^ s 
 
 C65 
 
 f%^^. //■ <i^^:^-^/t 
 
 
 DEPARTMENT OF AGRICULTURE, 
 
 SYDNEY, NEW SOUTH WALES. 
 
 Miscellaneous Publication, No. 116. 
 
 Agricultural Experiment Work: 
 
 BY 
 
 N. A. COBB. 
 
 SYDNEi' : WILLIAM APPLEGATE GULLICK, GOVERJvTMENT^PEINTER. 
 
 1S96. 
 11* 100—96 A 
 
ALBERT R. MANN 
 LIBRARY 
 
 New York State Colleges 
 
 OF 
 
 Agriculture and Home Economics 
 
 Cornell University 
 
S 541 res ""°" ""'"''"'"■"'"'■' 
 
 Agricultural experiment work. 
 
 3 1924 001 013 055 
 
Cornell University 
 Library 
 
 The original of tiiis book is in 
 tine Cornell University Library. 
 
 There are no known copyright restrictions in 
 the United States on the use of the text. 
 
 http://www.archive.org/details/cu31924001013055 
 
DEPARTMENT OF AGRICULTURE, 
 
 SYDNEY, NEW SOUTH WALES. 
 
 Miscellaneous Publication, No. 116. 
 
 ^ 
 
 Agricultural Experiment Work 
 
 BY 
 
 N. A. COBB. 
 
 SYDNEY : WILLIAM APPLEGATE GULLICK, GOVERNMENT PEINTER. 
 
 1896. 
 
 lib 100—96 A 
 
 AV 
 
334413 
 
Agricultural Experiment Work. 
 
 By N. a. COBB. 
 
 Introduction. 
 
 The following pages have been written in response to a request from the 
 Minister for Agriculture, made at the recent meeting of the experts of the 
 New South Wales Department of Agriculture held at the Hawkesbury 
 Agricultural College. 
 
 They are more largely directed against the old-fashioned plot system of 
 agi-icultural experiment, a system which I definitely abandoned in 1875, 
 than would be necessary in a country where the influence of the plot system 
 was less dominant than in Australia. I hope this will be regarded as a 
 BufiBcient excuse for the full exposure given in some places of what, to the 
 trained mind, may seem the most obvious fallacies. 
 
 The manner of experimenting in agriculture is gradually changing — • 
 always in the direction of the laboratory ; or, perhaps, it would be better to 
 say the laboratory is gradually enlarging its scope so as to include the fields 
 of the farmer, while the conservatory is gradually attaining the dignity of 
 an agricultural laboratory. This change is found to be of the greatest 
 use ill advancing agriculture, and I would fain see these Colonies take a 
 prominent position in hastening this general advance, not merely as a matter 
 of pride in scientific research, but as a iust due to their great agricultural 
 resources and growing agricultural importance. 
 
 Beyond question the glory of this age is its experiment work, and there 
 can be no higher study, and no study of greater economic value, than that 
 devoted to the nature and possibilities of experiment. Unfortunately, this 
 is not yet sufficiently well understood. 
 
 While it is perfectly true the whole world now-a-days listens with the 
 greatest respect to the theories and projects of almost any supposed scientific 
 authority, knowing as it does that the great advances made in the last 
 hundred years have been very largely due to the unveiling of the laws of 
 nature by men of science — while this is perfectly true, there is, sad to tell, 
 another side t j the picture ; for this very credulity which gives to-day to the 
 true men of science their great opportunity, offers also to the charlatan a rich 
 field of operations. Not until the mass of the people are well grounded in 
 the principles of science can they be expected to take an intelligent view of 
 scientific experiments. To expect them to fully appreciate the nature and 
 value of scientific research is to expect what is impossible. The great 
 
di Agricultural Experiment Work. 
 
 popularity of scientific research at the present time is born, not of knowledge, 
 but of ignorance. It is due to a respect for results obtained, and not lo 
 an understanding of the principles and methods of science. It is the praise 
 of the shallow person who, having undergone a great cure, praises only his 
 cure and the surgeon, giving no thought to the great body cf facts and laws 
 established by generations of scientific research, which alone made his cure 
 possible. 
 
 This I take to be a complete reply to those who decry experiment work 
 in agriculture, because, as they say, it is beyond the understanding of the 
 ordinary farmer, and who, partly at least, for the sake of popularity, would 
 support the discontinuance of such work. Those who expect the farmer to 
 take an intelligent interest in all that is going on in agricultural laboratories 
 show a woeful ignorance of the first principles not only of experiment work 
 but of education ; and those who seek to lessen the farmci-'s respect for 
 such work, by playing upon his ignorance concerning it, are guilty of a 
 great offence against the advance of agriculture. 
 
 It is evident, then, that these pages are written mainly for experimenters 
 in agriculture. I have tried, however, to mako the various explanations as 
 •clear as possible by means of diagrams, and I hope that farmers of an 
 -experimental turn of mind will gather from them some hints of value. 
 
 The Plot System. 
 
 Per some years the plot system of agricultural experiment has been, 
 growing in disfavour. Tnis is on account of the many drawbacks due to 
 unaccountable variations in natural soils. 
 
 My first experiments in agriculture were made in 1874, and were made by 
 means of the plot system. Two years later my observations had convinced 
 me that the plot system was very unreliable. I saw the results of one year 
 apparently upset by those of the next ; nay, worse, the results of one plot 
 ■contradicted by those of its next neighbour. "Why was this ? I propose by 
 a series of examples to show how these apparent contradictions arose, and 
 why I concluded that the plot system, as usually carried out, is a failure. 
 
 As I have frequently pointed out, modern agricultural experiments are, 
 for the most part, directed toward finding out whether some one method is 
 not slightly belter than another, and it is not to be expected that the differences 
 shown by the experiments will be great and striking. However much they 
 may be ignored, the general principles of agriculture are now well understood, 
 and it is ufeless to expect in future any frequent revolutions in practice due 
 to astonishing experimental results. From this it follows that the experi- 
 menter finds himself confronted by the necessity of adopting some method 
 that will show him with exactness the small variations in crops due to 
 differences between good varieties, to different methods of preparing the 
 soil, to different dates and modes of sowing, and so forth. I venture to say 
 that nothing in recent years has so impeded the progress of agriculture as 
 the differences of opinion among agriculturists themselves as to the best 
 rules of practice, and there can be little doubt that these conflicting opinions 
 are due largely to loose reasoning on deceptive plot experiments. However 
 complicated they may be, however ill-understood they may be, the laws of 
 growth are exact and invariable. We are, therefore, driven to conclude 
 when we see such a wide variance among the opinions of farmers that 
 the methods by which these opinions were acquired are faulty, and the' chief 
 of these methods — the plot method — must stand the brunt of this criticism. 
 
Agricultural Experiment Work. 
 
 How very unreliable the results of ordinary plot experiments are, is ■well 
 shown by the following facts : — If two plots, such as those shown at 1 and 
 
 2 in Fig. 1, be sown side by side 
 for comparison it will very fre- 
 quently be found that if plot 1 
 be divided into halves — A and 
 B — there will be more difference 
 between the halves A and B than 
 between 1 and 2. Of course, under 
 such circumstances it is useless to 
 compare 1 and 2 with any hope 
 of getting valuable results. It 
 is true that if the difference be- 
 tween 1 and 2 be very great, as, 
 for instance, when both plots are 
 very sandy and barren, and 1 is 
 manured highly, while 2 remains 
 unmanured, then the difference 
 between A and B is unlikely to 
 be greater than that between 1 
 and 2. Such striking differences,, 
 however, rarely occur in modera 
 experiments. 
 
 Let us then consider, one by 
 one, the various defects showa 
 by experiment crops arranged in 
 plots. 
 
 The extra force of the wind 
 on outside plots is a thing seldom thought of or allowed for, yet its- 
 effect is sometimes consider- 
 
 A ; e 
 
 1 2 
 
 Fijf. 1. — Two plots, the first of which has been divided 
 into halves, A and B, to sliow that the difference 
 between A and B is frequently greater than be- 
 tween 1 and 2. 
 
 able. As an illustration of 
 this point, let 1 and 2 in 
 the adjacent illustration be 
 two plots, one of which. No. 
 2, is an outside plot on the 
 exposed side of the field. The 
 constant wear of the prevail- 
 ing winds, whose direction is 
 shown by the arrows, will 
 affect the two plots differently. 
 This will be most apparent 
 during the ripening of crops 
 likely to shed their seed, but 
 is true in a smaller degree all 
 through the season. 
 
 Connected with this matter 
 is the extra exposure of out- 
 side plots in general, not only 
 to winds but to light, dirt, 
 passing of traffic, and many 
 other things. The effect of 
 these agencies is usually ap- 
 preciable, and is. ofte.n very 
 marked. 
 
 Fig. 2, — Plots of grain, showing the greater exjiosure of No. 2 
 to the prevailing winds, the direction of which is shown by 
 the arrows. No. 2 is an outside plot. 
 
Agricultural Expe^imeiit Worl\ 
 
 I have shown this graphically in Fig. 3. It will be found that it is much 
 less desirable to compare plot 1 with plot 2 than to compare plot 1 with the 
 blank plot. 
 
 Fig. 3. — ^Three plots, to show the unfairness of comparing au outside plot, as No. 2, ^\^th an inside 
 
 plot, OS No. 1. 
 
 The remedy, of course, is to place a certain amount of crop between the 
 plots and the outside of the field to serve in the capacity of buffer ; yet this 
 
 is almost never done. 
 
