"ijiiY:':.: ;.,i^ .■,-v'^:m'--"''-^^ UC-NRLF ^B b^ 5fll oA-^C—^ V "^^Z-^rL^c-^^^ On the Behavior of Nitrate in Paddy Soils. BY G. DAIKUHARA AND T. IMASEKI. A great many manuring experiments carried out by different authors prove that under certain conditions nitrate nitrogen is superior to ammonia nitrogen, while under other conditions the reverse is true. It is to be regretted, however, that the different conditions, as nature of soil, its chemical composition and the character of the manure whether acidic or alkaline or neutral, have not always been stated, since this knowledge would have assisted much in recognising certain regularities. For those tropical and subtropical countries, in which agricultural crops are raised in paddy soils or swamps, the question is also of vital importarxe, which of the two sources of nitrogen would be the more favor- able. The conditions in the paddy soil for nitrification and denitrification are very different from those in the dry land soil. In the first place, the transformation of ammonia into nitrite by the nitroso-bacterium is very much depressed not only on account of less air penetrating into the soil, but also on account of much organic matter, which is by no means oxidized as easily as in dry land, and which according to Winograd^ky, depresses the action of the nitrite-microbium ; but also for the nitrate-microbium the conditions are very unfavorable, since its action is very much depressed by traces of ammonia, Loehnis^^ however declared that only in the form of carbonate, ammonia is injurious for the nitrate-microbium not in the form of neutral salts. The paddy soil for rice culture in Japan has always been manured chiefly with excrements, fish manure, oil cakes, and frequently green manure ; superphosphate was applied occasionally as a mere supplement to i). Chem. Ztg. Repertorium 19:5, I, 5. 420044 \2.\\6\'ds() S(P^^ 3 G. Dalktiliara and T. Imaseki. those manures. These conditions being favorable for denitrificatlon might lead to much loss, if nitrate were applied as a part of the nitrogen manure. Until recently however no experiments with plants cultivated in swamps have been made. It was Prof. M. Nagaoka^^ who first carried out a number of careful experiments with plants cultivated by Japanese farmers in swamps as viz. rice, arrowhead {Sag-itiaria sagittaefolia) and Jiincus cffusiis. Sagittaria bulbs are used as food and Jiincus is used for the manufacture of rugs. The Plants manured with nitrate remained pale i:i color and small in size. With rice the yield of the nitrate plants was only in one case a little above that of the control plants, in most cases, however, far below ; in one case, where the soil received no lime and the ratio of 150 Kg N p. ha., the yield of the nitrate plants fell to about 2^ of that of ammonia plants. With Juncus the yield of the nitrate plants varied from 4,7-665^ ot the ammonia plants provided with equal doses of nitrogen. For Sagittaria the result was still more unfavorable. It is very remarkable that with Juncus the harvest of the nitrate phnts diminished gradually with the increased applications of the nitrate. Some experiments^^ carried out by one of us (Daikuhara) with rice plants in sand culture showed that the availability of nitrate is 42% of that of ammonia nitrogen. The average ratio for the manurial value of ammonia nitrogen to that of nitrate nitrogen calculated from different results of experiments carried out in our central and branch Stations during the last few years is 100 : 47. As to the reason why the paddy plants cannot utilize nitrate nitrogen so well as ammonia nitrogen, Nagaoka^* has proposed the two following factors : i). Bull. College of Agric, Imp. University, Tokyo, Vol. VI., No. 3. 2). This Bulletin Vol. I., No. r. 3). Nagaoka has also ascribed the pale yellowish color of the nitrate p'ants to the physiological influence of accumulated nitrate, but according to our observations the pale yellowish color appears in the first period of growth, 2-3 weeks after the application of nitrate, and recovers afterwards. It is very probable, therefore, due to the poisonous action of nitrite formed by the reducing action of certain bicteria. On tlie Behavior of Nitrate in Pad ly S^i'-s. q 1. Paddy plants do not accumulate a sufficient quantity of sugar in the leaves to convert all of the nitric acid absorbed into protein. 