantage of, in a rotation albuminoid ratio in . and nitrogen conditions of growth of . culture of . dry matter in . experiments with general conclusions from experi- ments with .... importance of . pre - eminently dependent on manure . reverting . sugar in . . yield of nitrogen iu . Root-nodules — and microbe-infection of soil and seed . . 139, 144, and sources of nitrogen 139, 144, 148, 161 13 . 20 . 19 21,58 . 66 . 19 19 20 66 19 67 19 58 20 66 101 ,150 360 INDEX. Root-nodules — continued, form and position of growth of . nitrogen in physiological meaning of various nodule - forming bac- teria Eoots and leaves of — mangels . sugar-beet swedes white turnips . Eotation crops and stock-feeding barley in . benefits of Boussingault's investigations constituents of crops removed from and consumed on land . consuming roots on land 203, 207, 211, 214, 228 Daubeny's researches . . 198 dry matter in rotation and con tinuous crops effects of manures illustrated experiments on faUowipg . four-course general conclusions . history of . importance of . leguminous crops in . Liebig's view , lime in rotation and continuous 152 157 159 151 152 . 53 . 34 25, 29, 30 23,27 17 255 205 197 198 255 218 204 199 203, 208 196 249 195 195 208 197 crops ... mineral matter in rotation and continuous crops . nitrogen in. rotation and con^ tinuous crops . . . 225 phosphoric acid in rotation and continuous crops . . . 231 potash in rotation and continu. ous crops .... 237 silica in . . . . 251 swedes in 200 wheat in 212 yield of crops in Britain and » foreign countries . . . 197 Rothamsted — Manor 2 Manor-Hou^e .... 2 11 illustration of facing 10 Trust 7 Sample-House, illustration of . Samples of soil, grain, &c. and produce of straw and soda in barley . and weight of barley character of good and bad effect of, on barley . 71, 76, 11 on mangels 1 1 on milk-production 11 on wheat . Sheep — compared with c.attle and pigs in feeding .... experiments with . 265, 305, 244 231 97 88 87 83 81,85 51 321 186 270 330 Silica in — barley- rotation crops Soda in — barley wheat Soil- analysis of for barley , for wheat . Soxhlet on food and feeding . Staff employed at Rothamsted Stock-feeding. See Feeding. Straw — season and produce of II and ash in silica in . strength of Sugar-beet — carbohydrates in carbon in ■ composition of leaf and root effects of manure - residue and crop-residue . experiments with nitrogen supplied to, in manure and regained in crop produce of, from direct manuring and residue action produce of, from dung and other manures II leaf II sugar from proportions of leaf and root sugar-production in . Sugar-production in^ mangels .... sugar-beet various roots Summary of subjects dealt with Superphosphate — and sugar-production for barley for mangels for spring crops for sugar-beet . for swedes for wheat . for white turnips Swedes — accumulation in root of . .25 and white turnips compared 25, 27 composition of roots and leaves of 27, 30 dry matter in rotation and con- tinuous crops . . .218 experiments with . . 24, 28 in rotations . . . . 200 nitrogen in rotation and con- tinuous crops phosphoric acid in rotation and continuous crops . potash in rotation and continu- ous crops .... produce of roots and leaves of . M with and without manures ... 25, 28 production and economy of leaf in 27 proportion of roots and leaves in 26 88,94 . 251 183 95 100 100, 187 299 97 97 3,97 97 106 106 34 39 31 48 38 32 34 43 34 37 37,43 66 17 44, 59 74, 78, 100 . 50 . 80 . 32 . 29 . 100, 173 . 22 225 231 239 25 INDEX. 361 Svfedes— continued superiority of . . . .27 Thaers, experiments on stock-feeding 257 Trustees of the Eotharasted Trust . 7 Turnips — abnormal root-development in . 19 advantages of, in a rotation . 21 and mangels compared . . 49 and nitrogen . . . .19 composition of, with and with- out manures .... 22 conditions of growth of . .19 continuous growth of 21 culture of . . . 20 effects of different manures on . 21 II soil-exhaustion on . 21 experiments with . . 19 importance of . . . .19 introduction of . . 196 produce of white . . 20 proportion of leaf and root in . 23 reverting 20 why different varieties are grown 27 with dung .... 20 without manure ... 20 Urea voided by animals under dif- ferent conditions as to food and exercise 341 Vetches — and nitrogen in soil . . . 131 experiments with . . 127, 147 Vines' views on the use of nitro- genous manures . . . 109 Volt's views as to sources of fat 285, 294, 312, 341 Weiske and Wildt on food and feeding 295 Wheat- application of nitrogen for . 174 carbohydrates in . . 106 carbon in . . . . 106 compared vrith barley . 68, 99 Wheat — continued. connection between nitrogen ac- cumulation, chlorophyll -for- mation, and carbon assimila tion continuous cropping with 166, 188 dependence on available nitro- gen dry matter in rotation and con- tinuous crops effect of artificial manures 172, II bad seasons II dung . . . 170, II soil and locality . II thorough tillage . experiments with general conclusions . habits of growth of . in rotations loss of nitrogen in drainage 169, 172, 178 manures requisite for 99, 100, nitrogen in crop-residue . It in rotation and con- tinuous crops . II in soil phosphoric acid in . II in rotation and continuous crops . potash in rotation and continu ous crops potash, soda, and phosphoric acid in grain and straw produce of, without manure 167 soils for . . . . 100, 187 yield of, and nitrogen and carbon in soil .... II of, from heavy manuring . M of nitrogen in crop . II of, on prairie-land White turnip — and swedes compared experiments with Wolff on food and feeding 269, 273, 295, 297, 312 175 174 218 188 186 188 187 170 166 188 99 212 168 179 225 170 183 231 239 183 ,188 189 177 169 192 25 20 PRINTED BY WILLIAM BLACKWOOD A^'D SONS.