 Pig. 4 is instructive as showing 
 how agencies entirely out of sight 
 and beyond control may affect a 
 crop in a most erratic manner. 
 There are large numbers of insects 
 which pass either the whole or part 
 of their existence beneath the sur- 
 face of the ground, feeding upon 
 the roots of crops, or sometimes 
 coming out to feed on the foliage 
 at night. The habits of these in- 
 sects remain totally or partially 
 unknown. The manner of their 
 distribution in the soil is beyond 
 our sight, and outside our know- 
 ledge. Besides insects there are 
 various other animals, such as 
 earthworms, nematodes, &c., that 
 affect crops injuriously or other- 
 wise, and these again are distri- 
 buted in the soil according to laws 
 which we do not understand. "With 
 
 Fig. 4.— Showina the uneven distribution of insects in these facta in minri T ,^-f+n,, ,„«„J„. 
 *he soil. The in.socts arc indicated by the small ! f^ mina, i Otten WOnder 
 
 elongated maridnga. at the presumption of those who 
 
 '.'■. 1 2 
 
Agricultural Experiment Work. 
 
 plant two plots side by side and expect to get results comparable with each 
 other. In the illustration, Fig. 4, the small elongated dots are supposed to 
 represent injurious subterranean grubs. Asthese occur almost entirely in plot 
 No. 1, the injury wiU, of course, be much greater to the crop growing in that 
 plot. Such cases as this might lead an experimenter entirely astray, the cause 
 o| the damage is so hidden from sight. Suppose, for instance, that a trial of 
 two sorts of seed was being made, and the variety in plot 1 yielded the less, 
 because of the attacks of the grubs. The grubs not being visible, the lower 
 yield of plot No. 1 would in all likelihood be set down to the inferiority of 
 the variety of seed used ; but the variety might in reality be the letter of 
 the two. Thus the experimenter would come to an entirely wrong conclu- 
 sion. I have mentioned only subterranean animals in this connection, but 
 subterranean plants, such as fungi and microbes, act in a similar manner. 
 
 One of the most uncontrollable losses in experiment work is that due to 
 the raids of birds. Babbits and some other small animals can be, and always 
 should be, fended out. But birds often defy our attempts to keep them 
 away. The most we can do is to place our plots so as to minimise the losses 
 due to their raids. One effective method is to place the experiment plots 
 near dwellings, or alongside a road where there is considerable traffic. The 
 
 
 
 M^ V ,> 
 
 
 
 
 ^M, 
 
 ~^ 
 
 
 ^ ■ 
 
 ^^J 
 
 &^ 
 
 
 ~- J^^ .A^l 
 
 PR^^^M 
 
 f^B^ 
 
 
 ■*-. ^ v."^"^ 
 
 Ob^B^^sH 
 
 jBKI^^ 
 
 V. 
 
 *•*■*■ "C 
 
 >^H 
 
 £^ 
 
 
 ^Ov v^ 
 
 \^^^M^a 
 
 sS^^p-, 
 
 
 t-'^'^ 
 
 ^^^ 
 
 ^^A 
 
 
 ^ 
 
 mB^^^^^^e 
 
 ^^^^aHBf 
 
 
 ^\' 
 
 ^^^Sj^M3 
 
 RlUI^ 
 
 
 ^ ,. ^ 
 
 
 
 
 V v^v'--^ 
 
 ;> -je^y^ 
 
 MT^ ' 
 
 
 
 ^^^? 
 
 ^ 
 
 
 v^ '• " 1. "^ £ 
 
 
 
 
 .^ ^x .■^ 
 
 ^^^^ 
 
 ^^^ 
 
 V 
 
 ^ "~ -^ -^ v- 
 
 J^^^WffiJ 
 
 ^^^^M^ 
 
 '^ V. 
 
 -,■' 
 
 X^^w^^^ 
 
 ^^%^^^ 
 
 v*"*- 
 
 
 j^ Ig^g 
 
 ^^^^s 
 
 
 r 
 
 "^8 
 
 ^w 
 
 1 
 
 2 
 
 ^m 
 
 m- 
 
 Fig. 5.— Two plots, 1 and 2, alongside forest land. The birds will frequent No. 2 more tliau No. 1. 3 
 
 plots should be placed at a distance from the haunts of the birds, or if they 
 must be placed near the haunts then the plots should be so located that the 
 losses will be equally distributed. It is manifest that if two plots alongside 
 woodland be arranged, as shown on this page, the birds will go more to No. 
 2 than to No. 1, and in consequence the loss will be more from No. 2 than 
 from No. 1, at any rate in cases where the crops ripen at the same time, and 
 are equally attractire to the birds. In the supposed case the losses would 
 be somewhat equalised if the plots had been placed with their ends towards 
 the trees. 
 
8 
 
 Agricultural Experiment Work. 
 
 :;!il!l 
 
 i!!;i-.!i, 
 
 iM 
 
 mm 
 
 I have spoken only of losses at harvest-tirao, but I hardly need say that 
 the losses occur also at eprouting-time. 
 
 The surface of the soil is rarely ,( 
 
 even, and surface-water from rain 
 or any other source collects more 
 or less in the depressions. In con- 
 sequence the condition of the soil 
 in such depressions soon differs 
 from the remainder of the plot or 
 field. Though this difference may at 
 first be mechanical only, it is none 
 the less eiFectual in changing the 
 character of the plants growing in 
 the depression. As the soil settles 
 together, owing to the soakage, its 
 chemical composition begins also 
 to change, and this, added to the 
 mechanical changes, has a very tell- 
 ing effect on the portion cf the crop 
 growing on such a depressed patch 
 of soil. I have attempted to indi- 
 cate such a patch in the adjacent 
 illustration, in which the crooked 
 arrow-headed line in the upper 
 right hand corner of plot 2 shows 
 
 the direction from which water due Fig- C. -Boggy land, unequally distributed in two plots. 
 
 to a shower of rain has flowed on 
 to a depressed patch, whose area is 
 
 
 'wiil liii'i' 
 
 II i 
 
 -^ij'^-i 
 
 UU/ Uiif Ul^ tUJl- 
 
 s:i. 
 
 indicated by the shading. In the 
 illustration this area is distributed 
 unequally in the plots I and 2, and 
 will affect the two differently, and 
 thus hinder a comparison of the 
 yield on the two plots, or prevent 
 it altogether. This uneven surface- 
 drainage is the source of much error 
 in ordinary plot experiments. 
 
 Closely connected with the fore- 
 going is the existence of swampy ' 
 areas on the field containing ex- 
 periment plots. Such a condition 
 of things is roughly mapped in Fig. 
 7, where the swampy patch is shown 
 as preponderating on plot No. 1. 
 The effect of the swamp would be 
 the greater therefore on plot No. 1. 
 Nothing seems more erratic than 
 the distribution of weeds. For 
 reasons it would be profitless to 
 seek after, weeds grow most un- 
 evenly, — a patch here, a patch there. ■ 
 The distribution of clouds in the 
 sky above does not seem more law- 
 less than the distribution of weeds on the ground below. "Wherever they 
 grow weeds bring about uneven conditions in the soil. They seize on the 
 
 Fig. 7.— Swampy land, unequally distributed in two 
 experiment plots. 
 
Agricultural Experiment Work. 
 
 9 
 
 soluble products of the soil, and take it away from the crop, and they 
 bring about unequal conditions of shade and moisture. Such disturbancea 
 are reduced to zero if all the weeds 
 
 
 ■fH'n 
 
 ■■^mW-- '■■■■■ 
 
 ■■'■ •■.■■y.iw'' . 
 
 8. — Showing the unequal distrihution of weeds 
 on two experiment plots. 
 
 are kept down in experiment plots, 
 
 but they are rot often kept thus f^e?J^ 
 
 completely down. "Where plots are 
 
 sown broadcast, as is often the 
 
 case, it is nearly or quite impossible 
 
 to attend to the weeds, and they 
 
 are allowed to grow as they will. 
 
 If they would grow evenly so as 
 
 to be distributed equally through 
 
 the experiment plots they might be 
 
 disregarded, but this they ne\er 
 
 do. 'J hey are much more likely to 
 
 be distributed as shown in Fig. 8, 
 
 where the left hand plot suffers a 
 
 large loss, while the other goes 
 
 nearly free. It would be useless 
 
 to expect to get a fair com])arison 
 
 between the yields of two plots so 
 
 unevenly conditioned. 
 
 The distribution of trees on land 
 
 before it is cleared is anything but 
 
 equal, and their unequal distribu- Fit; 
 
 lion gives to the soil a corres- 
 pondingly unequal composition. 
 
 "Whether we consider the unequal chemical composition brought about by the 
 
 absorption from the soil of certain 
 substances in the vicinity ot'thetrees 
 which are not absorbed elsewhere, 
 or whether we consider the unequal 
 physical condition brought about 
 by the presence of roots or fallen 
 leaves, or the long continuance of 
 shadow in one place, wo must admit 
 that trees pla}' a most important 
 part in bringing about unequal con- 
 ditions in the soil. Tet, who ever 
 hears of any allowance being made 
 in plot experiments for these in- 
 equalities ? Though they are most 
 obvious, and persist for years after 
 the land is cleared, the disturbances 
 they cause are passed over without 
 a thought. And such work is 
 dignified by the name of experi- 
 ment ! Astounding ! 
 
 Again I have attempted to show 
 the facts graphically in Pig. 9. 
 The trees shown are located to the 
 left of the centre of the figure, so 
 that of the subsequent plots Ko. 
 
 1 has had more "tree-influence" than its neighbour. The probability of 
 
 Fig 9 —Trees drawn in to show the difference in the 
 previous condition of experiment plots 1 and 2. 
 
10 
 
 Agricultural Uxperiment Work. 
 
 Fig. 10. — Fire drawii in to show varying previous 
 conditions of plots 1 and 2. 
 
 getting useful comparisons between tlie yields of tlie two plots is much 
 
 lessened by tlie former presence of the trees. 
 In cleariug tlie land the natural 
 
 disturbance in the soil due to the 
 
 growth of the trees is often added 
 
 to by burning the timber to get rid 
 
 of it. This operation lills the soil 
 
 near the fire with a great excess of 
 
 salts of potash and lime, and the 
 
 other substances occurring in wood 
 
 ashes. This excess of these salts 
 
 may oftimes be traced for years 
 
 afterwards in the extra luxuriance 
 
 of the crops in patches where fires 
 
 hare burned. If these patches fall 
 
 in one experiment plot more than 
 
 in another, as shown in Fig. 10, the 
 
 comparison of the results from the 
 
 two plots is rendered less valuable, 
 
 or even worthless. 
 