2. In paddy soils, denitrification and also formation of poisonous nitrites may take place. As to the first we have instituted some tests. It might be assumed that an injurious accumulation of nitrate may take place in the leaves of the rice plants manured with nitrate but as to this there is no positive proof yet. But the result of our analysis shows practically no difference not only between the leaves of paddy and dry land rice, but also between those of the paddy rice applied with ammonium sulphite and sodium nitrate as shown in the following tables : A. Sugar content in the leaves of the paddy and dry land rice. Kind of Name o( Period of Gluccse % Cane sugar Total sugar rice plant. varieties. Growth. % % ( Before flowering 1. 19 0.48 1.67 Shin-shu ...) Milky repening 2.04 0.74 2.78 I Yellow ripening 1.94 0.24 2.18 Paddy rice . Young leaves 0.89 0.48 1-37 \ Before flowering 1. 21 0.64 1.85 Suga-ippon./ Milky ripening 2.48 1-15 3.63 ^ Yellow ripening 2.47 0.55 3.02 ( Before flowering 1.19 0.56 1-75 Oiran ...3 Milky ripening 2.02 0.85 2.87 ( Yellow ripening 2.16 0.56 2.72 Dry land nee / Youn^ leaves 0.S9 0.52 1.41 Kiushu ...J Before flowering 1.29 069 1.98 j Milky ripening 2.00 1. 10 3.10 ' Ydlow ripening 2-59 0.46 3-05 lO G. Oaikuhai'u an<1 T. Iinacekl. B. Sugar content in the leaves of paddy rice with different manures. Kind of Manures. Period of Growth. Glucose % Cane sugar % Total sugar % ( Young leaves 0-99 0-33 1.32 Ammonium \ sulphate ^ Before flowering 1.60 0 56 2 i5 ( Milky ripening 1.56 0.66 2.22 ( Young leaves 0.87 047 1-3+ Sod'um j Nitrate \ Before flowering 1-35 0.50 1.85 I Milky ripening 1.36 075 2.11 Thus we see that the first assumption by Nagaoka was not justified. Some tests were made by Nagaoka with regard to the second cause^ rendering however further observations desirable and we have therefore further studied the behavior of swamp plants towards nitrate, being certainly of considerable theoritical and practical importance. With the denitrifying organisms in a wide sense the following groups may be distinguished after Jensen'^ : " I. Reduktion von Nitraten zu Nitriten und Ammoniak (Salpeter- reduktion). 2. Reduktion von Nitraten zu Nitriten und niederiegeren, gasformigen StickstofT-Sauerstoffverbindungen (N^O und NO). 3. Reduction von Nitraten und Nitriten unter Abspaltung elementaren Stickstoffes (Denitrifikation im engeren und eigentlichen Sinne). 4. Umb'ldung des SalpeterstickstofTes in organische Verbindungen (Salpeterassimilation). 5. Freiwerden von Stickstoff bei der Faeulnis organischer Stickstoff verbindungen." Denitrification will proceed much more energetically in moist land than in dry land and also nitrates will be reduced to nitrites more i). Lafer, Handb. d. Techn. mycologie Band III., S. 182. On Ike Behavior uf Xitrate in Paddy SoV. ij energetically by peculiar bacteria of the paddy soil, and these nitrites may- persist for a time before they are further reduced to ammonia. Certaini bacteria reduce nitrates very quickly to ammonia so that the intermediary^ step, the formation of nitrites, can not be recognized. Certain other bacteria,, however, produce considerable quantities of nitrites from nitrates. Nitrites however show a strong poisonous character, since ihey act very powerfully^ on the amido-groups and hence will also attack the amido-groups of proteins, and thus kill the living protoplasm. In order to decide whether denitrification takes place extensively in. paddy soil the following experiments were made : LABORATORY EXPERBIENT. I.^> One hundred g of soil were placed in an Erlenmeyer's flask of 300 cc capacity and added 1 50 cc. of a solution containing 2 g of sodium nitrate and 0,2 g of K2HPO4 which liquid kept the soil covered in a thin layer. A humy soil from our experimental field, a sandy loam soil from Kinai Branch Station and a heavy clayey soil from Kaga Province served for this experi- ment. The experiment was started on March 23, 1905 ; the flasks were shaken every day and tested from time to time for nitrite by the reaction of Griess. The room temperature during this experiment was i6-26°C. After 24 hours a very decisive reaction for nitrite was already observed witb the humus soil, much less than in the sandy soil, and only a trace