THE FIXATION OF ATMOSPHERIC 
 NITROGEN BY SWEET CLOVER 
 
 A THESIS 
 
 PRESENTED TO THE FACULTY OF THE GRADUATE SCHOOL 
 OF CORNELL UNIVERSITY FOR THE DEGREE OF 
 
 DOCTOR OF PHILOSOPHY 
 
 BY 
 
 LEONARD AMBY MAYNARD 
 
THE FIXATION OF ATMOSPHERIC 
 NITROGEN BY SWEET CLOVER 
 
 A THESIS 
 
 PRESENTED TO THE FACULTY OF THE GRADUATE SCHOOL 
 OF CORNELL UNIVERSITY FOR THE DEGREE OF 
 
 DOCTOR OF PHILOSOPHY 
 
 BY 
 
 LEONARD AMBY MAYNARD 
 
 [Reprinted from Bulletin No. 394, November. 191 7, of Cornell University Agricultural 
 Experiment Station, where it appeared under the title, "The Decomposition of Sweet 
 Clover (Melllotus Alba Desr.) as a Green Manure under Greenhouse Conditions."! 
 

CONTENTS 
 
 PAGE 
 
 Review of literature 122 
 
 Sweet clover culture 122 
 
 Decomposition of green manure 124 
 
 Experimental work 126 
 
 Soil used 126 
 
 The seed 128 
 
 The pots 130 
 
 Season of 1914 130 
 
 Season of 1916 139 
 
 Discussion of results 142 
 
 Production of dry matter and nitrogen 142 
 
 Percentage of fiber 143 
 
 Rate of decay 144 
 
 Summary 1 46 
 
 Bibliography 148 
 
 119 
 
THE DECOMPOvSITION OF SWEET CLOVER (MELILOTUS 
 
 ALBA DESR.) AS A GREEN MANURE UNDER 
 
 GREENHOUSE CONDITIONS ^ 
 
 L. A. Maynard 
 
 The practice of <j;reen-manuring as a method of restoring;' and main- 
 tainin<j; soil fertiHt\' has become more and more (general as evidence 
 regarding its value has accumulated and methods of procedure have been 
 worked out. The choice of a green manure best suited to given conditions 
 has been a subject of much investigation. The literature of the subject 
 abounds with data regarding alfalfa, crimson clo\'er, and various other 
 plants that gather nitrogen from the air. 
 
 A plant of this tyjje to which little attention has been given is sweet 
 clover. This has long been regarded as a weed because of its occurrence 
 in waste places, its general rank growth, and its ability to thrive under 
 conditions unfavorable for the development of other plants. The fact 
 seems to have been o\^erlooked that these very characteristics, coupled 
 with nitrogen-gathering power, are the features desired in a soil reno\^ator. 
 Recently several waiters have called attention to the possibilities of sweet 
 clover for this purpose, but experimental data are lacking. The present 
 investigation is a study of the ability of the plant to gather nitrogen, 
 and the rate with w^iich this nitrogen becomes a\-ailable when the plant 
 material is incorporated with the soil. 
 
 Sweet clover obtains its name from the peculiar sweetish fragrance 
 of its flowers, due to an ethereal oil, comnarin. The jjlant is known by 
 a variety of other names, such as Melilotus, Bokhara, giant clover, and 
 wild alfalfa. Three Sjjecies are common in the United States — the 
 white biennial {Melilotus alba Desr.), the large yellow biennial {Melilotus 
 officinalis Lam.), and the small yellow annual {Melilotus indica All.). 
 This in\'estigation concerns itself with the first sj^ecies, which is commonly 
 referred to as Bokhara, or merely as sweet clover. The name Bokhara is 
 obtained from a district in Asiatic Russia, supposedly the original home 
 of the plant. 
 
 Sweet clover is an erect, stemmy plant, reaching a height of from 
 eighteen to thirty inches the first year; a single ]jlant growing by itself 
 will tend to branch mfire than do ])lants growing together. When young 
 the plant resembles alfalfa, but its leaves are usually more broadly ovate 
 and its foliage is less dense. At bloom it is easily recognized by its 
 
 1 The work described in this bulletin was done in the Department of Agricultural Chemistry, under 
 the direction of Professor George W. Cavanaugh. 
 
122 Bt'lletin 394 
 
 perfume, a eharacteristie not so inarketl in early .growth unless the i:)]ants 
 are dried. A lon^^ taproot is early developed, which becomes fleshy 
 toward the end of the first season due to the large quantity of reserve 
 material stored in it. During the second season the plant makes a 
 growth of from five to twelve feet in height. In the middle of the summer 
 white flowers, borne on long, slender racemes, are produced. Death 
 occurs when the seed has matured. 
 
 In common with other legumes sweet clover possesses the power, 
 thru the agency of proper bacteria, of storing up atmospheric nitrogen 
 thru the nodules on its roots, thus enriching the soil in which it is grown. 
 
 REVIEW OF LITERATURE 
 
 A survey of the literature of the subject shows that very little definite 
 experimental work has been undertaken regarding sweet clover. Such 
 work as has been done has been limited, for the most part, to the study 
 of the plant as a forage crop and as a honey plant. Data are lacking 
 regarding its value as a green manure. Recently se\^eral authors have 
 pointed out its utility as a farm crop, but their writings are based on 
 general observations rather than on experimental results. 
 
 SWEET CLOVER Cl'LTl'RE 
 
 An excellent discussion of sweet clover is given by Westgate and Vinall 
 (191 2).- This publication contains a survey of the distribution of the 
 crop thruout the United vStates, together with some data regarding yields 
 and the purposes for which the crop is used. Methods of cultivation 
 are discussed, as well as the value of the plant for hay, for pasture, or 
 for soil improvement. The authors state that sweet clover possesses 
 a wider adaptability to soil types and climate than any of the true clovers 
 and probably than alfalfa. Attention is called to the fact that the plant 
 thrives in the most humid as well as the semi-arid sections of the country, 
 and produces a satisfactory growth on both the acid soils of the East 
 and the alkali soils of the West. It is stated further that a good stand 
 is obtained on soils too low in humus for the favorable growth of most 
 other legumes. Failures in obtaining a good stand are considered to be 
 due largely to faulty culture methods and to poor germination of the 
 seed. The authors conclude that sweet clover, properly handled, is a 
 valuable addition to the farm crops of many sections. 
 
 Another comprehensive study of sweet clover is contained in a bulletin 
 by Lloyd (191 2). A survey of the distribution of the crop in Ohio is 
 given, together with notes from twenty-nine foreign countries and thirty- 
 
 - Dates in parenthesis refer to bibliography, page 148. 
 
Decomposition of Sweet Clover as a Green Manure 123 
 
 four of the American experiment stations. The habits, soil adaptations, 
 and agricultural value of the plant are discussed. It is concluded that 
 the plant is peculiarly suited to adverse conditions, that it will thrive 
 where other legumes will not, and that as a green manure it adds more 
 humus-forming material to the soil than do any of the other nitrogen- 
 gathering plants. Lloyd states further that the best growth is secured 
 in soils rich in lime, and that inoculation with the appropriate bacteria 
 may be necessary. 
 
 Investigations to test the manurial value of sweet clover have been 
 limited to a study of its effect on a succeeding crop. Orth (1892) con- 
 ducted an experiment to note its effect as a green manure on succeeding 
 crops of oats and potatoes. In May sweet clover was sown in rye, and 
 in the summer of the following year the crop was plowed under. The 
 yield of potatoes was doubled in comparison with that of check ]jlats 
 which received no green manure. Better yields were secured where sweet 
 clover was turned under than on similar plats which received eight tons 
 of stable manure instead of green manvire. The yield of oats, grain and 
 straw, was increased by 60 per cent and 90 per cent, respectively. Green- 
 manuring with sweet clover also increased the growth of maize. 
 
 Westgate and Vinall (1912:30) describe an experiment conducted in 
 Alabama showing the effect of sweet clover on a succeeding corn crop. 
 On poor, run-down soil, 6672 pounds of sweet clover hay per acre were 
 produced the first year and 7048 pounds the second year. The stubble 
 was then plowed under and the field was planted to corn, 22.7 bushels 
 per acre being produced as compared with 16.2 bushels where sweet 
 clover had not been sown. 
 
 Hopkins (1910:219-220) mentions some Illinois investigations with 
 sweet clover regarding yield and nitrogen content at the end of the 
 second season's growth. Figures are given for roots and to]3S separately, 
 showing 86 per cent of the total nitrogen to be present in the latter. 
 It is concluded from the results obtained that sweet clover gives great 
 promise as a green-manure crop. 
 