 In all the cases we have thus far 
 
 described it is natural causes that 
 
 have so entered into the previous 
 
 history of the land as to bring it 
 
 into an uneven condition. We 
 
 come now to artificial causes, such 
 
 as should be within the knowledge of the experimenter, and for which he 
 
 might more easily allow. The 
 
 > y j t ; t ■ , reader will, however, agree with 
 
 me, I think, in saying that these 
 causes are, almost equally with 
 natural causes, fertile sources of 
 error in plot experiments. Take, 
 for instance, the case illustrated 
 in Fig. 11. Last year plot 1 was 
 cropped with peas and plot 2 
 with cabbages. These two crops 
 act in markedly difierent ways 
 on the soil, so that even if the 
 soil of these two plots was uni- 
 form last year, it has now become 
 divided into two sorts. From 
 this it follows that to plant the 
 two plots with similar seed for 
 the purpose of comparison would 
 end in failure, for who could tell 
 how much to allow for the differ- 
 ence of soil in the two plots? 
 Mistakes of this sort have been 
 extremely common in agricul- 
 
 Fipf. 11.— To show the previous history of two plots. tural experiments. * 
 
 No. 1 bore peas, while No. 2 bore cabbae-es. To show \ c:»^;i„„ /i«w„-« ^r^,,! J K* j„ 
 
 the unfairness of locating uvo plots, as shown at 1 and .A simiJar error would be made 
 
 2, if these plots are afterwards to be compared with each in case the land of one of thenlots 
 other, unless the object of the comparison is to show the i j i j. -i , , . * , 
 
 relative effects of oBti and cabbage crops on the land. "aQ -last year been Cultivated, 
 
 
 ■%.^^-^i^^ 
 ^•^=^%-li% 
 
Jfjriddlural F.xperiment IFoik. 
 
 M 
 
 while that of the other j lot had not. 
 A supposed case is shown in diagram 
 12, where plot No. 1 bore maize last 
 year, while plot No. 2 was unculti- 
 vated. 
 
 I have many times seen experi- 
 ment plots arranged as shown on this 
 page at Fig. 13. Nos. 1 and 2 are 
 t^vo plots of 
 Ifguminosa? to 
 bo compared 
 vi i t li e a e li 
 other, but the 
 
 V i^y « .s .*. ^ -o «• i» .*. 
 
 .^ t* Mj » ft I* ,1, «. >'/ w 
 
 I, V y, ** « I. U '• •> " 
 
 * 1* O 't •*''<«' i» ■" " 
 
 Showing;' the iu«.\^l!S^hi3tory of two plots, 
 " . roore maize, while Ko. 2 
 
 *r^-i!^;:3»5a:_:^^'f*;,,-;— --—; --_ • ^^ -^^^ 2 is fallow, 
 
 Fi^^-. 13. — Two leguminous plots, numbered 1 and 2, growing alongside sorcrlium. WXllle that next tO 
 
 No. 1 is occupied by a tall 
 graminaceous crop. Hence, the 
 outside plants of the two plots, 
 if I may po call them, are under 
 entirely different conditions so 
 far as the soil is concerned. 
 Again, supposing the left-hand 
 to be east, the tall crop alongside 
 No. 1 keeps the morning sun 
 from that plot (No. 1) much 
 longer than from No. 2. The 
 continuance of these conditions 
 through an entire season is sure 
 to have a decidedly different 
 effect on the two plots. II o vv can 
 anyreasonable person expcctpro- 
 fitablyto compare with each other 
 the yields of these two plots ? 
 
 Old sheep-camps leave almost 
 
 ' no trace by which they can be 
 
 recognised after the ground is 
 
 ploughed. Consequently, plots 
 
 1 
 
 z 
 
 rii 
 
 14.— Two plota which now are of unequal value for 
 experiment purposes, because of the foi'iner presence of a 
 sheep camp on the dotted part. 
 
 for experiment purposes are often found located as shown in Pig. 14!, The 
 
12 
 
 Agricultural Experiment Work. 
 
 
 % 
 
 ■A-vr-y 
 
 stippling represents the location and extent of an old sheep-camp or other 
 stock-camp, all trace of which has, however, heen removed by P ougI»ng. 
 The plots have been so placed that plot No. 1 contains more of the o d 
 camp than plot No. 2. The result from an experiment so arranged would 
 probably be useless, because the extra manure or altered conditions in plot 
 No. 1 would render worthless the comparison of the results obtained on the 
 
 ^ ' The presence of a road or 
 
 path, cither now or in the past, 
 often materially alters the crop 
 on a given piece of laud. The 
 soil of a road, track, or path 
 becomes very compact, and when 
 broken up with a plough works 
 quite differently from ordinary 
 soil. The location of an old road 
 can often be traced on a culti- 
 vated field in dry climates for 
 some years after it is first broken 
 up, merely by this mechanical 
 difference. The soil remains hard 
 and lumpy on the site of the 
 road, though it may work well 
 enough elsewhere. This dif- 
 ference in physical condition 
 between the soil of an old road 
 and the adjacent soil has its 
 effect on the crop. A hypo- 
 thetical case is shown in the 
 adjacent illustration, where a 
 road formerly passed obliquely 
 across the left-hand plot, while it 
 
 crossed only the upper left-hand corner of the other plot. This would give 
 
 rise after ploughing to very unequal conditions in the soil of the two plots ; 
 
 -SO unequal as to render a comparison of the two plots useless. Such cases 
 
 are much more common than would at first be supposed. 
 
 ,1 / ^, 
 ft "/ff -f 
 
 
 Figr- 15. —Two plots of unequal experiment value, because 
 of the foniier presence of a road located as shown. 
 
 Seed. 
 
 While it is true that the variations in the soil, which I have so fully 
 described, are principally responsible for failures in agricultural experiments, 
 there are other causes of failure which need to be noted. Among these 
 minor factors, variation in the seed is of common occurrence. Seeds may 
 vary in age, size, and vitality. I have often cross-questioned a so-called 
 experimenter as to his seed somewhat as follows : — 
 
 Q. " You say you are comparing Smith's Wonder with Brown's Cham- 
 pion?" 
 
 A. "Yes. 
 
 Q. " Where did you get your seed ;" 
 
 A. " I bought the Smith's Wonder from Smith, the seedsman, and my 
 neighbour, Thompson, gave me the Brown's Champion." 
 
 Q. " How old was the Smith's Wonder seed ?" 
 
 A. " I don't know. Last year's, I suppose. Smith said it was good seed." 
 
 Q. " How old was the Brown's Champion seed ?" 
 
Agricultural Uxperiment Work. 13 
 
 A. " "Well , I don't know that either ; but I'll ask Thompson. ' I say, 
 
 Thompson, how old was that seed you gave me?' (Thompson, in 
 
 reply) : ' I have forgotten whether it was last year's or that of the 
 
 year before last.' " 
 Q. " And which variety do you think is the best?" (Successfully concealing 
 
 the sarcasm of the question.) 
 A, " Oh, the Brown's Champion. Look at it !" 
 Q. " Dp you know what sort of land either of the samples of seed was 
 
 raised on?" 
 A. " I don't know vrhere the Smith's "Wonder was raised, but the Brown's 
 
 Champion came from that piece of land down there on the river 
 
 bank." 
 Q. " Do you know how the crops were cultivated from which the seed 
 
 came?" 
 A. " ]N"o ; but Thompson is a good farmer." 
 Q. " Did you grade the seed ?" 
 A. " I don't know what you mean." 
 Q. " I mean to ask whether you removed the small seeds and used seed 
 
 all of one size ?" 
 A. "Oh, no." 
 Q. " Then there might have been more small and shrivelled seeds in one 
 
 samjjle than in the other r" 
 A. " Tes, there might have been ; but I did not examine particularly." 
 Q, " Then, for ought you could say, the Smith's "Wonder might have been 
 , _ grown four years ago on poor sandy land, without cultivation, in a 
 
 location or climate unsuited to it, while the Brown's Champion might 
 
 have been grown either last year or the year before, on a rich river 
 
 bottom, in a favourable season, and with the best of neighbour 
 
 Thompson's cultivation ?" 
 A. " Tes ; that's true." 
 
 Now, if the above conversation were typical of certain rare exceptions I 
 ■would not have introduced it. "Unfortunately, however, if the whole truth 
 were extracted concerning the data of so-called agricultural field experi- 
 ments the world over, there is reason to fear that some such exposure would 
 be the result. 
 
 "Where one variety of seed is being compared with another, great care is 
 necessary in securing and preparing the seed. Again, when the subject of 
 experiment is the efEect of different manures, or different methods of cultiva- 
 tion, the same care is necessary. It is of no use to compare manure A with 
 manure B if the seed used with A is one year old while that used with B is 
 three years old, nor if that used with A contains many more small grains 
 than that used with B, nor yet if the source of the seed used with A is widely 
 different from that used with B, as, for instance, when one is home-grown 
 and the other imported from a different climate. 
 
 Some of these matters are so important, and so generally overlooked, that 
 it will be advantageous to dwell upon them more particularly. 
 