was found in the clayey soil. After 5 resp. 6 days, the maximum point of formation of nitrite (0,003 resp. 0,002% KNOJ-^ was attained in the humy and the sandy soil, whereupon the nitrite decreased, disappearing after 3 resp. 4 weeks, much nitrate still remaining in the solution. In the clayey soil, however, the reduction of nitrate was very slow and the maximum point was reached (0,01% nitrite) after 4 weeks; a little nitrite was still present even after 5 months. From the start of the experiment ammonia was occasionally tested for and reactions obtained in all of the i). Laboratory experiment I, II, III, IV, V and VI (d) were carried out by G. Daikuhara, and. Laboratory expt. VI (a), (b) and (c) by T. Imastki. 2). By colorimetric determination. 12 («■ Daikuhara and T. Iinaaekl. flasks, especially with the sandy soil, somewhat less so with the clayey soil and still weaker in the humus soil. With the humus and sandy soils a little sodium acetate was added to accelerate the bacterial action, after the nitrite reaction had disappeared. Thus again a gradual increase of the nitrite was observed and finally it reached a concentration of 0,1% resp. £),o$% nitrite 12 days after that addition, much foam being developed. No -ammonia was observed in this case. Further to ascertain whether the reduction of nitrate wis due to bacteria, the mixture of soil and the nitrate solution mentioned above, was sterilized with steam, with chloroform and with mercuric chloride. Oc- •casional testing for nitrite showed the absense of nitrite even after several weeks. This proves clearly that the reduction of nitrate in soils in the ipaddy state certainly is due to the action of denitrifying organisms. LABORATORY EXPERIMENT. II. Since in Japan much organic manure is commonly used in the paddy field, it is of vital importance to observe the influence of organic matter upon the reduction of nitrate in the paddy state. 150 cc. of the following nitrate solution was mixed with 100 g of air dry soil and kept at a temperature of 30-3 2°C. It was found that all the nitrate was reduced •entirely after 48 hours. Na-acetate (or glycerine) 0.5 % NaNOs 0.2 % K2HPO4 0.2 % MgS04 + 7H20 .' ... 0.02% On April 15, 1905, 100 g of the three different soils above mentioned were mixed with the same nitrate solution containing glycerine and kept at 'room temperature (r3-26''C). The occasional tests for nitrite and ammonia ^ave the following results : On the Behavior of Nitrate in Paddy Soil. 1$ (a). Test for nitrite.^^ From the start. Humy clay soil. Sandy loam soil. Clay soil. After 3 days 0.0013?^ KNO2 o.cx)4?^ KNOa 0.002% KNa2 ,, 4 .. 0.00105^ „ 0.005 % .1 0.003 % ., 6 „ 0.0033?^ „ 0.0055^ 0.007 % » 9 ., trace trace o.coo85^ „ ,, 12 „ 0 0 trace (b). Test for ammonia. From the start. Humy clay soil. Sandy loam soil. Clay soil. After 3 days little moderate moderate. ., 4 ., .. >i » „ 6 „ »» .. >» ,. 9 .. »» " .. ,. 12 ,, trace trace little The above result shows that by the presence of a certain amount of organic matter the reduction of nitrate in the paddy soil is much accelerated and all the nitrate added (0,2%) reduced entirely after 9-12 days even at room temperature in spring and after 2 days when kept at 30-3 2°C. LABORATORY EXPERIMENT. III. Former experiments carried out by various investigators show that fresh stable manure favors denitrification more than a well rotten one, but in regard to fresh and rotten oil cakes no special observations seem to have been made, therefore we have made the following experiment : (a). On Aug. 14, 1905, 100 g of soil from our Station at Nishigahara 2 g of rape cake (i) and 5 cc. of glycerine (2) and 150 cc. of the nitrate i). The amount of nitrite was also here approximately determined colorimetrically by Griess test. A standard solution of pure KNO^ served for the comparison. «4 Ci. Dtikuliaia an<1 T. Iinaseki. solution of the following composition and kept at room temperature (20-3 2°C) with occasional testing for nitrite and nitrate : NaNOg 2.C0 g'\ K,HPO, 0.20 gP°'"^^ '" MgSO, + 7H.O coagj ^5° cc. H.O. After 24 hours a stronger nitrite reaction was observed in (i) than in (2) (0,05% resp. 0,005^ nitrite as KNO^,). The maximum amount of nitrite was attained after 3 days with (r) and af:er 5 days with (2) but after 4 resp. 7 weeks all the nitrate was