 The bibliographical evidence indicates that sweet clover thrives under 
 a great variety of soil and climatic conditions, and that it is peculiarly 
 adapted to adverse situations. Data regarding yield and nitrogen content 
 show that the plant comi^ares favorably with the other legumes in its 
 ability to fix nitrogen. Further, surveys indicate that good stands have 
 been obtained where other legumes have failed. In view of all this 
 evidence in favor of sweet clover, it is jjertinent to ask why the crop 
 has been so little utilized in general farm j^ractice. One reason is that 
 the plant has long been regarded as a pest, difficult to eradicate. This 
 view has now been disjjelled, as a ])erusal of either Lloyd's or Westgate 
 
124 *<■ Bulletin 394 
 
 and Vinall's bulletin will show. A more valid objection to sweet clover 
 has been the difficulty experienced in obtaining a satisfactory stand, 
 especially the first season. This has been demonstrated to have been 
 due in part to faulty methods of cultivation, for a hard, compact seedbed 
 is required. The difficulty has probably been due in a larger degree to 
 poor germination because of hard seed, especially when combined with 
 lack of inoculation. Further, the fact that sweet clover is a biennial 
 has curtailed its use. Two years has seemed too long a time to give up 
 a field to a soil renovator, particularly as the woody nature of the plant 
 at maturity indicates that it would decompose slowly in the soil. The 
 lack of accurate ex]3erimental data has doubtless been another factor in 
 limiting the use of the crop. 
 
 A survey of the factors that have limited the utilization of sweet clover 
 in general fami practice indicates that some exact data are needed 
 relative to the following points; the ability of the plant to thrive on a 
 worn-out soil; methods of increasing the germinating power of the seed; 
 the value of lime; the possibility of securing a satisfactory yield the 
 first season; the rate of decay of the plant material when incorporated 
 with the soil; and the proper time for turning under to combine a satis- 
 factory yield with rapid decay. It was with the object of throwing 
 some light on these questions that the present investigation was undertaken. 
 
 DECOMPOSITION OF GREEN MANURE 
 
 No record has been found of any study of the rate of decay of sweet 
 clover as a green manure. The value of green manures has been more 
 frequently studied from the standpoint of their effect on succeeding crops 
 than by measurements of their rate of decay. This is natural, inasmuch 
 as the decomposition of organic matter is a complex process of many 
 stages and no really satisfactory methods have been devised for measuring 
 it. Decay of organic matter is caused by microorganisms, as first pointed 
 out by Wollny (1884), and this fact has been utilized in the methods 
 devised for studying the process. 
 
 Wollny (1886) measured the carbon dioxide evolved during decay. 
 This method leaves open the question as to the form into which the 
 nitrogen of the organic matter is converted. Inasmuch as plants take 
 up this element principally as nitrate, it is desirable to know to what 
 extent nitrification is going on, or at least to know that conditions are 
 favorable for this process. It is obvious that conditions which favor 
 initial decomposition and those favorable to nitrate formation are not 
 identical. Pagnoul (1895) has shown that under certain conditions 
 ammonium salts may form in abundance while nitrification ])ro])er is 
 
Decomposition of Sweet Clover as a Green Manure 125 
 
 retarded. It is also known that carbon dioxide may be used by micro- 
 organisms as a source of carbon, under which conditions its evolution 
 would not be a proper measure of the rate of decay. Hutchinson and 
 Milligan (1914) improved Wollny's method by supplementing the carbon 
 dioxide measurements with nitrate determinations. 
 
 In consideration of the fact that the decay of organic matter is a 
 bacteriological process, the problem of its measurement has been attacked 
 from that standpoint. This mode of attack was first suggested by Remy. 
 The underlying principle of the method is the determination of the kind, 
 and the intensity of the functions, of the bacteria concerned. This is 
 accomplished by inoculating nutrient solutions with portions of the soil 
 in question and detennining the metabolic products after a given period 
 of incubation. Many recent investigators have modified the original 
 method. Altho the Remy scheme and its modifications have been the 
 most widely adopted of the methods deyised for studying the decay of 
 organic matter, the procedure contains many weaknesses, as pointed 
 out by Allen and Bonazzi (1915). 
 
 The stage of decay which gives the most information as to the avail- 
 ability of a green manure is nitrate formation, since it is as nitrate that 
 nitrogen is taken up by the plant. If nitrification has occurred, it is 
 evident that other processes have preceded it, to furnish forms of nitrogen 
 capable of being nitrified. The accumulation of nitrates in soils under 
 a green-manure treatment has been used as a measure of the availability 
 of the material turned under. 
 
 Inasmuch as the measurement of nitrates formed was the method 
 finally adopted in the present investigation, it seems desirable to cite 
 instances when this procedure has been followed and considerations that 
 have led to its adoption. Early investigators believed that the presence 
 of organic matter retards nitrification. If this were so, the measurement 
 of nitrate accumulation would be a poor method of studying the decom- 
 position of a green manure. Recent work, however, seems to have 
 disproved the earlier views. An account of the experimental work on 
 this subject is given by Hill (19 15), who inade a thoro study of various 
 green manures on several types of soils under greenhouse conditions. 
 He found that organic matter, as bluegrass, clover, and alfalfa, passed 
 over into nitrates. The addition of the green manures increased the 
 rate of nitrate formation. Brown and Allison (1916) have shown that 
 the application of the common humus-fonning materials in maximum 
 amounts for farm conditions increases ammonification and nitrification 
 to a considerable extent. Wright (19 15), working with vetch and rye, 
 found that plowing under green manures resulted in rapid decay accom- 
 panied by vigorous nitrification. The work of Hutchinson and Milligan 
 
126 Bulletin 394 
 
 (19 1 4) has been referred to. These investigators studied the rate of 
 decay of sann hemp, using the accumulation of nitrates as the measure. 
 The plants were turned under at different stages of growth. The per- 
 centage of nitrification was found to decrease markedly with age. 
 
 The work of the last-named investigators, together with that of Hill, 
 shows the application of the study of nitrate formation to the problem 
 of the rate of decay of a green manure. The method is open to several 
 objections, but so are the other methods devised. The process of nitrate 
 formation is very sensitive to soil and atmospheric conditions, but this 
 objection to the method is less serious in pot experiments in the green- 
 house, where conditions can be carefully controlled and no nitrates are 
 lost by leaching. 
 
 In the present investigation the crude-fiber content of sweet clover 
 was considered. It is believed that the amount of carbohydrate material 
 present has a bearing on the rate of decay. Hill (191 5) found that pure 
 cellulose and paper retarded nitrification. Paterson and Scott (19 12) 
 found that starch and sugar retarded nitrification. Hutchinson and 
 Milligan (19 14) state that the rate of decay of plant material depends 
 on the percentage of vascular ring and lignified parts. Brown and 
 Allison (19 16) found that straw did not increase nitrification as much 
 as did green manures; however, they found no relation between the 
 nitrogen-carbon ratio and nitrification. Wright (19 15) found that plow- 
 ing under materials such as old hay, leaves, and strawy manure, and 
 even green manure which had become dry, reduced the quantity of 
 available nitrogen in the soil. It apj^ears that crops at later stages of 
 growth decay less rapidly, not only because they contain more fiber but 
 also because they contain less water. 
 
 EXPERIMENTAL WORK 
 
 The experimental work described in this paper extended over two 
 seasons. In 19 14 a general study was made of crop yield and rate of 
 decay at different stages of growth and under different conditions as 
 regards liming. In 19 16 a further study was made of the period of 
 growth which seemed most desirable on the basis of the results obtained 
 the previous season. 
 
 SOIL USED 
 
 The soil used in the experiments was classified by the United States 
 Bureau of Soils as Volusia silt loam. The Volusia series occupy an 
 estimated area of about 10,000,000 acres in the United States, ])rincipally 
 in New York, Ohio, Indiana, and Pennsylvania. Th^se soils occur 
 in New York State to the extent of over 20 ])er cent of the total area, 
 comprising about 35 per cent of the tillable land. Regarding the general 
 
Decomposition of vSweet Clover as a Green Manure 127 
 
 character of the X'okisia soils, Carr (i(;o(;: 11) states: " In recent years 
 extreme difiiculty has been exjjerienced in seeihni; clover, amounting,' 
 in many cases to comjjlete failure. Corn seldom ^ives any yield of 
 mature grain; wheat yields have become so low that attempts to grow 
 this crop have been abandoned. . . . In the region occupied by the 
 silt loam many farm homes are abandoned." Carr concludes that' this 
 unproductivity is due to poor physical condition, lack of organic matter, 
 and general unfavorable conditions for the development of bacteria. 
 Hence this type of soil, because of its general distribution and run-down 
 character, is an excellent one on which to try the restorative powers of 
 sweet clover. 
 