 The grading of seed 13 almost wholly neglected by farmers in most of the 
 great agricultural countries, and this very fact is one that is sometimes 
 pointed out in defence of not grading the seed. I have heard it said, in 
 substance, by a well-known teacher of agriculture— a man whose word is 
 respectfully listened to by thousands of farmers— that there could not be 
 much in using graded seed or farmers would not so generally neglect the 
 matter. Add to this thiit one may find, in almost any farmers' meeting of 
 
11 
 
 uigricuUural Experiment JForlc. 
 
 any size, advocates of small ami pinched seed, and we have, indeed, a strange 
 state of affairs. One might pertinently ask tl)i^ teacher of agriculture how, 
 if a practice is to be its own sufficient defence, any improvement is ever to 
 take place, and iiKjuire of the advocates of pinclied seed, why dame Nature 
 does not provide all her plants with pinched and puny ;^eed, if they are so 
 much belter. 
 
 The truth is that, other things being equal, plump seed is much better than 
 shrivelled sceJ, and that where the results obtained from shrivelled seed are 
 better than those obtained from plump need, the result is due to other factors 
 than the quality of the seed. A fanner some d.iy fows some badly-pinched 
 seed and reaps from it a good crop, and concludes that pinched seed is as 
 good as any, forgetling to note that the conditions have been such that 
 almost anything would have grown well. Ho may e\e:i sow plump seed in 
 
 Rubbish. 
 
 Fig. 18. 
 
 Fie. 21. 
 
 Fig. 22. 
 
 Figures showing the size of tlie gr.iins belonging to each ;:-ia(lL- yielded by the sieves shown in 
 Fi,^Hrcs 2_J lo 28. The proper grade numbers are placed above each illustration. 
 
 the same stason, and get a poorer harvest from it than from the shrivelled 
 seed, and in this case feel highly fortified in his good opinion of pinched 
 seed. The mistake he mnkea is in attributing his result to the one obvious 
 difference of seed, entirely forgetting a score of other things that would 
 account for the difference in result. Anyone who will take a sample of seed 
 and grade it, as shown in the illustration above, and sow the same in drills 
 side hy side, can easily convince himself that the produce of a plant depends 
 
 Fig. 23. 
 
 Figures of six sieves described in the text, .and used for grading v. heat. The sieves are square 
 wooden frames, carrying sieve bottoms niatle of half-rountl bi*ass wire soldered ill jilacc to 
 otie, two, or three cvos^ wires. The iliffcrent grades are know n as 325, ^ W, 275, 250, 22,'), and 
 2fiu, meaning 3-25 nnllinutres, 3 OO millin.eires, 2-7r» niiirmetre>^, 2.'>0 millimetres, 2-25 
 millimetres, and 2*00 inillinieties. These six sie\es i;ive sever: gr.uies of wheat, to which tlie 
 CHI-responding mesh-nnmbers are t;i\ en. The -ruin coming tluough the 20fl mesh is ealletl 
 rubbish. The proper grade nunilji r is iil.ued l>elnw each of the figures, which are about one 
 twenty-fifth full tize. 
 
 from which it springs. The 
 much inferior to those from 
 
 in a marked degree upon the size of the seed 
 ])]auts from grade 2uO, for example, will be 
 grade 325. 
 
 This fact should be taken specially into account in experiment work, and, 
 except when the object of tlie experiment is to ascertain the difterence due 
 
Agricultural Experiment Work. 
 
 15 
 
 to size of the seed, the seed should be carefully graded and brought to as 
 uniform a standard as possible. There are good machines manufactured 
 for grading seed on a large scale for commercial and agricultural purposes, 
 but these are not, as a rule, adapted to accurate experiment purposes, so 
 that the experimenter is, for the most part, driven to the use of hand 
 appliances, and these must, very often, be of his own devising. In general, 
 it may be said that only spherical seeds can be successfully graded with 
 square or round meshes. Elongated seeds require long meshes with parallel 
 sides. Ifc is needless to add that for experiment work the meshes should 
 be quite uniform and accurate as to size. 
 
 Of course, the grading by means of sieves yields seed of uniform size only. 
 Of all the seeds of one size, however, some will be plump, some shrivelled, 
 some bad, some good, and it is, therefore, necessary, in many cases, not only 
 to grade, but also to hand-pick and otherwise sort out the seed for experi- 
 ment purposes. Here, again, is a field for the inventive talent of the 
 experimenter. 
 
 The vitality of seed diminishes with age, hence it is necessary in compara- 
 tive experiments, except such as are devoted to ascertaining the rate of 
 decrease in vitality due to age, to make sure that the seed used in the 
 different plots or rows is of the same age. " 
 
 The vitality of seed is also markedly affected bv the method of storing. 
 Seed kept in confined air-tight space often diminishes more rapidly in vigour 
 than that which has dry air 
 circulating through it. Por 
 example, seed-wheat in stop- 
 pered bottles loses vitality 
 faster than in bags. For 
 these reasons it is necessary 
 for the experimenter in agri- 
 culture to see that his seed 
 used in comparative experi- 
 ment work has all been stored 
 in the same manner. 
 
 The source of the seed used 
 iu agricultural experiments 
 has to be taken into careful 
 consideration. The nature 
 of the land, cultivation, season 
 and climate used in producing 
 the seed all require to be 
 attended to, and, if necessary, 
 allowed for. The accompany- 
 ing figures will serve to make 
 this clear. Suppose 1 and 2 
 to be two ordinary plots or 
 rows to be compared with 
 each other. It is evident that 
 the comparison of the crops 
 will be rendered less useful , ., i, . ^ t fn^ o ;« 
 
 if the seed used on No. 1 is derived from A, while that used for No. 2 is 
 derived from B. I have designed the illustration to show how varying 
 conditions would give rise to differences in the seed derived from plots 
 A and B respectively. A is supposed to have been formerly under tobacco 
 while B was under turnips ; A was, therefore, cultivated quite differently 
 
 Fig. 29. 
 
IG 
 
 Agricultural Experiment Work. 
 
 from B, and would liave had a different demand made on the constituents of 
 its soil ; A is supposed to be one mile from B and two thousand feet lower, 
 and to have a different sort of soil, subject to different seasons and rainfall. 
 It is evident that any one of these variations in condition would give rise to 
 noticeable differences in the seed, while certain combinations would cause 
 a most marked difference. 
 
 The Row System. 
 
 Having in the preceding pages pointed 
 out how defective the plot system of 
 experimenting is, and how, even under 
 the best of conditions, it is very diflacult 
 to correctly allow for the various disturb- 
 ing factors, let us now consider whether 
 some other system is not better adapted 
 to ordinary open-air conditions. 
 
 For years past I have pursued with 
 success a modification of the plot system, 
 which I call the row system. The whole 
 matter is explained by saying that each 
 plot is reduced to a single row. The row 
 may be as long as desired, or as long as 
 the land allows. The two plots to be 
 compared become reduced to two adjacent 
 rows, and it will be at once seen that 
 this arrangement reduces to a minimum 
 many of the errors due to variations in 
 the soil. The distance between two adja- 
 cent rows, say fifteen inches, is so small 
 compared with the size of the variations 
 in the soil that the comparison is generally 
 a fair one. No matter how much, or for 
 what reason, the soil varies, the effect will 
 be about the same on each row. The 
 adjacent diagram has been prepared to 
 ill ustrate this fact. A and B are two rows, 
 fifteen inches apart, to be compared with 
 each other. The illustration is so pre- 
 pared as to show the supposed previous 
 history of the soil, and it will be seen 
 that the variations in the soil due to 
 that cause will affect the two rows about 
 equally. Pires, floods, insect pests, pn- 
 vious crops, stock camps, trees, decay- 
 ing vegetable matter, old paths, weeds. 
 
 
 rows, A and B, are drawn with a double line 
 for the sake of plainness. 
 
 Fij<. 30. — To show that no matter what the ^ ^ . . 
 
 previoM history of the ground may have g-nramps, will, as in the Case of rectangular 
 
 been, it is fairly safe to compare two °""'"'1'"> " i i p .1 - j. 
 
 adjacent rows with each other, provided they plots, aflect the grOWth Ot the tWO parts 
 
 are located as shown at A and n. Each of the ^£ ^^^ experiment, but with this im- 
 
 jjorlant difference — they will affect them 
 
 about equally. 
 
 It is not claimed for the row system that it is perfect, but that it is a great 
 improvement on the ordinary plot system so largely used. After trying the 
 row system for years, I am prepared to say that it gives a complete denial 
 to the statements of those who claim that no outdoor method of experiment 
 
Agricultural Bxperiment Woi^le. 
 
 17 
 
 on ordinary soil can give scientifically useful results. It is true that in some 
 cases I find that two adjacent rows supposed to be under identical conditions 
 show a wider difference than the two rows that are to be compared with 
 each other, and the experiment thus made abortive, but such cases are 
 much less common than among ordinary plots. It is just as necessary to 
 have check experiments, that is, check rows, as ever, but in experiments con- 
 ducted by means of the row system, these checks do not, as a rule, have 
 the disheartening effect of showing the experiment to be useless. 
 
 Another great advantage of the single row system is the ease and accuracy 
 with which replanting can be done. If the first planting is done accurately, 
 any misses can be instantly detected and easily 
 remedied. In many experiments the failure of 
 the seed to germinate evenly makes the final 
 result questionable, and any method that will 
 permit of a ready method of eliminating this 
 uncertain factor, in some instances at least, 
 has everything in its favour. Suppose in the 
 rows shown in Fig. 31 failures to germinate 
 occur at A, A, A. These failures can be seen 
 very early on account of the regularity of the 
 planting, and new seed inserted. From expe- 
 rience, I can say that this is often the saving of 
 the experiment, which would otherwise be lost. 
 It should go without saying that if the result 
 sought by the experiment depends in any way 
 on the germinating power of the seed, such 
 replanting should not be done. Where, for 
 instance, one variety is being tested against 
 another, it is evident that the germinating 
 power is one of the factors that control the 
 result. In case, however, one manure is being 
 tried against another, the germinating power 
 of the seed may have nothing to do with the 
 result, so that any failure to grow, due to the 
 quality of the seed, could be eliminated by 
 replanting. 
 