reduced. Thus we see that fresh rape •cake favors denitrification much more than glycerine. (b). Some rape cake and soy been cake were left to putref)' in a warm place with a moderate supply of water and after 2 months rottening, samples were dried and finely pulverized. Two g of each, fresh cakes and the same amount of dry matter of the rotten cakes were mixed respectively with 100 g of air dry soil and 150 cc. of the foUo'.ving nitrate solution, and kept at 25°C. NaNOo 1.75 g) ^, ^^ f dissolved in K2HPO4 0.50 gV ,, „_ I 150 cc. of HvO. MgSO^ + yHvO 0.05 g) ^ The tests showed that with fresh rape cake all the nitrate' was reduced after 2 weeks, but with rotten cake much nitrate was still present even after 10 weeks. With soy been cake which is poorer in carbohydrates and richer in protein the difference was much smaller and the nitrate was reduced entirely in the case of fresh and rotten cake after 10 and 12 days respectively. Thus we see that the fresh rape cakes favor denitrification much more than the rotten. LABORATORY EXPERIMENT. IV. To see the effect of inoculation of denitrifying organisms, pure cultured, into the soil in the paddy state, the following 3 organisms were selected : f« n I lie Bchaviiti- of Xitrate in Paddy Suil. j c a). Bact. denitrificans I. ] b). „ „ II. J c). „ nitrovorum. One hundred g of soil from our Experiment Station were kept in an Erlenmeyer's flask of ca. 300 cc. capacity with 100 cc. of nitrate solution as in Expt. Ill, (a). After repeated sterilization the denitrifying organisms were inoculated on Aug. 16, 1905, and kept at room temperature. Of the supernatant solutions small doses were withdrawn from time to time with a sterilized pipette, and tested for nitrate, nitrite and ammonia with the following results : (a). Bact. denitrificans I. Much nitrite was found after 24 hours and after 6 days the amount of nitrite reached the maximum (0,125^ KNO^). But even after 2 months later both nitrate and nitrite were still present. (b). Bact. denitrificans II. The nitrite reaction was observed after 2 days from the start and after 9 days this reaction showed the maximum. All the nitrate and nitrite were reduced after 1 3 days from the start. (c). Bact. nitrovorum. Nitrite was observed also after two days, the maximum point of the nitrite reaction after 6 days, and the disappearance of both nitrate and nitrite after 16 days from the start. A moderate ammonia reaction was obtained in each mixture after 3 days, (a) yielded much, (b) and (c) a weak ammoniacal reaction after 5-10 days, and still a moderate reaction after 2 weeks. Further with the flask (a) much pasty film was noticed on the surface of the soil, but with flasks (b) and (c) this was not the case ; the liquids became gradually turbid and of a light reddish-brown color. LABORATORY EXPERIMENT. V. To observe the influence of starch, compost, straw and rape cake upon the reduction of nitrate in paddy soil and to determine the loss of nitrogen by denitrification the following experiment was made. On Sept. 18, 1905, 28 Erlenmeyer's flasks of ca. 300 cc. capacity were i). These two organisms were isolated from the soil in the field of this Station by S. Machida. Burri and Stutzer were the first to isolate denitrifying bacteria in pure culture. i6 G. Dalkuliara'. and T. Iinacekl. filled with 200 g of humy soil from Nishigahara containing 30.53% H^O and mixed with the following salts,^' dissolved in 250 cc. H^O, and organic manures.^^ As general nutrients 0,5 g di-potassium-phosphate and 0,05 g magnesium sulphate were added to each flask and kept at room temperature (i5-26°C). I. Ammonium sulphate-) II. Sodium nitrate , III. Ammonium nitrate {Sqdium rtitrate Starch ( Sodium nitrate ^Compost well rotten (a) sterilized fSodium nitiate (Compost well rotten (b) not sterilized {Sodium nitrate Rape cake {Sodium nitrate Straw in powder Starch X. Compost well rotttn (a) sterilized... XL Compost well rotten (t) sterilized... XII. Rape cake XIII. Straw in powder , XIV. Control VI. YII. VIII. IX. 1-25 g- 2.00 „ 1.25 ,. 6.C0 „ 1-25 ., 6.00 „ 1.25 ,. 2.00 „ 1.25 ,. 