 The soil was obtained from the fanii of Mi\s. Ellen Crutts, at Varna, 
 New York, and came froin a field that had recently received no application 
 of manure or commercial fertilizer. The soil contained approximately 
 35 per cent of pebbles too large to ])ass a half-inch mesh. A sample was 
 analyzed according to the methods described b\' the United States Bureau 
 of Chemistry.^ The mineral constituents were determined in the solution 
 obtained on strong acid digestion. Phosphorus was weighed as mag- 
 nesium pyrophosphate, and calcium as the oxide. Potassium was deter- 
 mined by the Lindo-Gladding method. The loss on ignition over a 
 Bunsen was considered volatile matter. As a measure of the organic 
 matter present, the organic carbon was determined by combustion with 
 chromic acid, using the modification described by Cameron and Breazeale 
 (1904). The inorganic carbon was first determined and deducted from 
 the total obtained on combustion. A determination of humus was also 
 made according to the procedure described by the United States Bureau 
 of Chemistry. Dry ammonium carbonate was added to the ammoniacal 
 solution to flocculate the clay particles before filtering. Nitrogen was 
 obtained by the Gunning method, modified to include the nitrogen of 
 nitrates, using copper wire as a catalytic agent. The nitrates were 
 determined on a dry sample by the phenol-disulfonic-acid method. Com- 
 parisons were made in the colorimeter designed by Schreiner. The 
 results obtained on analysis of the soil used in the first year's work 
 were as follows, the percentages being based on the moisture-free fine 
 soil: 
 
 Potassium oxide o . 3683 per cent 
 
 Calcium oxide o . 475 per cent 
 
 Phosphorus pentoxide o. i486 per cent 
 
 Volatile matter 6.05 per cent 
 
 Organic carbon i ■ 36 per cent 
 
 Humus I . 1 5 per cent 
 
 Nitrogen o. 183 per cent 
 
 Nitrates 57 parts per million 
 
 3 Official and provisional methods of analysis, Association of Official Agricultural Chemists. Bulletin 
 107, U. S. Bureau of Chemistry. igo8. 
 
128 BULLKTIN 3Q4 
 
 In addition to the detemiinalions listed above, the hme requirement 
 was obtained by the Veitch method. One gram of moisture-free fine 
 soil was satisfied by 0.00103 gram of calcium oxide, corresponding to a 
 requirement of about 3600 pounds per acre-foot. 
 
 The soil used in the second year's work was obtained from the same 
 place as that on which the above analyses were made. It was examined 
 for nitrogen, nitrates, and lime requirement, with the following results: 
 
 Nitrogen 0.1857 per cent 
 
 Nitrates 3 parts per million 
 
 Lime requirement o.ooi i gram CuO per gram of soil 
 
 The determination for nitrates was made on the moist soil as placed 
 in the pots, and the result obtained was figured to the moisture-free basis. 
 
 The results of the chemical analysis show the soil to be deficient in 
 calcium oxide, and there is a lack of organic matter indicated by the 
 determinations for organic carbon and humus. The figure for volatile 
 matter is included merely for comparison with the two determinations 
 mentioned above, as a means of showing that the loss on ignition is no 
 measure of the organic matter present in a soil of this type. The soil 
 is acid according to the Veitch method. The soil as analyzed contained 
 considerable organic material still retaining a definite cell structure. 
 This fact explains why the figure obtained for ]3ercentage of organic 
 carbon is higher than that for htrmus, and, when considered with the 
 result obtained for lime reqtiirement, indicates that conditions were 
 unfavorable for decomposition to proceed. Thus, not only is the soil 
 deficient in organic matter and nitrogen, but conditions are such that 
 the conversion of ]K)tential into actual ]jlant food must be occurring at 
 a very slow rate. 
 
 A striking difi^erence in the two samples of soil is noted as regards 
 nitrates. Tho both samples were taken at about the same time of year, 
 the second season was a much later one. The spring months were' 
 characterized by cold, wet weather, a condition inhibiting the formation 
 of nitrates. The crop grown on the soil the year previous to that in 
 which the first sample was taken was potatoes; the field was in grass 
 when the second sample was obtained. The inhibiting effect of mixed 
 grasses on the formation of nitrates has been shown by Lyon and Bizzell 
 
 (1913)- 
 
 THE SEED 
 
 The seed used in the experiment was furnished by the Bokhara Seed 
 Company, of Falmouth, Kentucky, and was obtained from a crop grown 
 the previous year in Pendleton County, Kentucky. Inasmuch as one 
 of the causes of failure in obtaining a satisfactory stand of sweet clover 
 
Decomposition of Sweet Clover as a Green Manure i2g 
 
 had been fouiitl Uj be due to poor ^'ennincilioii, it seemed desirable to 
 
 run germination tests on the seed used, and also to try the sulfuric-acid 
 
 treatment. This method of treating seed to increase the i^ercentage of 
 
 germination is described by Love and Leighty (191 2). It consists 
 
 essentially in allowing the seed to stand in contact with concentrated 
 
 sulfuric acid until the hard seed coat is softened or removed. In these 
 
 experiments sufficient acid was added to make sure that all seed was 
 
 coated with it, and the mixture was allowed to stand for twenty minutes. 
 
 It was then dumped into a large volume of water in order to prevent 
 
 the heat of dilution from injuring the seed. Washing by decantation 
 
 caused the removal of most of the loosened hulls. Finally the seed was 
 
 washed on a cheesecloth filter until it was free from acid, and allowed to 
 
 dry. The treated and the untreated seeds were then tested by gemiinating 
 
 between filter paper. One hundred seeds were used for each trial. The 
 
 results were as follows: 
 
 Percentage of 
 Untreated germination 
 
 seed after four 
 
 days 
 
 Lot I 28 
 
 Lot 2 30 
 
 Lot 3 32 
 
 Treated 
 seed 
 
 Lot I 97 
 
 Lot 2 98 
 
 Lot 3 97 
 
 Lot 4 95 
 
 It is seen that the percentage of germination was more than trebled 
 by the acid treatment, and that practically complete germination was 
 obtained. Thirty i)er cent is probably a fair average of the results obtained 
 by other investigators in germination tests on untreated seed. A treat- 
 ment by which the quantity of seed required may be reduced by from 
 one-half to two-thirds, should be worth while. The procedure requires 
 little time, and the method is inexpensive since commercial acid may 
 be used. 
 
 There mav be some question whether the average farmer would be 
 successful in treating seed by this method. In this connection it seemed 
 desirable to investigate the eftect of age on seed given the acid treatment. 
 The following results were obtained ten months after the treatment 
 was used: 
 
130 Bulletin 394 
 
 PercetiUige of 
 Untreated germination 
 
 seed after four 
 
 days 
 
 Lot I 28 
 
 Lot 2 29 
 
 Treated 
 seed 
 
 Lot I 94 
 
 Lot 2 95 
 
 Lot 3 96 
 
 These results are important in showing that it is not necessary to reserve 
 the acid treatment until just before the lot of seed is to be used, but that 
 seed so treated may be kept for months with no appreciable decrease 
 in the percentage of germination. Thus the process need not be left 
 to the farmer, in whose hands it might not be a success, but may be 
 carried out at the seed farm or by the dealer.'* 
 
 THE POTS 
 
 The pots in which the experiments were conducted were 9.5 inches 
 in diameter and 10 inches deep. The soil was sifted to remove the larger 
 pebbles, and a portion equivalent to about 25 pounds of moisture-free 
 fine soil was put into each pot. The soil was inoculated with the sweet- 
 clover organism by the pure-culture method. The cultures were obtained 
 from the Laboratory of Plant Ph}-siology of the New York State College 
 of Agriculture. The pots were kept in the greenhouse and maintained 
 at a moisture content of 30 per cent. Quartz sand was spread over the 
 surface of the soil to lessen evaporation. 
 
 SEASON OF 1914 
 
 In accordance with the procedure just described, 36 pots were set up 
 in 1914. In order to test the effect of lime on the growth of sweet clover, 
 the pots were divided into three series. In one series sufficient slaked 
 lime was added to satisfy the lime requirement, while in another an 
 equivalent quantity of finely pulverized ground limestone was used. In 
 order to accurately meet the requirement the lime used was analyzed. 
 The third series received no lime. The seed was placed in the pots on 
 May 20. In sowing, the seedbed was rendered compact and the seed 
 
 * Since this work was done, Professor Hughes, of the Iowa Agricultural Experiment Station, has devised 
 the Ames hulling and scarifying machine for the treatment of hard seed. This machine is described in 
 Farm and Fireside. June 19, 1915. By treatment with it. germination is increased from 50 to 90 per 
 cent. Many seed firms have installed the machine and have found it practical in every respect. 
 
Decomposition of Sweet Clover as a Green Manure 131 
 
 was covered v/ith only a thin layer of soil. At the end of ten days the 
 plants were thinned, leaving twelve in each pot. 
 
 The pots were divided into three groups as regarded tiine of harvest. 
 At each period, four of the pots limed with hydrate, four of those limed 
 with carbonate, and four of the check pots, were harvested. From six 
 of these pots the roots and the tops were removed and a sample of the 
 soil was taken. The plants from the remaining six pots were chopped 
 up and incorporated with the soil, pending later examination for rate of 
 decay. Before the plant material was added a sample of the soil was 
 taken. The pots in which the plants were turned under were kept at a 
 moisture content of 30 ]jer cent, in order that conditions might be 
 favorable for decomposition. The pots from which the plants were 
 removed were kept under similar conditions, so that they might serve 
 as checks on any measurements made later regarding the rate of decay 
 of the material turned under. 
 