 Closely associated with the foregoing is the power given the experimenter 
 by counting the plants in liis crop to make allowances for any defects due to 
 the attacks of grubs, birds, &c. The counting of even large numbers of 
 plants is rendered both easy and quick by the fact that it is only necessary 
 to count the misses, these, of course, in well conducted experiments being 
 usually few. 
 
 It is easy to see why two rows at even so small a distance from each other 
 as one foot are often afiected unequally by the unequal condition of the soil. 
 It is only necessary that the various causes which have been mentioned as 
 interfering with ordinary plot experiments should act on a small scale 
 to cause the same difiSculties that have been already fully explained m 
 the numerous cases cited in the previous pages. A dead beetle, a piece 
 of charcoal, a few sticlvs, a wisp of straw, at the base of a plant are sufficient 
 to cause it to grow differently from its neighbours which ^^e situated 
 differently. Nevertheless, all our deductions from what we know about the 
 nature of soil lead us to the conclusion that variations in the soil will be 
 less disastrous to row experiments than to plot experiments, and what to 
 
 116 100— 9G B 
 
 Fig. 31 
 
18 
 
 Agricultural JExperiment Work. 
 
 my mind is far more to the point, many thousand trials conducted by me 
 have sustained these deductions. 
 
 The length of the rows may be so great as to bring the area up to any 
 desired size. Where the rows are two links apart a row a mile long would 
 occupy more than an acre. Many of my experiment plantings are half this 
 size. 
 
 It is, of course, not necessary to have the crop all in a single row. Thus 
 in Fig. 32, the rows lettered A are to be compared with those lettered B. The 
 
 Fig. 32. — Showing a succession of short rows of two alternating sorts arranged for comparison 
 
 sum of all the A's, and of all the B's, even though each one of them be 
 short, may be made to rise to any area desired by eimply increasing their 
 numbers. 
 
 The row system presents no particular difficulties, although it has pecu- 
 liarities of its own which I shall now endeavour to explain. 
 
 The usual aim of a row experiment is to give all the plants the same 
 conditions with one exception, and to note the variation in yield due to this 
 one cause. The method adopted will vary according to the nature of the 
 experiment, but the sowing, in all experiments not devoted to ascertaining 
 the effect of different methods of cultivation, will proceed somewhat as 
 follows : — 
 
 1. The land is ploughed deep enough at a time (dependent on the 
 
 season) to suit the crop to be put in. 
 
 2. It is then harrowed, or rolled and harrowed, until reduced to the 
 
 right tilth, and finally rolled. 
 
 3. The rows are then marked off and opened, if the sowing is to be in 
 
 drills. 
 
 4. The seed is then sowed and covered. 
 
 Xet us consider each of these operations in detail, — 
 
 1. The aim in ploughing should be to thoroughly turn over the laud at the 
 proper time, and in this connection much depends on the season in a 
 country with a small and uncertain rainfall. The best rule in such a case 
 is to have a large team in good order and await an opportunity, and when 
 the opportunity occurs make the utmost of it. To plough when the ground 
 is so dry as to turn up in large clods is in many cases inadvisable. The depth 
 to which the land should be ploughed depends on the depth of the soil, the 
 crop to be put in, and the nature of the season. 
 
 Eight inches is a good average depth. If this depth cannot be reached 
 every year, there is all the more reason to plough deep when a favourable 
 year does occur. 
 
Agricultural Experiment Work. 
 
 19 
 
 Harrowing twice, or tbree times if necessary, brings the top of the soil 
 
 into a fine and uniform condition, and harrowing across the furrows accom- 
 
 ^plishes this object quicker 
 
 ' \l \V 
 
 Fig S3.— Marking the location of 
 dulls by means of a survej'or's 
 chain. 
 
 than harrowing with the fur 
 row. The soil should not 
 be allowed to remain lumpy, 
 for the obvious 
 reason that such 
 soil is less likely 
 to be uniform in 
 quality. "When the 
 soil is lumpy a roller 
 should follow the 
 harrow, and if the 
 lumps are very hard 
 the roller may be 
 heavy and be fol- 
 lowed up with a 
 
 spading-harrovv. Both ploughing and har- 
 rowing should be done in a uniform manner, 
 so as to secure the greatest regularity in 
 the mechanical condition of the soil. If 
 one part is ploughed or harrowed differently 
 from any other part, the plants will be grow- 
 ing under different conditions in the two 
 
 parts, and such plants cannot, therefore, be compared with each other to 
 advantage in other respects. If, for instance, the yield is different in the 
 two parts it will be impossible to say how much of the variation is due to 
 difference in the manner of ploughing and harrowing. To facilitate laying 
 down the experiment rows with quickness and precision, it is best to follow 
 the last harrowing with a light roller. 
 
 Fig. 31.— Opening drills with a wheel hoe. 
 
 The position of the future rows is easily marked out by laji ig down a 
 chain, and marking from the links on the rolled soil with a stick in the same 
 
20 
 
 Agricultural Experiment Work. 
 
 manner as in laying out vegetable and flower beds. If the rows are one, two, 
 or other exact number of links apart, the future results will readily reduce 
 
 Fi;;. 35.- Closintr drills with a wheel hoc. 
 
 to yield per acre, a matter of con.iide.ablc consequence when the number of 
 of experiments is large. The drills 
 should be straiglit, both because if 
 they are not straiglit, even, square, 
 and parallel, some rows will have 
 better chance than other.--, 
 and because the subtcqiieut 
 labour of cultivating and 
 keeping down the weeJs is 
 much less if the rows are not 
 crooked. The drills may be 
 conveniently opened with a 
 Planet Jr. DoubJe-wheel Hoe 
 as shown in the adjacent 
 sketch. AVlth an orchard- 
 cultivator and pair of horses 
 several drills may be opened 
 at once, but it is difficult to 
 get the drills as straiglit a.s 
 when they are made one at a 
 time by hand. 
 
 Much unnecessary truvel- 
 ling may be saved bv having 
 two field lines, or wires, by 
 which to open the drills. 
 The lines are laid down at 
 A A and C C as shown in 
 Kg. 37— the drill B B being 
 missed. Open drill A A, be- 
 g'nning at the to;) and then 
 
 Fit,'. 3C.-0pening drills with 
 double wheel hoe. 
 
AgriciiUnral Experiment JVorh. 
 
 21 
 
 move tte line by nliieh A A was opened to the position A B Pig. 38. 
 Open C, beginning at tVie bottom, and move the line .by -which C was 
 
 A B C D 
 
 k B C D 
 
 Fig. 37. 
 
 A B C D- fi. B C O E 
 
 if ' 
 
 fi, B C D.A B C D E 
 
 HBCOE\fKBCDEFQ 
 
 I I I I 
 
 'a b c d e f c 
 
 I I > 
 A B C D £ F G 
 
 Fig. 38. Fig. 39. Fig. 40. Fig. 41. 
 
 To sliow the metliod of using two field lines in opening drills. 
 
 opened to C E Kg. 39, and the first line to B B Fig. 40. Then open B B 
 Kg. 41, and move the line by which it is opened to B D Kg. 42, and the 
 other line to E E, and open E E, and so continue. 
 
 Opening the drills across the furrows gives 
 more uniform results than opening them with 
 the furrows. The harrow 
 fails to obliterate the fur- 
 rows. If, therefore, the 
 
 i 
 
 drills are made 
 
 in the same 
 
 direction as the 
 
 furrows, it may 
 
 happen that 
 
 one drill lies 
 
 on the top of 
 
 what was a fin 
 
 row while its 
 
 neighbour licb 
 
 between the 
 
 furrows, and the two resulting rows of plants would be 
 
 likely to differ on that account. TVe must always keep 
 
 in mind that we wish to give all the rows the 
 
 same chance. 
 
 When the drills are opened the seed should be 
 dropped by hand. Iso machine has been made, 
 so far, that will drop seed with sufficient accuracy 
 for experiment purposes. The only reliable way ii to sow by hand, and to 
 entrust the sowing to reliable persons only. The sketch below gives a cood 
 
 a surveyor's chaia marked to links. 
 
 ri^. 44.— Punch for putting in 
 seed. A, shoulder governing 
 the depth to which the seed ia 
 pushed into the ground ; b, a 
 brass ferrule ; c, seed. 
 
22 jigncuUural Experiment Work. 
 
 idea of one of the best methods of dropping seed accurately alongside a 
 chain or tape divided into links or otherwise graduated. When sown the 
 seed should be covered uniformly. It is better that the same man should 
 cover all the rows that are to be compared with each other. 
 
 There is another method of sowing for experiment purposes, vphich, thougli 
 not quite so expeditious as that just described, nor yet so similar to the 
 ordinary method of sowing, is nevertheless superior in point of accuracy. 
 No drills are opened. A chain is laid down on the rolled surface of the soil 
 
 ::c::: 
 
 
 r* 1 1 1 
 
 Fig. 45. — One hundred square experiment plots arranged in a square. 
 
 and the seeds, laid in position alongside- the chain, are punched into the 
 ground by means of a hollow-ended wooden punch having the form shown 
 in the adjacent figure. The shoulder on the punch enables the sower to 
 place each seed at the same depth, and if necessary the same side up. 
 
 rinally, there is another advantage in the row system in the e'ffective way 
 it can be used to educate those who inspect the experiments. Each row is an 
 experiment by itself, and can be readily compared with its neighbours. 
 Thus, in the accompanying engraving of a series of manure experiments on 
 
Agricultural Experiment Work. 23 
 
 the row system, the results are rendered far more striking than they 
 would be if conducted on the ordinary plot system. By the time one had 
 inspected three or four plots the strain on the memory would be so great 
 that the mental comparison would be hazy. It is usually the case in plot 
 experiments that only one plot can be well seen at a time. It is only 
 necessary to imagine an observer while examining plot No. 10 trj'ing to 
 recollect how plot No. 5 looked to understand at once how much superior 
 the row system is, when, in all probability, rows 10 and 5 could be seen 
 almost at one and the same time. 
 