6.00 „ 2.00 „ 6.00 „ 6.03 ,, 2.CO „ 6.00 „ The experiment was carried out in duplicate. While one series served for occasional testing the other served for the final determination. It was observed that flasks containing starch, rape cake and straw showed the strongest reaction for nitrate, and with the starch flask after 10 days from start, i). Nitrogenous manures subjected to this experiment contained the following percentage of N : a) Ammonium sulphate 20.50 % N b) Ammonium nitrate .. c) Sodium nitrate f ) Rape cake e) Compost we' 1 rotten .. f) Straw... 2). One g of precipitated magr.efiun 34-80 „ „ 15-55 .. M 5-7° .... 1.41 .. .. 0.57 „ „ c; rbonate was added to neutralize the acidity of sulfuric acid. On tlie Behavior of Nitrate in Paddy Soil. 17 with the rape cake flask after two weeks and with the straw flask after four weeks, all the nitrate aud nitrite were reduced. In the other flasks^ however, to which the nitrate was added much nitrate was still present after 6 weeks and in the ammonium sulphate flask also strong reactions for nitrate and nitrite were noticed after three weeks. With Compost flasks (a) and (b) moderately strong nitrite reactions were observed after 2-3 days from the start but the reaction decreased gradually. The flasks for the final determinations were kept for 45 days with frequent shaking until the nitrite reaction had entirely disappeared. 100 cc. of supernatent clear solution were withdrawn from each flask to determine the soluble nitrogenous compounds, and the remainder in the flask was evaporated to dryness after adding a little sulphuric acid. The total nitrogen^ nitric and ammoniacal nitrogen were determined as usual with the following^ results : i No. of Total air dry soil. In 100 parts of air dry soil. [n 100 part of the supernatent solution Total amount of Flasks. Total N. A'nionia N. Nitric N. A'monia N. Nitric N. Total N. g A'monia N. g Nitric N. g I. 144.4 0.456 0098 0.005 0043 0.C04 0.705 0.184 0.0 1 1 11. 147.9 0.403 0.034 0.082 0.003 0.072 0.671 0.053 0.193 III. 147.0 0495 0.053 0.094 6.043 0.072 0.843 0.120 0.210 IV. 142.3 0.3 1 1 0.015 0 0.004 0 0.447 0.025 0 V. 153-3 0443 O.OIO 0-075 0003 0.068 0.750 0.017 0.183 VI. 151.4 0-439 0.005 0.071 0.C02 0.C68 0-735 O.OIO 0.175 VII. 147.6 0.381 0025 trace 0.C06 trace 0.568 0.043 0 VIII. 1529 0.348 0.005 0 O.COI 0 0.533 0.009 0 IX. 143.0 0.307 0.020 0 0001 0 0.440 0.030 0 X. 148.5 0.401 0.025 0 0.003 0 0.598 0.038 0 XI. 152.0 0.395 0.007 trace 0.003 trace 0.603 C.OI2 0 XII. 147-5 0-377 0.015 0 0.C05 c 0.561 0.026 0 XIII. 153-5 0.340 0.004 0 0002 0 0.524 0.007 0 XIV. 145-7 0.322 0.C08 0.024 O.OOI 0.008 0.478 0.013 0.043 i8 G, Daikiiliitfu, and T. Iinaseki. From the above figures the loss of nitrogen by the reduction of nitrate was calculated with the result shown in the foUowingr tables : No. of flasks. Kind of N- manures applied. Remaining a- mount of soluble N applied. Amount of solu- ble N applied. percentage of soluble N remaining. Loss of soluble N applied. I. (NH4)2S04 0.227 g. 0.256 88.7 % 1 1.3 % II. NaNOg 0.193 ,- 0.194 99-5 .' 0.5 ., III. NH4NO3 0.365 „ 0-435 83-9 „ 16.1 „ IV. NaN03 + starch 0.007 „ 0.194 3-6 „ 96.4 ,. V. ,, + compost a) 0.152 „ 0.194 78.3 ,, 21-7 ., VI. „ + ,. b) 0.135 .. 0.194 68.6 „ 31-4 .. VII. NaN03 + rape cake 0.007 1' 0.194 3-6 „ 96.4 „ VIII. „ + straw 0.009 " 0.194 4-6 „ 95-4 .. This result confirms the observations made by occasional testing as above stated and clearly shows that the starch, rape cake and straw very much favor denitrification in the paddy soil, while well rotten compost especially when sterilized has far less influence upon denitrification. But little loss of nitrogen took place in flask II to which nitrate alone was applied, perhaps by the accidental absense of the denitrifying microbes proper. LABORATORY EXPERIMENT VI. To observe whether nitrate formation and denitrification^' take place in the dry land state, the following experiments were carried out. A. Dry land top soil. One hundred g of the following three samples of soil were placed in Erlenmeyer's flasks of ca. 300 cc. capacity with 2 g of sodium nitrate in solution and kept in a moderately dry condition, and in control flasks the same amount of distilled water was applied : i). Nitrite formation is frequency the first step of denitrification with loss of nitrogen. In certain cases however the nitiite may be completely reduced to ammonia. Oil the Behavior of Nitrate in Paddy Soil. 19 i). Humy soil of a field, from Nishigahara, 2). Clay soil from Kaga province. . 