 According to the above procedure, the thirty-six pots were harvested, 
 in groups of twelve, on July 21, August 17, and September 15. These 
 dates represented periods of growth of sixty-two, eighty-nine, and one 
 hundred and eighteen days, or approximately two, three, and four months, 
 respectively. At each time of harvest a photograph was taken of repre- 
 sentative pots in order to give an idea of the amount of growth and the 
 differences noticeable according to the different treatments of the soil. 
 These photographs are shown in figures 9, 10, and 11. A foot rule is 
 shown in each pot. 
 
 It is seen that at the end of the first period the plants had reached a 
 height of from eight to twelve inches. Differences due to lime are not 
 noticeable. At the end of three months a height of from fifteen to twenty 
 inches had been attained and many lateral branches had been sent out. 
 During the final period the height increased from five to eight inches 
 more. This period was especially characterized, as the figure indicates, 
 by a development of dense foliage. In both the second and the third 
 period the effect of lime is evident. 
 
 In taking up the plants at the different periods, some observations 
 were made regarding root development and nodule formation. During 
 the first two periods slender taproots were developed, many attaining 
 a length of four or five feet. During the final period the roots grew 
 somewhat in length, but the most noticeable development was a thickening 
 and branching a few inches below the root crown. At the first harvest 
 it was noticed that practically all the roots bore nodules, the roots in 
 the limed pots being more plentifully supplied. This is in accordance 
 with what has previou.sly been observed regarding the influence of lime 
 on nodule formation. In no case, even at the end of the final period. 
 
132 
 
 Bl'LLETIN 394 
 
 l.'nunif.i Slakr,] Inr,. Cmiin.l 1 mmc-.i . nic 
 
 Fig. 9. RESULTS from different soil treatments, on plants -grown for 62 DAYS 
 
 Fig. 10. results i'kom diiferent soil treatments, on plants grown for 89 days 
 
Decompositiox of Sweet Clover as a Green Manuj?e 133 
 
 Lnlinicd Slaked lime Grouiul limestcjiie 
 
 Fig. I I . RESULTS from different soil treatments, on plants grown for I 18 DAYS 
 
 were large nodules formed. Howe\'er, nodules were present in abundance, 
 as is shown b\- the fact that as many as sixty were counted on a single 
 root system. 
 
 The plants removed from the various pots were allowed to dry in the 
 air and were weighed. They were then pulverized and bottled for later 
 analysis. Similarly the soil samples were sieved and bottled after drying. 
 These samples were analyzed for nitrates according to the inethods 
 previously described. The crops harvested were examined for nitrogen 
 and cnide fiber. The nitrogen was determined by the Gunning method, 
 with the addition of copper wire as a catalyst. The crude fiber was 
 determined according to the procedure outlined by the United States 
 Bureau of Chemistry.'^ This determination was deemed important 
 because the rate of decay of any plant material must be largely governed 
 by the amount of woody tissue present. Thus it is desirable to know 
 the percentage of crude fiber present at those stages when the plant 
 might be used as a green manure. 
 
 ' Official and provisional methods of analyses. Association of Official Agricultural Chemists. Bulletin 
 107, U. S. Bureau of Chemistry. 1908. 
 
134 
 
 Bulletin 394 
 
 The results of the examination of the ])lant material are given in 
 tables 1 , 2 . and 3 : 
 
 TABLE I. Yield and Composition of Crops 
 
 (Duration of growth, 62 days) 
 
 Pot 
 
 I 
 
 2 
 
 5 
 6 
 
 9 
 10 
 
 Treatment 
 
 Unlimed 
 
 Un]imed 
 
 Slaked lime 
 
 Slaked ime 
 
 Ground limestone 
 Ground limestone 
 
 Dry matter produced 
 
 Grams 
 
 Percentage 
 
 of 
 
 nitrogen 
 
 41 
 29 
 
 ■52 
 .40 
 
 43 
 53 
 
 Nitrogen 
 
 produced 
 
 (grams) 
 
 0.08150 
 o . 08620 
 o . 07920 
 0.08670 
 0.09673 
 0.08860 
 
 Crude 
 
 i'llx-r 
 (per cent) 
 
 20 . 50 
 20.01 
 21 .04 
 23.19 
 21.13 
 21.78 
 
 TABLE 2. Yield and Composition of Crops 
 (Duration of growth, 89 days) 
 
 Pot 
 
 13 
 14 
 17 
 18 
 21 
 22 
 
 Treatment 
 
 Unlimed 
 
 Unlimed 
 
 Slaked lime 
 
 Slaked lime 
 
 Ground limestone 
 Ground limestone 
 
 Dry matter produced 
 
 Grams 
 
 9.61 
 
 8.81 
 
 12 10 
 
 10.50 
 
 II 45 
 n . 10 
 
 Percentage 
 
 of 
 nitrogen 
 
 Nitrogen 
 
 produced 
 
 (grams) 
 
 0.3287 
 o 3145 
 0-4344 
 0.3906 
 0.4340 
 0.4285 
 
 Crude 
 
 fiber 
 
 (per cent) 
 
 25 33 
 2?> ^3 
 25 . 82 
 25 63 
 23.84 
 2451 
 
 Pot 
 
 TABLE 3. Yield and Composition of Crops 
 
 (Duration of growth, 118 days) 
 
 Treatment 
 
 Dry matter produced 
 
 Grams 
 
 Percentage 
 
 of 
 nitrogen 
 
 Nitrogen 
 produced 
 
 (grams) 
 
 Crude 
 fiber 
 
 (per cent) 
 
 25 
 26 
 29 
 30 
 
 33 
 34 
 
 Unlimed 
 
 Unlimed 
 
 Slaked lime 
 
 Slaked lime 
 
 Ground limestone 
 Ground limestone 
 
 .42 
 .29 
 30 
 
 ■25 
 ■03 
 •19 
 
 0.6416 
 0.6399 
 0.8451 
 0.7881 
 0.8175 
 0.8020 
 
 52 
 40 
 66 
 73 
 76 
 79 
 
Decomposition of Sweet Clover as a Gkeen Manure 135 
 
 It is seen in table 1 that little dry matter was ijroduced in the first 
 two months of growth and that no differences due to liming are discernible. 
 During the third month the growth increased fourfold on the unlimed 
 pots and an even greater growth resulted in the soil to which lime had 
 been added. The dry matter produced during the first three months 
 was doubled during the fourth, and here again the effect of lime was 
 manifest. The results indicate that in order to obtain the largest yield 
 the lime requirement of the soil should be satisfied; the yields for the 
 two kinds of lime do not differ sufficiently, however, to permit any 
 suggestion as to w^hich is the better form to apply. 
 
 The results for nitrogen content show that sweet clover compares 
 favorably in this respect with other legumes. The percentages do not 
 vary widely, but the figures indicate that the nitrogen content per gram 
 of dry weight reaches a maximum before the end of the four-months 
 growth. The figures showing the amount of nitrogen produced follow 
 those for dry matter rather closely. 
 
 The percentage of fiber appears to be somewhat greater at the end 
 of the second period of growth than at either of the other times of 
 examination. This seems rather surprising, but probably may be explained 
 on the basis of the cultural observations made. It was observed that 
 during the second period the greatest increase in height and thickness 
 of stems occurred, while the final period was characterized by a develop- 
 ment of foliage. Thus in the last period a proportionally smaller amount 
 of woody tissue was fonned and the percentage of crude fiber dropped. 
 A comparison of reported analyses of sweet clover for fiber shows a wide 
 variation in the results obtained, depending on the locality, the time 
 of cutting, and other factors. In general, analyses show a crude fiber 
 content of from 30 to 35 per cent at blooining time the second season. 
 From the standpoint of rate of decay, the crop should be of much more 
 value at the end of the first season, when the fiber content is from 10 
 to 15 per cent less. 
 
 Figuring the results obtained in the pot experiments to the acre basis, 
 it is found that at the end of the four-months growth the yield from the 
 pots limed with ground limestone corresponds to about two tons per 
 acre. The nitrogen produced corresponds to 115 pounds per acre, an 
 amount to supply which would require ten tons of stable manure. It 
 is realized that this computation is of limited value, in that considerable 
 error may enter into the calculation. Further, it is realized that results 
 under field conditions might vary considerably from those obtained in 
 pot experiments. However, the figures arc iiseful for the interpreta 
 tion of the results given in the tables, and they indicate what sweet clo\'er 
 would do in the same soil in the field under optimmn conditions. 
 