 10 20 30 40 50 60 70 80 00 100 
 
 Fjff. 46.— One hundred experiment plots arrang^ed in rows. 
 
 Added to this convenience in comparison is the small labour necessary in 
 inspecting rows in comparison with that of inspecting plots. Let figures 45 
 and -16 represent one hundred plots, in one case arranged in squares and in 
 the other in rows. It will be seen at once that the observer has to travel 
 six times as far to inspect the plots as to inspect the rows. In case of the 
 rows he has only to pass along by the ends of the rows. In the other case 
 the shortest road to a complete inspection, marked by the line in the 
 diagram, is about six times as great. 
 
24i 
 
 Agricultural Experiment Work. 
 
 The Pot System. 
 
 In 1880, Professor "Wagner, Director of the Darmstadt Experiment 
 Station, initiated a system of pot experimentation on an extensive scale, 
 being driven thereto by his previous failure to secure scientifically accurate 
 results in his field experiments with manures. This system has been care- 
 fully and systematically developed at Darmstadt by Professor Wagner, and 
 has become widely known as the Wagner System of Pot-culture. 
 
 The pot-culture system has been developed on the assumption that no 
 system of field culture can be made to yield scientifically accurate results in 
 any reasonable length of time. Though we may not accept such an assump- 
 tion it will nevertheless be profitable for us to consider carefully the 
 methods and results of the pot-system. 
 
 The sketches below show how this system is carried out. The two glass- 
 houses are for the protection of those particular pots which are carried in and 
 out on trucks run on permanent steel tracks. The houses are utilised at night 
 and in inclement weather to protect these movable pots. Platforms along- 
 side the tracks permit of the pots being removed from the trucks for experi- 
 ment, or for the sake of separation, so as to give each pot more air and sun- 
 light. 
 
 The cylindrical water-tight pots, over 1,000 in number, and varying in size 
 from three feet {icross, by four feet deep, to two feet across, by two feet 
 deep, are made of zinc, and are supplied on the outside with a tube by 
 which water can be poured into the bottom of the pot to be absorbed by 
 
 4. >.f-*^w^* , -f'-f^^T' ti 
 
 Fig. 47.— Sketch from a photograph of the Daimstadt Experiment Station, ahowingpots on trucks. 
 
 the soil within. The soil is, of course, carefully made and prepared, 
 according to the requirements of the experiment. The pots inserted in the 
 ground have no bottom. The pots not inserted in the ground have each 
 three legs, by which they are raised two and one-half inches. The trucks 
 
Agricultural Experiment Work. 25 
 
 ^oi*?rff??'*n''? now well known on large cane and other plantations 
 coSXuVJ^^OO t"m-i?"* ^'' ^^^^•'' ^^' *'^ ^'^'^^ ^- *° corrLpondTd 
 abJut'£4?o'°Thf« ^P\^^t^f/« large as the Darmstadt plant would be 
 tw»^f ^t 1 T^ '^°" "^ '°''^^'^^ °°« glass-house, twenty-five bv fifty feet 
 wK.^*;?'' *^^«-«:;g^<« f f "^ile of t'-aclr, and several hundred poTs ' 
 Where the exact nature of the soil is a telling feature of the experiment 
 and this IS usualljtbe ease at Darmstadt, the soil is- prepared cSllTl' 
 that Its exact physical and chemical conditions are known^ Soils arfnraW? 
 cally exhausted of their potash, phosphoric acid and sn fnr+l-, L Ci- 
 thern in cement pits, a/d grov;iPng m th"m cont nuou ly c ps'^'S^^^ 
 large amounts of potash or phosphoric acid, until the Jron wm no lon^ef 
 grow This requires several years. The experimenter thus Kins within 
 nessT, dlr^'"*^ '"?■ '"'^ tl^^^^ghout the experiment the sam^rthoroih 
 ness and accuracy can be preserved, because the various conditions are «lll 
 rTteW reSied '^3"''P'^ of water air, and sunlight c^nte and islccu! 
 exclude'^ Th ^ thp J^'^ # -^T ^^^ ''^^'^'^' i°^<^'=t«' ^"d birds are 
 AW wS all JhVf ''•*'"' ^'°"''' '^ '""^' "* laboratory experiments 
 Along with all the foregoing advantages the Warner system has acouired 
 many of the disadvantages of the system of water culture^ the cWef of^which 
 
 r/\^~i^ 
 
 
 iHnfi^^ip5^u^ 
 
 Fij. 48.— Sketch-from a photograph of the Darmstadt Exporiment Station, showin" pots 
 insei-ted in the ground. * 
 
 IS the unlikeness of the conditions to those of common agriculture, and the 
 conseqiient diflSculty in reasoning from the results so as to make them useful 
 m guiding ordinary field operations. These objections are least troublesome 
 when the experiments are closely connected with root absorption. 
 
 Tbe wider introduction of pot experiments is inevitable, and the advan- 
 tages of the wider adoption of the system in this Department will be two- 
 fold. 
 
26 
 
 Agricultural Uxperiment Work. 
 
 First, they will be of high educational value ; they are cheap and easily 
 carried out in many instances ; they occupy little space, and they are very 
 demonstrative. These qualities are such as to recommend them highly 
 wherever students are to be shown conclusive examples of the laws of plant 
 nutrition and of plant diseases. A few dozen pot experiments at each of the 
 Department's educational institutions would be a very desirable addition. 
 
 Second, the conditions in this country in respect to rainfall are so peculiar it 
 seems to me probable that a repetition of many of the European experiments, 
 with special attention to Australian conditions, would be profitable. Add 
 to this that the field open to pot experiments has hardly as yet begun to be 
 explored, and we have a strong argument in favour of our undertaking 
 experiments of this sort, especially when we remember that we have officers 
 abundantly able to carry them on. 
 
 Educational Experiments. 
 
 There are many inexpensive educational experiments that should be carried 
 out at all agricultural colleges. These are merely annual repetitions of old 
 experiments, and are designed to give ocular evidence of the principles of 
 agriculture. The effect of manures on various crops and in various soils can 
 be made the subject of very demonstrative experiments, as can also the 
 
 m i 
 
 ■^w 
 
 -^^ 
 
 Fig. 49.— Shomng experiment rows, Wagga Wagga Experiment Farm; two manured, followed by 
 three not manured, then two more manured differently, &c., &c. Part at the left has been 
 harvested. 
 
 application of manures to late-sown crops. The known relations of certain 
 crops to certain classes of manure, as of potatoes to potash manures, would 
 also be a suitable experiment for educational purposes. The relation of 
 certain soils to certain crops could be shown in a series of pot experiments. 
 
 Entomological experiments showing the value of friendly insects, of iu- 
 secticides, and of the value of sound sets or tubers, as against insect- 
 infested sets, would be of value. 
 
 Eows of accurately named weeds would be an instructive lesson from the 
 botanist. 
 
 The relative value of large and small, old and fresh, plump and shrivelled 
 seed can be satisfactorily shown by a series of rows. The relative effect of 
 early and late sowing can be illustrated in the same way. The wheat- 
 nomenclature plot, at Wagga Wagga, is a good example of what can be done 
 by a careful study of annual crops. It would be a fine work to attack oats, 
 barley, potatoes, and other annual crops on similar lines, and have nomen- 
 clature plots of all. Such a series of plots would have an exceedingly high 
 educational value. 
 
 The relative value of different methods of planting and cultivating could 
 be shown in a series of rows and plots. 
 
Agricultural Experiment Work. 27 
 
 Examples of. pruning side by side, good and bad, would teach a valuable 
 lesson. 
 
 Examples of disease, and the benefit of treatment by spraying and other 
 methods could also be made instructive. 
 
 All these experiments, and hundreds of others, very few of which have 
 been actually carried out, show how much we have yet to do before we can 
 claim to be up with the times. 
 
 The Cost of Experiment Work. 
 
 New and valuable knowledge is rarely obtained except by the expendi- 
 ture of much thought and money, and this applies to agriculture as well 
 as to everything else. Tet the investment is a good one. The most 
 rapid way of acquiring new knowledge is by experiment. Tet it is 
 necessarily expensive work. It means the thoughtful work of minds that 
 have to be paid for. It must not be thought that agricultural experi- 
 ments can be "rough" and inexpensive. That has been one of the chief 
 of agricultural fallacies in the past, and even now there are very few men 
 in the world who have found out how much of most careful thought and 
 patience it requires to carry out exact experiments in agriculture. To a 
 large extent you can control the factors in the experiments in most 
 sciences, but when it comes to outdoor experiments in agriculture you find 
 some matters are quite beyond control, and the necessity of allowing for 
 these uncontrollable variations makes the experiments oftimes most difficult. 
 
 I think the cost of experiment work should be kept separate from the 
 ordinary accounts of the farm. The accounts of public experiment farms 
 should be kept as carefully as the accounts of any commercial house, and 
 should show exactly at what cost the results have been obtained. It seems to 
 me to be advisable, in order to show the nature of the work, that the accounts 
 should be systematically audited so as to show cost of the experiment work, 
 while the accounts of crops produced at a profit should be shown by them- 
 selves. The cost of the educational work should also be kept separate. These 
 accounts should be published annually, including cost of implements and all 
 working expenses. Every article should be put in, and the depreciation from 
 year to year calculated, so that anyone may see at a glance how the farms 
 are worked, and with what financial results. 
 
 If it is asked will all this careful experiment work pay, the reply can only 
 be that, unless careful work is done, it would be better not to do any- 
 thing. What is an experiment farm for ? It is to show the people who 
 have been growing certain crops in a certain way how much more profit 
 they could make by adopting improved methods which have been discovered 
 by means of these careful experiments. It is to show how the cultivation of 
 crops that are new— like those described in the series of articles by Mr. 
 Turner some years ago— can best be undertaken, and the probable results as 
 applied to different districts. The carrying out of experiments of this 
 kind need not, however, involve an extravagant outlay. Take, tor instance, 
 the experiments that have been carried out at "Wagga Wa,gga, to show the 
 relative merit of pure seed, and seeds of different sizes. It is very 
 evident that the seed-wheat that is used in this Colony is much mixed, 
 and is rarely graded. The Wagga experiments show that the results trom 
 graded seed are greater than from seed that has not been graded and they 
 will, if properly continued, induce people to adopt the practice of grading, 
 and so produce crops that will be more profitable. We want to keep these 
 facts with regard to graded seed clearly before the mmds of the farmers ot 
 
28 Agricultural Experiment Work. 
 
 the Colony, so that gradually tliey will begin to follow the practice they 
 ought to follow. Anyone who goes to Waggajand sees the results of planting 
 the different grades of seed used there, can see how from one grade large, 
 fine plants are produced, and from another, puny ones. But, as I have 
 pointed out before, you can only definitely determine this by having careful, 
 somewhat costly, but not extravagant experiments. I thoroughly believe 
 that there are a great many lines of experiment that it will pay well to carry 
 out in this way, and, as I said before, it involves no extravagant outlay. 
 For educational purposes the ordinary old-fashioned plot experiments would 
 not be sufBcient. 
 
 Century Plots. 
 
 I have often thought it would be a very good idea to set aside on some of 
 our experiment farms a series of century plots — that is, plots that are to receive 
 a uniform treatment for one hundred years. The results of carefully-planned 
 experiments, carried out systematically for such a length of time, would have 
 great value. It would not, of course, be necessary to wait a century before 
 the results were available. In a few years even, the value of the experiments 
 would be evident. The idea is to carry out experiments on a definite plan 
 for a long series of years, and the object of doing this, according to the best 
 agricultural authorities, is to secure reliable results, such as are not obtainable 
 except through experiments carried on for a long series of years. When I 
 undertake an experiment I take every precaution to work out all the little 
 details, for instance, sometimes even taking care to put all the seeds in the 
 same side up ; yet when I do get my result, it only proves that it was so 
 in that season. There will never be another season precisely like that one. 
 The experiment must be repeated through a series of years to get a reliable 
 average. 
 
 The celebrated experimeuts of Lawes and Gilbert have been conducted in 
 this uniform and systematic manner, some of them for fifty years, and the 
 results are admittedly of high value. 
 
 The lines upon which such century plots should be conducted are clearly 
 suggested by the principles set forth in the previous pages of this article, 
 and by the experiments already conducted in England. It is very probable 
 that a repetition of some of the English experiments under Australian con- 
 ditions would be advisable. The continuous growth of one crop on the same 
 pieco_ of land for a long series of years, the continued u^e of a particular 
 rotation of crops on the same piece of land for a long series of years, and 
 many other similar experiments are suggested as worthy of trial. 
 
 We of to-day might start these experiments, and make careful pro^^sion 
 for their maintenance, and long after we have passed away the experiments 
 would continue, and doubtless earn the gratitude of our successors. 
 
 These experiments could be so planned as to nearly or quite pay their 
 way. They should be plots of large size, and they should be well cultivated 
 in the ordinary way. It would be necessary to keep a complete and minute 
 history of each such plot from year to year, and this might be published (say) 
 once in five years. 
 
 Model Farm. 
 
 There is another idea I should like to see carried out. We have a number 
 of experts in this department, and I have often heard people say that it 
 would be a good plan to let them go on a farm, and run it themselves at a 
 profit, implying that they would not be successful. Why should we not set 
 
Agricultural Experiment Work. 29 
 
 apart a farm and have it under the care of an ordinary farmer, telling him 
 precisely what to do ? The farmer should select, clear, build, break up, 
 plant, cultivate, reap, and market, according to the advice of the depart- 
 ment's experts. Let us see what the result would be. 
 
 I think it would be an easy matter to get a suitable man to run such a 
 place. It would be a matter of policy where to place him, but the idea of 
 having a man to act under the advice of the experts strikes me as a good 
 one. It would be necessary to keep the exact history of the place. I can 
 conceive of it being a most successful thing. Certain criticisms that I have 
 heard have led me to propose this answer to them. 
 
 Sydney ; William Applegate Gxillick, Government Printer. — 1896. 
 
BY THE SAME AUTHOR. 
 
 1. list of Plants found Growing Wild within 30 miles 
 
 of Amherst. 
 
 51 pages— in Atlas of Hampsliire County, Mass., 1887. 
 
 2. list Of Plants. 
 
 Separate edition, S. E. Bridgeman & Co., North- 
 hampton, Mass., 1887. 
 
 3. Beitrage zur Anatomie and Ontogonie der KTema- 
 
 toden. 
 
 36 pages, with 32 original illnstrations on -3 litho- 
 graphic plates Jenaisehen Zeitschrift fiir Natur- 
 wissensohaft, XXIII Bd., N.F. XVI, 1888. 
 
 4. Inangnral-Sissertation zur Drlangen der Sok- 
 
 torwurde. 
 
 Gnstav f^ischer, Jena, 1888. Same as the above, 
 6. The Differentiator. 
 
 Sydney, 1889. Beprint in ^liiterioaii Naturalist, Phila- 
 delphia, 1889. 
 
 6. Keport on Occupation of the Tahle of the British 
 
 Association at the Naples Zoological Station. 
 
 Heport of the British Assbciation, London, 1889^ 
 
 7. Kene parasitische B'ematoden. 
 
 3 pages, with 9 original illustrations on one plate, 
 Archir fflr Naturgeschichte, Berlin, 1B89, 
 
 8. Two new Instruments for Biologists. 
 
 11 pages, with 5 original figures on one lithographic 
 plate. I*rocee1iiigsof the Linnsean Society of Kew 
 I South Wales, Sydney, 1890. 
 
 ,9. Oxjruris larrai; hatched in the human stomach 
 under normal conditions. 
 18 pages, with 5 original figures on one lithographic 
 plate. ^Pi^oceedinga of the Linneeau Society of 
 New South "Wales, Sydney, 1890. 
 
 10. A Kematode Formula. 
 
 5 pages, with description of Oncholaifmus inAeJi,' 
 AgricvXturkl Gazette, 1890, 
 
 11. Eeprint of the ahove. 
 
 Department of Agriculture, Sydney, 1890. 
 
 12. Coutrihutions to an Economic Knowledge of the 
 
 Australian Busts (Uredinese). 
 
 ' In course of publication. So far issued, 154 pages, 
 with 303 original illustrations in the text, and 1 
 plate. Agricultural Gazette, Sydney, 1890-94. 
 
 13. Arabian Kematodes. 
 
 20 page^. Proceedings of the Linnsean Society of 
 New South Wales, Sydney, 1890. 
 
 14. Tylenchns and Soot-gaU. 
 
 30 pages, with 21 original illustrations in the text, and 
 1 plate with 9 illustrations after Neal. 
 Agricultural Gazette, Sydney, 1890. 
 
 |1S, Anticoma : A genus of flree-living nematodes, 
 
 10 pages, with 12 original illustrations in the text. 
 Proceedings of the Linnflean Society of New South 
 Wales, Sydney, 1890. 
 
 06, Notes on Diseases of Plants, 
 
 3 pages. Diseases of the yine, pear tree, strawberry 
 ^ plant, and hollyhock. Agricultural Gazette, 1891. 
 
 p. Pathological Notes. 
 
 2 pages, with 4 original illustrations in the ,text. 
 Disease of lucerne. Agricultural Gazette, Sydney; 
 1891. 
 b. Rust in Wheat : Report to Sydney Conference on. 
 ' 7 pages. In "Bust in Wheat" (Report of Conference 
 on). Legislative Assembly, New South Wales, 2u& 
 ' Session, 1891. 
 p. Dialogue concerning the manner in which a 
 poisonous spray does its work in preventing 
 or checking blight. 
 8 pages, with 8 original illustrations in the text, 
 Ajricullural Gazette, Sydney, 1891. 
 
 Most of the above 
 
 New Sa'JTH Wales, either 
 
 20. Eeprint Of the same. 
 
 Department of Agriculture, Sydney, 1891. 
 
 21. Notes on the Diseases of Plants. 
 
 Diseases of apple, strawberry, and wheat. 
 Agricultural Gazette, Sydney, 1891. 
 
 22. Onyx and Dipeltis: New nematode genera, with 
 
 a note on Dorylaimus. 
 
 16 pages, with 17 original illustratidns in the text. 
 ProoeedingB of the Linnsan iSooiety of New South ' 
 Wales, Sydney, 1891. 
 
 23. Maize for the tahle. 
 
 11 pages. ,lgi*fcultwal(3a2«fte, Sydiiey, 1891. 
 
 24. Notes on the Diseases of Plants. ' 
 Diseases of the apple and vine. Agricultural Gazette, ' 
 
 Sydney, 1891. 
 
 25. Smut. 
 
 6 pages, with 9' original illustrations , in the text. 
 Agricultural Gazette, Sydney, 1891, 
 
 26. Smuts. 
 
 Separate and enlarged with more illustrations. 
 Department Agriculture, Sydney, 1891. ' 
 
 27. Hair-worm. , 
 
 2 pages. Agricultural Gazette, 1891. 
 
 28. Notes on Diseases of Plants. 
 
 Diseases of horse-radish and apples. Agri^idtural 
 GazeUe, Sydney, 1891. 
 
 29. Insect-larva eating Bust rn Wheat and Flax. 
 
 4 pages, with 4 original illustrations in the text. 
 Conjoint paper. Agricultural Gazette, Sydney, 1891. 
 