136 Bulletin 394 
 
 It has already been stated that no method has been devised which 
 will measure the rate of decay of organic matter thru all its stages, and 
 that none of the methods in use at the present time for studying this 
 problem are free from objections. The question arose as to whether a 
 more satisfactory method could be devised. In this connection it seemed 
 possible that some modification of the alkaline permanganate method as 
 devised by Jones (19 12) might be applicable. In this method a sample 
 containing 50 milligrams of water-insoluble organic nitrogen is washed 
 free from soluble salts, and digested below the distillation point with a 
 definite amount of alkaline permanganate of definite strength for thirty 
 minutes. The temperature is then raised, and during the next sixty 
 minutes 95 cubic centimeters of the sam]jle is distilled over into standard 
 acid. The nitrogen thus obtained is designated as active water-insoluble 
 nitrogen. It is considered to be that portion which will rapidly become 
 available for plant use. The method has demonstrated its value in 
 differentiating between high- and low-grade sources of nitrogen in com- 
 mercial fertilizers. Its disadvantages are that it is empirical and may 
 give varying results with different investigators. As regards its adapt- 
 ability to the determination of the availability of soil nitrogen, it seemed 
 possible that with a given soil and a given source of organic matter the 
 rate of decay might be indicated by the increasing amounts of ammonia 
 distilled from the permanganate solution. It seemed that if the manipu- 
 lations were carried out with exactness, results comparable with one another 
 might be obtained. The possibility was considered worthy of experiment. 
 
 After a number of trials of the alkaline permanganate method, it was 
 found to be inapplicable to the pur]30se desired. A few of the results 
 obtained are given, because of their bearing in explaining the failure 
 of the method. A sample of soil containing the prescribed amount of 
 organic nitrogen was treated according to the procedure outlined for a 
 fertilizer material. Similarly, a sample of plant material was examined. 
 Next, a composite sample of soil and plant material containing the 
 required amount of organic nitrogen was run. The active nitrogen 
 obtained was about 25 per cent below the theoretical amount on the 
 basis of the materials examined separately. Using a larger amount of 
 permanganate solution, the results were somewhat higher but still did 
 not equal the theoretical. Samples of soil were next examined with 
 varying amounts of the reagent. The following results were obtained: 
 
 Amount of reagent Active nitrogen 
 
 (cubic centimeters) (grams) 
 
 100 (prescribed amount) o .01876 
 
 125 0.02072 
 
 150 o .023 10 
 
 200 0.02501 
 
Decomposition of Sweet Clover as a Green Manure 137 
 
 Thus the amount of active nitrogen increased with the amount of the 
 reagent. On examination it was found that the supernatant hquid in 
 the Kjeldahl flask after digestion was colorless in each case. This showed 
 that all the permanganate had been used up; thus it, and not the organic 
 matter, became the limiting factor in determining the amount of nitrogen 
 obtained. It became evident that the reagent was reacting with other 
 substances besides the organic matter, notably the iron of the soil. A 
 trial on the portion of the soil remaining after ignition showed this to 
 be true. It was thus seen that the alkaline pennanganate method was 
 not applicable to the determination of the availability of soil-organic 
 nitrogen unless the organic matter could be separated from the mineral 
 part. In this connection the method of separating the organic from the 
 mineral constituents of commercial fertilizers as devised by Jones and 
 Anderson (191 4) suggested itself. Jones has applied this method to 
 muck soils high in organic matter. It was not found to work satisfactorily 
 for light soils, and was not applicable to the work with Volusia silt loam. 
 
 Attention was next turned to the methods previously mentioned as 
 used in measuring rate of decay ; and, inasmuch as these experiments were 
 conducted under well-defined conditions, it was decided to adopt the 
 method of determining the nitrates formed. It was further decided to 
 study only the ])ots in which the crops had been turned under at the 
 end of the second and tliird periods, since the dry matter produced at 
 the end of two months growth proved to be too little to make profitable 
 the turning-under of a crop at this period. At harvest, the nitrates were 
 deteiTnined in all the pots of the last two series. The jjots were then kept 
 under constant conditions for four months and the nitrates again deter- 
 mined. At the beginning of the period an effort was made to have the 
 soil at the same degree of compactness so that the oxygen supply might 
 not vary, and a moisture content of 30 per cent was maintained thruout. 
 
 It is realized, of course, that nitrates accumulated under a green 
 manure may not necessarily ha\'e resulted from its decomposition. It 
 may be considered that the green manure is changed to nitrates, or it 
 may be considered that the organic matter increases the nitrification 
 of sources of nitrogen already ]3resent in the soil. Doubtless both these 
 processes obtain. However, where check pots having no green manure 
 are used, as in these experiments, the net gain in nitrates as a result of 
 adding the green manure can be shown. The production of available 
 nitrogen as a result of the addition of the organic matter is thus shown, 
 and this is really the object (^f an exjjeriment of this kind. 
 
 The samples taken for the nitrate determination were allowed to 
 air-dry beff)re the analysis was made and the results were then figured 
 to tlie moislurc-frce basis, ll is rcahzcd llial in drying in air, nitrate 
 formation is accelerated, and thai as a consequence the results so obtained 
 
138 
 
 Bl'LLETlN 394 
 
 are higher than when the examination is made immediately after sampHng. 
 Circumstances prevented the following of the latter ])rocedure. Inasmuch 
 as the samples taken when the crop was turned under, and after the period 
 of four months, were handled identically, the figures for the difference 
 in nitrate content at these periods should not be greatly affected by the 
 increase due to drying. . 
 
 The nitrates formed during four months in the pots having the plants 
 removed or turned under after a growth of 89 days, are shown in table 
 4, and in table 5 are shown the same results for the ii8-days period: 
 
 TABLE 4. Formation of Nitr.a.tes 
 (Crop turned under after 89 days growth) 
 
 Pot 
 
 Treatment 
 
 Treatment of crop 
 at harvest 
 
 Nitrates 
 
 at 
 
 harvest 
 
 (parts per 
 
 million) 
 
 Nitrates 
 
 after 
 
 four months 
 
 (parts per 
 
 million) 
 
 13 
 
 14 
 
 Unlimed 
 
 Unlimed 
 
 Removed 
 
 Removed 
 
 Turned under 
 
 Turned under 
 
 Removed 
 
 Removed 
 
 Turned under 
 
 Turned under 
 
 Removed 
 
 Removed 
 
 Turned under 
 
 Turned under 
 
 72 
 
 59 
 
 55 
 
 69 
 
 76 
 
 87 
 
 89 
 
 124 
 
 149 
 
 140 
 
 160 
 
 129 
 
 59 
 53 
 45 
 60 
 
 15 
 
 t6 
 
 Unlimed 
 
 Unlimed . . 
 
 17 
 18 
 19 
 
 Slaked lime 
 
 Slaked lime 
 
 Slaked lime 
 
 70 
 
 87 
 148 
 212 
 209 
 181 
 339 
 270 
 
 20 
 21 
 
 22 
 
 23 
 24 
 
 Slaked lime 
 
 Ground limestone 
 
 Ground limestone 
 
 Ground limestone 
 
 Ground limestone 
 
 
 
 
 TABLE 5. Formation of Nitrates 
 (Crop turned under after 1 1 8 days growth) 
 
 Pot 
 
 Treatment 
 
 Treatment of crop 
 at harvest 
 
 Nitrates 
 
 at 
 
 harvest 
 
 (parts per 
 
 million) 
 
 Nitrates 
 
 after 
 
 four months 
 
 (parts per 
 
 million) 
 
 
 Unlimed 
 
 Unlimed 
 
 Unlimed 
 
 Removed . . 
 
 54 
 
 61 
 
 72 
 
 66 
 
 99 
 
 145 
 
 135 
 
 147 
 
 70 
 
 66 
 
 140 
 
 124 
 
 76 
 
 26 
 
 Removed 
 
 64 
 
 27 
 28 
 29 
 30 
 31 
 32 
 
 33 
 34 
 35 
 36 
 
 Turned under 
 
 Turned under 
 
 Removed 
 
 Removed 
 
 Turned under 
 
 Turned under 
 
 Removed 
 
 Removed 
 
 .Turned under 
 
 Turned under 
 
 185 
 
 Unlimed 
 
 Slaked lime 
 
 Slaked lime 
 
 Slaked lime. . . 
 
 122 
 
 185 
 225 
 319 
 
 Slaked lime . 
 
 422 
 
 Ground limestone 
 
 Ground limestone 
 
 Ground limestone 
 
 Ground limestone . ... 
 
 I3f^ 
 107 
 
 444 
 367 
 
 
 
Decomposition of Sweet Clover as a Green Manure 139 
 
 It is noted tliat in each period there were four pots treated alike as rej^ards 
 liming, and that at harvest the erop was removed from two of these pots 
 and turned under in the other two. It is natural that nitrification should 
 have occurred in all the pots, with the possible exception of the unlimed 
 ones, due to the favorable conditions maintained. 
 