 30. The same. 
 
 Eeprint in The Annals and Magazine of Natural History, 
 London, 1891. - 
 
 31. Pathological Notes. 
 
 Diseases of maize and the apple. Agncultural Gazette, 
 Sydney, 1891. 
 
 32. Notes on the Diseases of Plants. 
 
 Diseases of the apple, pear, peach, and flax, 
 Agniultural Gazette, Sydney, 1891. 
 
 33. Strawberry Bunch. 
 
 11 pages, with 4oi;iginal illustrations in the text and 1 
 plate with 9 illustrations, after. Bizema Bos. 
 A disease of the strawberry plant. Agricultural 
 Gazette, Sydney, 1891. ■ , 
 
 34. Beprint of the same. ^ 
 
 Department of Agriculture, Sydney, 1891. 
 
 35. Parasites in the stomach of a cow. 
 
 2 pages, with 2 original illustrations in the text,^ 
 Agricultural Gazette, Sydney, 1891. 
 
 36. Notes on the Diseases of Plants. 
 
 9 pages, .with 8 illustrations (mostly original) in the 
 text, and one plate with 2 illustrations. Diseases 
 of the potato, onion, tobacco. Agricicltaral Gfazetfe, 
 Sydney, 1891. 
 
 37. Reprint of the above. 
 
 Department of Agriculture, Sydney, 1891. 
 
 38. Devastating Eel-worm. 
 
 5 pages, with 7 original illustrations in the text. 
 Agricultural Gazcttcj Sydney, 1891. 
 
 39. Devastating el-worm. 
 
 The abo7e with 2 additional illustrations. 
 Department of Agriculture, Sydney, 1891. , 
 
 40. Plant Diseases, arid How to Prevent them. 
 
 28 pages, with 30 origin&l figures in the text, and 7 
 figures on 4 plates. Diseases of the apple, fear, 
 apricot, grape, rose, stra^^barry, pumpkin, cabbage* 
 and turnip. Agricultural Gazette, Sydney, i892. 
 
 41. Miscellaneous Puhlication.-Ni. 12, 
 
 Department of Agriculture, Sydney, 1393, is a rsprint 
 of the aboife. ■ ' 
 
 Department or Asricultuhe, Sydney. 
 
 Publications can be had from the 
 
 Free on Application, oa at Prises frd.vi Sixp;nc:e Upjuaros. 
 
BY THE SAME AUTHOR. 
 
 42. Notes on the Diseases of Plants, 
 
 2 piij^es. Biaeaao (tf the ornuge. Afiricnlhiyal Oazette, 
 eydni'j-, 1892. 
 
 43. Plant Diseases and How to Prevent them, 
 
 14 pages, with 25 ori^nnal illustrations iu the text, anil 
 2 original illustrations on one plate. The diseason 
 " Take-all ," ami " Dry Blight." Agricultural Oaidlt, 
 Sydney, !I<!I2. 
 
 44. Kepriiit of thesame. 
 
 Department of Agriculture, Sydney, 1893. 
 
 45. Cold storage of Apples, from the Vegetable 
 
 Pathologist's point of view. 
 
 5 pages. A^-icuUvml QazeUe, Sydney, 1892. 
 
 46. Reprint of the above. 
 
 Department of Agriculture, Sydney, 1892. 
 
 47. Eust in Wheat : Keport to Adelaide Conference on. 
 
 7 pages, with 21 original illustrations in the text. 
 C. B. Bristow, Adelaide, 1892. 
 
 48. Plant Diseases, and How to Prevent them. 
 
 4 pages, and 5 illustrations, mostly original, in the 
 text. Diseases of the vine-root, bean, and apple- 
 tree, .l./r.ciiltural Gazette, Sydney, 1892. 
 
 49. Reprint of the above. 
 
 Department of Agriculture, Sydney, 1892. 
 
 50. Gumming of Sugar-cane ; Preliminary Report. 
 Annual Report of the Department of Mines and 
 
 Agriculture, Sydney, 1893. 
 
 51. Bulletin of the Royal Gardens, Kew, 1894. 
 
 Contains a reprint of the ahore. 
 
 52. Revue Agricole et Journal de la Chambre 
 
 d'Agriculture, Kaurice, 1894. 
 Also contains a reprint of the above. 
 
 3. 'Nematodes, mostly Australian and Fijian. 
 
 59 pages, with 24 original illustrations in the text,, 
 and 125 original figures on seven lithographic plates, 
 Macleay Memorial Volume, Linnsean Society of 
 Kew South Wales, Sydney, 1893. 
 
 64. Host and Habitat Index of the Australian Fungi, 
 44 pa,ges. Department of Agriculture, ^dney, 1893. 
 £6. New Australian Fungi. 
 
 12 original illustrations in the text— Cyflt^us dimor- 
 phus, Tolyposporiitm anthiftirtce, ^Agricultural Gazette. 
 Sydney, 1893. 
 
 56. Reprint of the above, 
 
 Department of Agriculture, Sydney, 1893. 
 
 57. Tricdma and other hew Nematode Genera. 
 
 33 pages, with 50 original illustrations in the text. 
 Linnroan Society of New South Wales, Sydney, 1893. 
 
 58. Plant Diseases and their Remedies ; Diseases of 
 
 the Sngar-cane. 
 56 pages, with over 70 original illustrations in the 
 text. AgriculturaliGazette, Sydney, 1893. 
 
 59. Reprint of the above. 
 
 Department of Agriculture, Sydney, 1893. 
 Eevue Agricole, Maurice, 1894. 
 
 60. Diseases of the Sugar-cane. 
 
 Special oition for the Queensland Department of 
 AgHoulture. Department of Agriculture, Brisbane, 
 1893. 
 
 61. Nematodes, mostly Australian and Fijian. 
 
 59 pages, with 24 original illustrations in' the text, 
 and 115 original figures on seven lithographic 
 plate's, as rej-rinted by Agricultural Department, 
 Sydney, 1893. 
 
 62. A new Australian Fungus. 
 
 Single I age, with 2 original illustrations. FezUa 
 lyovfiia: Agncvlturnl Ijimltr, 1894. 
 
 63. Reprint of the same. 
 
 Department of Agriculture, Sydney, 1894. 
 
 64. Notes on the Diseases of Plants. 
 
 11 pages, with 14 nrivrinal illuKtrations iu the text. 
 Disease! of the ln'uii, tomato, peach, and munin> 
 ^lai-;ci(»r(raU,'«,Ti/(i;, Sydney, 1894. 
 66. Reprint of the same. 
 
 Department of Agriculture, Sydney, 1894. 
 
 66. Improving Wheats by Selection. 
 
 18 pages, 5 illustrations. Department of Agricultore. 
 Bydner, 1H!)4. 
 
 67. The Cause of Gumming in Sugar-cane. 
 
 7 I'ages, with 2 illustrations iu the text. Agricultural' 
 Gazette, Sydney, 1895. 
 
 68. Reprint of the same. 
 
 Department of Agriculture^ Sydney, 1995. 
 
 69. Notes on the Form and Size of the Grains in- 
 
 Different Varieties of Wheat. 
 
 7 p.i.ges. Afrricultural CazeUe, Sydney, 1895. 
 
 70. Reprint of the same. 
 
 Department of Agriculture, Sydney, 1895. 
 
 71. The Hot-air Treatment of Bunt or Stinking 
 
 Smut. 
 
 2 pages. Agricultural Qazettt, Sydney, 1896. 
 
 72. Diseases of Plants and their Remedies. ^ 
 
 10 pages, with 13 illustrations iu the text. Diseose^^ 
 
 of the s^iuash, lemon, apple, cott'ee-plant. Ag)'icul^^- 
 turaX Gazette, Sydney, 1896. 
 
 73. Reprint of the same. 
 
 Department of Agriculture, Sydiiey, 1896. 
 
 74. Notes on the Threshing of Wheat. 
 
 5 pages, no illustrations. Agricultural Gazette, Sydney. 
 1896. ' 
 
 75. Reprint of the same. 
 
 Department of Agriculture, Sydney, 1896* 
 
 76. The Hardness of the .Grain in the principal 
 
 varieties of Wheat. 
 
 22 pages, with 28 illustrations- .<4gi-/cul(itraf Gazette, 
 Sydney, 1806. 
 
 77. Reprint of the same. 
 
 Department of Agriculture, Sydney, 1896. 
 
 78. The relative Hardness of Australian and American 
 
 Fife Wheats. 
 
 8 pages, and 28 original illustrations in the text. A^- 
 crutural Gazette, Sydney, 1896. 
 
 79. Notes on the colour of the Grain iu different 
 
 Varieties of Wheat. 
 4 pages, no illiistrations. Agrictiltural Gazette, Sydney, 
 1896. 
 
 80. Reprint of the same. 
 
 Department of Agriculture, Sydney, 1896. 
 
 81. The Common Crow. 
 
 14 pages, with 1 original illustration in the tejw 
 AffricuUurat Gazette, Sydney, 1896. 
 
 82. Reprint of the same. 
 
 Department of Agriculture, Sydney, 1890. 
 
 83. Experiment Work in Agriculture. 
 
 25 i>ages, with 49 original illustrations in the text 
 Aiiricultural Gazette, Sydney, l.^!>6. 
 
 8^. Reprint of the same. 
 
 Di'liartment pt Agriculture, Sydftey, 1896. 
 
 In preparation. 
 The Prevention and Cure of the Blights of Crops. 
 
 With, several hundred orii:inal illustrations iu tli^ 
 text, and numerous coloured plates. 
 
 Parasites of Sheep and other Domestic Animals (and 
 incidentally of the Australian Land-fauna). 
 
 With several hundred original illustrations. 
 
 Most of the above Publications can be had from the Department of Agriculture, Svdney,,5 
 IMevk South Wales, either Free on Application, oh at Prices from Sixpence Upwards. 
 
 Sydney ; WilUuui Applegate Gullick, Uoverament riintcr. — lyiKJ.