 The original soil contained 57 parts of nitrate per million on the 
 moisture-free fine basis, or 53 parts per million in the dry soil as put 
 into the pots. Thus by noting the figures for nitrates at harvest it is 
 seen that there was an increase during the period of growing the crop, 
 an increase more marked in the limed pots. In table 4 it is shown that 
 in the unlimed pots the nitrates did not increase in the four months 
 following harvest. It may be assumed that decomposition occurred to 
 some exteAt, as evinced by the fact that at the end of the period the 
 plant material, with the exception of a few roots, could not be distinguished 
 in the soil, but that conditions were not right for complete decay. In 
 table 5 it is seen that following harvest the nitrates increased somewhat 
 in the unlimed pots from which the crop was removed, and that there 
 was a decided increase in the pots in which the material was turned 
 under. Nitrate formation was much greater in the limed than in the 
 unlimed pots. 
 
 The examination of the original soil placed in the pots when the culture 
 experiments were started indicated that conditions had been unfavorable 
 for decomposition processes to go on. It may be considered that the 
 physical conditions and the water relationships for the soil in the pots 
 were more favorable for decay than the conditions obtaining in the field 
 from which the soil was taken. However, a comparison of the results 
 obtained for the limed and the unlimed pots indicates that lime was the 
 larger factor in bettering the conditions favoring decomposition, due to 
 its effect in producing a more favorable medium for the activity of micro ■ 
 organisms. 
 
 A study of tables 4 and 5 shows that the results vary rather widely 
 in duplicate pots, and makes evident the sensitiveness of nitrate formation 
 to slight variations even under accurately controlled conditions. In the 
 discussion at the end of this paper the probable errors of the results are 
 computed, in order to show how large variations may be expected under 
 the conditions of these experiments. 
 
 season of 1916 
 The experimental work conducted in 19 14 was repeated in part m 
 19 16. The data obtained in the first season indicated that, from the 
 standpoint of amount of available green manure produced, the four- 
 months period gave the best results. Consequently, it was decided to 
 
T40 
 
 BiTLLETIN 304 
 
 i^row Llic ijlaiiLs lor tliis ])cri()(l in i(;i6 before iurninj.^' under. In view 
 of the previous season's work it seemed desirable that a repetition of the 
 experiment should ]:)roduce results of greater value in drawing conclusions 
 as to rate of decay. It was hoped that the variations in the results for 
 nitrate formation might be reduced by a repetition of the work. 
 
 The soil used in the second season's work is described on page 126. 
 Seed was again obtained from the Bokhara Seed Company. Four pots 
 limed with slaked lime, four limed with ground limestone, and four 
 unlimed, were set u]j. The seed was sown on Jtme 2. 
 
 The cultural observations made were similar to those reported for 
 1 9 14. During the first six weeks diiferences due to lime were not 
 noticeable. Thereafter, however, more rapid growth on the limed soil 
 became increasingly evident. Toward the end of the growings period the 
 plants on the unlimed soil were characterized by an unhealthy color and 
 by the dying of the leaves. 
 
 The plants were harvested after a growth of 116 days. The roots and 
 the tops were removed from all the pots and a sample of the soil was 
 taken. As in 1914, the plants from two pots of each series as regards 
 liming were saved for analysis. The plants from the remaining pots 
 were incorporated with the soil. All pots were then maintained at a 
 moisture content of 30 per cent, pending later exainination for ra e of 
 decay. 
 
 In taking up the plants it was evident that the roots did not bear as 
 many nodules as in 1914, but the nodules were somewhat larger. The 
 limed plants were more plentifully supplied and their root systems were 
 much better developed. This was markedly shown by branching at 
 the crown. 
 
 The plants removed were allowed to air-dry and were then analyzed, 
 as in 1 9 14. The re.sults are given in table 6: 
 
 TABLE 6. Yield and Composition of Crops 
 (Duration of growth, 116 days) 
 
 
 Treatment 
 
 Dry matter produced 
 
 Nitrogen 
 produced 
 
 (grams) 
 
 Crude 
 
 Pot 
 
 Grams 
 
 Percentage 
 
 of 
 
 nitrogen 
 
 fiber 
 (per cent) 
 
 40 
 41 
 
 44 
 45 
 
 48 
 
 49 
 
 Unlimed 
 
 18.76 
 17-77 
 33-65 
 30 94 
 29 -45 
 25-51 
 
 2.77 
 
 2.79 
 3.22 
 3.00 
 2.97 
 332 
 
 0-5197 
 0.4958 
 I 0835 
 0.9282 
 
 0.8747 
 . 8469 
 
 23-13 
 
 Unlimed 
 
 Slaked lime 
 
 24.24 
 23 34 
 
 vSlaked lime 
 
 Ground limestone 
 
 22.05 
 23-85 
 
 Ground limestone 
 
 24 ■ 85 
 
 
 
Decomposition of vSwert Clover as a Green Manitre 141 
 
 These results suljslantiale (hose njjiained in 1(^14. A c<;niijarison of 
 tables 3 and 6 shows that lime ijrodueed a somewhat K'rcater effect in 
 1916 than in the previous season. The figures for percentage of nitrogen 
 are similar exce])t as regards the imlimed pots. It was noted that the 
 plants growing on the unlimed soil appeared less vigorous in 19 16 than 
 in 19 14. That they were poor in color, indicating lack of nitrogen, has 
 already been stated. The figures for fiber content agree fairly closelv 
 with those obtained the previous season. 
 
 The samples of soil taken at harvest were examined for nitrates in 
 the moist condition. Inasmuch as the drying of a soil is known to affect 
 its nitrate content, it was thought that this procedure might give more 
 accurate and more uniform results than were obtained in 1914. Following 
 harvest the i3ots were maintained at a moisture content of 30 per cent 
 for four months. During this period all possible care was taken to keep 
 conditions similar in the different pots. In sampling the soil at the end 
 of the four-months period, no tops still showing their original form were 
 found; however, several imdecomi^osed root crowns were found in each 
 of the pots in which the plants had been turned under. The samples of 
 soil taken at this period were also examined for nitrates immediately 
 after sampling. The results of these determinations are given in table 7 : 
 
 TABLE 7. FoRM.\TiON of Nitr.\tes 
 (Crop turned under after 116 days growth) 
 
 Pot 
 
 Treatment 
 
 Treatment of crop 
 at harvest 
 
 Nitrates 
 
 at 
 
 harvest 
 
 (parts per 
 
 million ) 
 
 Nitrates 
 
 after 
 
 four months 
 
 (parts per 
 
 million) 
 
 40 
 41 
 42 
 
 43 
 44 
 45 
 46 
 
 47 
 48 
 
 49 
 50 
 51 
 
 Unlimed 
 
 Unlimed 
 
 Unlimed 
 
 Unlimed 
 
 Slaked hme 
 
 Slaked lime 
 
 Slaked lime . . 
 
 Removed 
 
 Removed 
 
 Turned under 
 
 Turned under 
 
 Removed 
 
 Removed 
 
 Turned under 
 
 Turned under 
 
 Removed 
 
 5 
 7 
 5 
 4 
 17 
 II 
 
 9 
 
 17 
 8 
 
 14 
 8 
 
 9 
 
 20 
 24 
 47 
 25 
 60 
 
 41 
 283 
 230 
 52 
 60 
 240 
 228 
 
 Slaked lime 
 
 Ground limestone 
 
 Ground limestone 
 
 Ground limestone 
 
 Removed 
 
 Turned under 
 
 Turned under 
 
 
 
 The soil used in 19 16 contained 3 parts of nitrates per million. It 
 is seen from the table that there was an increase as a result of growing 
 the crop, and that the increase was more marked in the limed pots. 
 The results obtained for nitrate formation due to turning under the crop 
 follow, in general, those obtained in the previous season. A statistical 
 studv of this table is made in the general discussion of results. 
 
142 
 
 Bulletin 394 
 
 DISCISSION OF RESTLTS 
 
 A brief discussion of the results furnished by this investigation has 
 been given following the tables in which the data obtained are listed. 
 For a consideration of the results to determine what conclusions may be 
 drawn from the experimental work, it is necessary to study the data for 
 both seasons as a unit, and in so far as possible to apply statistical methods 
 in this study. 
 
 Production of dry matter and nitrogen 
 
 A glance at table i (page 134) shows that the different methods of 
 treatment did not result in any evident differences in production for 
 the first period of growth. A comparison of the yields listed in this 
 table with those given in tables 2 and 3, indicates that it is not desirable 
 to turn under the crop at this stage of growth. 
 
 That the effect of lime is evident during the second period of growth 
 is indicated in table 2. In order to decide how conclusive the results 
 were, the probable errors of the means were computed according to 
 Peter's formula,'' 
 
 ± 0.8453 
 
 V^ 
 
 in which - ( + d) denotes the sum of the deviations of every observation 
 from the mean, their signs being disregarded, and n represents the number 
 of observations. In comparing two means the probable error of the 
 gain due to a given treatment was computed according to the formula 
 
 E= vfEr' + Ea' , 
 
 in which Ei is the probable error of one mean and Eo of the other. The 
 results of these computations are given in table 8, in which the different 
 treatments are lettered A, B, and C: 
 
 TABLE 8. Statistical vStudy of Table 2 
 
 Treatment 
 
 Mean grams of 
 dry matter 
 
 Mean grams of 
 nitrogen 
 
 (A) Unlimed 
 
 (B) Slaked lime.-.. . . 
 
 (C) Ground limestone 
 
 Gain, B over A 
 
 Gain, C over A 
 
 9 21 ± .34 
 II .30 dz .68 
 II .28 ± .14 
 
 0.3216 ± .0060 
 0.4125 ± .0185 
 0.4313 ± .0022 
 
 2.09 
 2,07 
 
 76 
 
 37 
 
 0.0909 d= 
 o. 1097 =fc 
 
 .0194 
 .0064 
 
 «A probable error based on two determinations does not necessarily represent strictly the actual 
 frequency diagram, and this must be borne in mind in considering many of the probable errors computed 
 in this bulletin. A discussion of the probable error when the number of observations is small is given 
 in an article in Biometrika, volume 6, entitled The Probable Error of a Mean. 
 
Decomposition of Sweet Clover as a Green Manure 143 
 
 Assuming that odds of 30 to i represent a chance which is practical 
 certainty, a gain must be 3.12 times its probable error to make certain 
 its significance. On this basis it is seen that the gains calculated in 
 table 8 are significant, except that for dry matter from the slaked-lime 
 pots in comparison with the unlimed pots. Here the gain falls just 
 below the value postulated to denote certainty. It is believed, however, 
 that table 8, taken as a whole, shows that the effect of lime was evident 
 during the second period of growth. 
 
 In considering the results obtained for four months growth, tables 
 3 and 6 are available. Thus four yields for a given treatment are 
 averaged to obtain the mean. In computing the probable errors of 
 these means the same formulas were used as for table 8. The results 
 are given in table 9: 
 
 TABLE 9. Statistical Study of Tables 3 and 6 
 
 Treatment 
 
 Mean grams of 
 nitrogen 
 
 (A) Unlimed 
 
 (B) Slaked lime 
 
 (C) Ground limestone 
 
 Gain, B over A 
 
 Gain, C over A 
 
 Gain, B over C 
 
 o 5743 ± .0324 
 091 12 ± 0462 
 0,8353 ± .0148 
 
 0.3369 ± .0566 
 o 2610 ± 0362 
 o 0759 ± .0484 
 
 A perusal of table g shows that the gains due to lime are significant 
 but when the two kinds of lime are compared the difference is not 
 sufficient to allow of any conclusion. The limed crop produced, roughh', 
 50 per cent more dry matter and nitrogen than did the unlimed. 
 
 Percentage of fiber 
 
 A study of the tables does not reveal any differences in fiber content 
 due to different methods of treatment. It is indicated, however, that 
 
 TABLE 10. ST.A.TISTICAL Study of Crude-Fiber Content 
 
 Duration of growth 
 (months) 
 
 2 
 
 3 
 
 4 
 
 Gain, 3 months over 2 months 
 Gain, 4 months over 2 months 
 Gain, 3 months over 4 months 
 
 Mean percentage 
 of crude fiber 
 21 .28 zb .32 
 
 24.79 ± .:i2 
 23.36 ± .25 
 
 3.51 ± -45 
 2 . 08 ± .41 
 
 1 43 ± 41 
 
144 
 
 Bulletin 394 
 
 there is a difference due to time of harvest. Therefore, the averages 
 of the results obtained for the different periods are Hsted, with their 
 probable errors computed according to Peter's formula. For the four- 
 months period the results of both years are included. The figures are 
 given in table 10. 
 
 From this table it is evident that the fiber content was least at the 
 end of the first period of growth. There was also a small decrease during 
 the last period as compared with the second. While this may not be a 
 general characteristic, that it did obtain under the conditions of these 
 experiments was indicated by cultural observations as well as by analysis. 
 It may be concluded that from the standpoint of fiber content there is 
 no disadvantage in allowing sweet clover to grow for the four-months 
 period before turning it under for green manure. 
 
 Rate of decay 
 
 The variations in the values tabulated as a result of the nitrate deter- 
 minations make evident the necessity of computing probable errors before 
 discussing these values. A statistical study of nitrate formation, made 
 up from the data listed in tables 4, 5, and 7, is given in table 11. For 
 
 TABLE II. Statistical Study of Nitrate Formation 
 
 Treatment 
 
 Unlimed 
 
 Unlimed 
 
 Slaked lime 
 
 Slaked lime 
 
 Ground limestone 
 Ground limestone 
 
 Unlimed 
 
 Unlimed 
 
 Slaked lime 
 
 Slaked lime 
 
 Ground limestone 
 Ground limestone 
 
 Unlimed 
 
 Unlimed 
 
 Slaked lime 
 
 Slaked lime 
 
 Ground limestone 
 Ground limestone 
 
 Duration 
 
 of growth 
 
 (days) 
 
 118 
 118 
 118 
 118 
 118 
 118 
 116 
 116 
 116 
 116 
 116 
 116 
 
 Treatment 
 
 of crop at 
 
 harvest 
 
 Removed .... 
 Turned under 
 Removed .... 
 Turned under 
 Removed .... 
 Turned under 
 Removed .... 
 Turned under 
 Removed .... 
 Turned under 
 Removed .... 
 Turned under 
 Removed. . . . 
 Turned under 
 Removed .... 
 Turned under 
 Removed .... 
 Turned under 
 
 Mean gain 
 in nitrates in 
 four months 
 
 None I 
 
 None / 
 
 None \ 
 
 74 ± 12.3 / 
 
 51 ± 8.0 1 
 
 1 60 ± 1 6 . 1 j 
 
 13 ± 8.0 1 
 
 85 ± 24 . I f 
 
 «3 ± 2.5 \ 
 
 230 ± 38 . 5 i 
 
 54 ± I o . 6 1 
 
 274 ±25.8 I 
 
 16 ± 0.8 \ 
 
 ?,2 ± 8.9 j 
 
 ,S7 ± 5-5 \ 
 244 ±25.8) 
 
 45 ± 0.8 
 226 ± 5.5 
 
 Net gain 
 due to crop 
 turned under 
 
 None 
 
 74 ± 12.3 
 109 ±18.0 
 
 72 ±25.4 
 147 ±38.6 
 220 ±27.9 
 
 16 ± 8.9 
 207 ±26.4 
 181 ± 5.6 
 
 a given pot the nitrates at harvest were subtracted from those found after 
 four months. The values obtained for each pair of pots treated alike 
 at harvest were averaged, resulting in the column headed Mean gain in 
 nitrates in Jour months. Next, for a given treatment as regards liming, 
 
Decomposition of Sweet Clover as a Green Manure 145 
 
 the mean gain in nitrates for the pot from which the crop was removed 
 was subtracted from the value for the pot in which the material was 
 turned under. This calculation gave the net gain in nitrates due to the 
 green manure added. These figures appear in the last colimin of the 
 table. The probable errors were computed according to Peter's formula. 
 
 In the unlimed pots of the three-months period, no nitrates were 
 formed following harvest, but on the contrary denitrification occurred. 
 For the longer periods, nitrate formation took place to some extent in 
 the unlimed pots both in 1914 and in 19 16, but the probable errors 
 attached to the values for net gain are so large as to render them of 
 doubtful significance. The favorable effect of lime is indicated in the 
 table, but no significant differences appear for the two kinds. The 
 figures indicate also that a larger net gain in nitrates is obtained by 
 growing the crop for the longer period. 
 
 The percentage of the nitrogen in the plant material added, which was 
 changed to nitrates during the four-months period following harvest, 
 is shown in table 12. The figures for grams of nitrogen added are taken 
 
 TABLE 12. Percentage of Nitrogen Added in Crop, Ch.\nged to Nitrate in 
 
 Four Months 
 
 Treatment 
 
 Unlimed 
 
 Slaked lime 
 
 Ground limestone 
 
 Unlimed 
 
 Slaked lime 
 
 Ground limestone 
 
 Unlimed 
 
 Slaked lime 
 
 Ground limestone 
 
 Duration; Nitrogeji added 
 
 of in plant 
 
 growth material 
 
 (days) (grams) 
 
 89 
 
 118 
 118 
 118 
 116 
 116 
 116 
 
 o 3216 ± 
 0.4125 ± 
 o 4313 ± 
 
 0.6408 zb 
 
 0.8166 ± 
 
 0.8098 ± 
 
 0.5078 ± 
 
 I . 0059 ± 
 
 0.8608 ± 
 
 . 0060 
 .0185 
 .0022 
 
 0006 
 .0241 
 .0066 
 
 OIOI 
 
 .0656 
 
 .0117 
 
 Net increase 
 
 in nitrate 
 
 nitrogen 
 
 (grams) 
 
 None 
 
 0.18839 d= 
 0.28065 ± 
 0.18551 ± 
 0.37749 ± 
 0.56685 ± 
 o- 03557 ± 
 0.45060 ± 
 0.3821 1 ± 
 
 .03141 
 04603 
 . 06508 
 . 09876 
 .07142 
 02233 
 
 07559 
 . 02560 
 
 Percentage 
 nitrified 
 
 None 
 45 67 
 
 7.89 
 
 07 ± 10.68 
 95 ± 10. 16 
 23 ± 12 17 
 
 00 ± 8 84 
 
 00 zt 4.40 
 
 80 ± 8 . 06 
 39 ± 3 04 
 
 from the tables in which the yields are recorded. It is asstmied that 
 these figures, obtained b>' analysis of the crops remo\^ed, should approxi- 
 mate the nitrogen added in the material turned under in the duplicate 
 pots. The values for net increase in nitrate nitrogen were obtained from 
 those showiiTig the net gain in nitrates in parts per million due to the 
 crop turned under. The later values were changed to grams of nitrogen 
 by taking account of the amotmt of soil present in each pot. It is 
 realized, of course, that all of this nitrogen may not have been fonned 
 from the ]3lant material added. However, the figures do represent the 
 net increase in nitrate nitrogen as a result of turning under the crops. 
 
146 Bulletin 394 
 
 and thus it seems proper to use them in computing the value for percentage 
 nitrified. The probable error of a given value for the latter was obtained 
 according to the formula given by Mellor/ 
 
 \(x)+ b^ 
 
 E = ± ^^\ 
 
 in which A is the divisor with a probable error a, and B is the dividend 
 with a probable error b. 
 
 It appears unwise to make any very definite statements regarding 
 the data in table 12, both because of the computations involved in con- 
 structing the table and because of the large probable error attached to 
 many of the values. It is believed, however, that the table is useful 
 in showing the approximate amount of nitrification resulting from turning 
 under the various crops. The values for percentage nitrified for the 
 unlimed pots vary widely, and nothing can be said about them except 
 that they show that the added material nitrified less rapidly than in the 
 limed pots. There is no consistent difference shown by the two forms 
 of lime, neither is there any evidence to show that the crop grown for 
 the shorter period nitrified more rapidly. For the three periods of 
 growth the figures showing the percentage nitrified for the crop treated 
 with slaked lime agree very closely; on the other hand, the figures for 
 the pots receiving ground limestone vary rather widely. As a rough 
 figure to indicate the percentage of the added nitrogen nitrified in the 
 limed pots, 50 per cent might be chosen. 
 
 SUMMARY 
 
 In an investigation such as the one described in* this paper, repetitions 
 of the experiments under field conditions are desirable before general 
 conclusions are drawn. The results obtained in the present study must 
 be interpreted with respect to the experimental conditions maintained. 
 This fact should be borne in mind in considering the summary given. 
 
 These experiments show that sweet clover will make a satisfactory 
 growth for use as a green manure in three or four months on a worn-out 
 soil, provided the lime requirement is satisfied. When the crop is harv^ested 
 at either of these periods it compares favorably in nitrogen content with 
 other legumes, and sufficient fiber has not developed to inhibit rapid 
 decay. Growing the crop for the longer period does not result in an 
 increased proportion of fiber. The plant responds readily to inoculation 
 with the appropriate organism. To secure a good stand, the seedbed 
 
 ' Mellor, J. W. Higher mathematics for students nf chemistry and physics, p. 529. 1905. 
 
Decomposition of vSweet Clover as a Green Manure 147 
 
 should be compact and treated seed should be used. The use of treated 
 seed is important also from the standpoint of economy. Treatin<,' the 
 seed with acid increases the percentaj^je of gemiination threefold, and 
 seed so treated does not lose its increased germinating jjower for at least 
 ten months. Satisfying the lime requirement of the soil was found to 
 increase the yield 50 per cent for the crop grown four months. 
 
 A crop of sweet clover grown for three or four months decays rapidly 
 when used as green manure. It was found that in the limed pots sufficient 
 nitrates had been produced four months after harvest to account for 
 approximately 50 per cent of the nitrogen added in the material turned 
 under. From the standi)oint of the amount of available plant food, 
 it is desirable that sweet clox^er, to be used as a green manure, should be 
 grown for at least four months. 
 
 The measurement of nitrate fonnation in pot experiments is subject 
 to a large probable error. This fact is a real objection to the method as 
 a quantitative measure of rate of decay. 
 
148 Bulletin 394 
 
 BIBLIOGRAPHY 
 
 Allen, E. R., and Bonazzi, A. On nitrification. Ohio Agr. Exp. Sta. 
 
 Technical bul. 7:1-42. 1915. 
 Brown, P. E., and Allison, F. E. Influence of humus forming materials 
 
 of different nitrogen-carbon ratios on bacterial activities. Iowa Agr. 
 
 Exp. Sta. Research bul. 36:1-30. 1916. 
 Cameron, F. K., and Breazeale, J. F. The organic matter in soils 
 
 and subsoils. Amer. Chem. Soc. Journ. 26:29-45. 1904. 
 Carr, M. Earl. A preliminary report on the Volusia soils, their prob- 
 lems and management. U. S. Bur. Soils. Bul. 60:1-22. 1909. 
 Praps, G. S. Studies in nitrification. Amer. chem. journ. 29:225-241. 
 
 1903. 
 Hill, Harry H. The effect of green manuring on soil nitrates under 
 
 greenhouse conditions. Virginia Agr. Exp. Sta. Technical bul. 6:119^ 
 ^ 153- 1915- 
 Hopkins, Cyril G. Soil fertilit}^ and permanent agriculture, p. 1-653. 
 
 1910. 
 Hutchinson, C. M., and Milligan, S. Green manuring experiment, 
 
 1912-13. India Agr. Research Inst., Pusa. Bul. 40:1-31. 1914. 
 
 Jones, C. H. Activity of organic nitrogen as measured by the alkaline 
 
 permanganate method. Journ. Indus, and eng. chem. 4:438-441. 
 
 1912. 
 Jones, C. H., and Anderson, G. F. A procedure for separating organic 
 
 ammoniates from the mineral portion of commercial fertilizers. Journ. 
 
 indus. and eng. chem. 6:580-581. 1914. 
 Lloyd, W. A. Sweet clover. Ohio Agr. Exp. Sta. Bul. 244:589-684. 
 
 1912. 
 Love, Harry H., and Leighty, Clyde E. Germination of seed as 
 
 affected by sulphuric acid treatment. Cornell Univ. Agr. Exp. Sta. 
 
 Bul. 312:293-336. 1912. 
 Lyon, T. Lyttleton, and Bizzell, James A. Some relations of certain 
 
 higher plants to the formation of nitrates in soils. Cornell Univ. Agr. 
 
 Exp. Sta. Memoir 1:1- iii. 1913. 
 Mi'NTz, A. Du role des engrais verts comme fumure azotee. Acad. 
 
 Sci. (Paris). Compt. rend. 110:972-975. 1890. 
 Orth, . Melilotus as green manure for heavy soils. Cited in Exp. 
 
 sta. rec. 5:701-702. 1894. Front Braunschw. Landw. Ztg. 60^^:160. 
 
 1S92. 
 Pagnoul, . Recherches sur I'azote assimilable et sur ses trans- 
 formations dans la terre arable. Acad. Sci. (Paris). Compt. rend. 
 
 120:812-815. 1895. 
 Paterson, John W., and Scott, P. R. Influence of certain soil con- 
 stituents upon nitrification. Victoria Dept. Agr. Journ. 10:393-400. 
 
 1912. 
 
Decomposition of Sweet Clover as a Green Manure 149 
 
 Veitch, F. p. The estimation of soil acidity and the lime re(|uirement 
 of soils. Amer. Chem. Soc. Journ. 24:1120-1128. 1902. 
 
 Westgate, J. M., and Vinall, H. N. Sweet clover. U. S. Dept. Agr. 
 Farmers' bul. 485:1-39. 1912. 
 
 WoLLNY, E. Ueber die Thati^^keit niedriger Organismen in der Ackererde. 
 
 Centbl. Agr. Chem. 13:796-814. 1884. (Abstracted from Deut. landw- 
 
 Presse. vol. 10, 1883, and vol. 11, 1884.) 
 Untersuchungen uber die Zersetzung der organischen 
 
 Substanzen. Journ. Landw. 34:213-320. 1886. 
 Wright, R. Claude. The influence of certain organic materials upon 
 
 the transformation of soil nitrogen. Amer. Soc. Agron. Journ. 7: 193- 
 
 208. 1915.