3). Alluvial sandy loam soil from Arakawa, near Tokyo. The experiment was begun April 11, 1905, and lasted 50 days, every day a portion of the contents of each flask was withdrawn to test for nitrite. The room temperature was i7-25°C. During 12 days from the start slight reaction for nitrite was observed in the soil (i) and (2) and then the reaction disappeared. In the soil (3) no trace of nitrite was found through the whole period. Thus we see that little or no nitrite formation takes place in the top soil of the dry land state, especially in the sandy soil and when nitrate alone was applied. B. Top and Sub-soil in dry land state. The same soils as in Experiment (A) just described served also for this experiment. They yielded the following chemical and physical data : ; (i) Humy soil. (2) Clay soil. (3) Sandy loam soil. Hygroscopic water 12.53 1.81 2.90 Loss by ignition 16.23 4.09 5-" Nitrogen 0.35 0.22 0.19 Wdghtof fin loose state... 60.40 93-5° 98.25 100 cc. . »^in compact .state. 105-55 146.50 150.15 . , ^„ . fin loose state ... Absorptive I 106.67 49-77 48.62 rower for H.jO 1 ■ ' -" '■m compost stste. 70.25 31.81 31.28 45O) 550 and 600 g of these soils resp. were mixed with 4 g of sodium nitrate in a moderate concentration and put in glass cylinders of ^,6 cm. diametre and 2/6 cm. height open at both ends, the upper half of the soil in a loose condition and the lower half in a compact condition, the lower end of the cylinder being closed with linen smeared with melted paraffin. Besides, the cylinders were covered with black paper, sun-light being allowed to reach only the surface of the soils. 20 G. Daikiiliara. and T. Iinaseki. The experiment was begun April 12, 1905, and every 2 days in the first period and every 5 days in the later, a portion was withdrawn from the upper and lower layers and tested for nitrate, nitrite and ammonia with the following result : (i). In the humy soil to which sodium nitrate was added a little nitrite was observed after 2 weeks, increasing a little afterwards, and the reaction was always stronger in the sub-soil than in the top-soil. The reaction for ammonia was observed after one week and was always stronger in the top-soil. In the control case, however, no nitrite was found through the whole period ; after one month a trace of ammonia and after 50 days a trace of nitrate was observed. (2). In the clay soil to which sodium nitrate was added, a little nitrite was observed after 2 days from the start, the reaction in the sub-soil increasing gradually, reaching the maximum after 2 weeks and remaining the same through the whole period, while in the top-soil no further increase was noticed. Much ammonia was found after two days from the start, the reaction of which was always stronger in the top-soil, just as with the humy soil (i). In the control case, neither nitrate nor nitrite was formed through the whole period, but a trace of ammonia was found after one month, a little more in the top-soil than in the sub-soil. (3). In the sandy loam to which sodium nitrate was added, a trace of nitrite was observed in the sub-soil after 3 days from start, and the reaction increasing gradually, reached the maximum after 2 weeks, while in the top-soil only a slight reaction for nitrite was observed after 2 weeks. With regard to ammonia, only a trace of it was found after one week which remained constant through the whole period. In the control case, neither nitrate, nor nitrite was found through the whole period and not even a trace of ammonia was found. The second experiment was carried out with the same soil and cylinders for a longer period viz. five months (June 26 — October 28, 1905), the result of which exactly coincided with that of the former experiment above stated. Thus we see that denitrification takes place more or less in the On tlie Behavior of IVitratt* in Patldy Soil. 21 sub-soil in dry land state, while in the top-soil when kept in a loose condition almost no reduction of nitrate occurs if nitrate alone is applied to the field. (C). The dry land state with application of some organic substance. In order to determine the influence of the presence of some organic matter in the soil in diy land state upon denitrification, the following experiment was made with the three soils mentioned below : i). Humy top-soil from a field of Nishigahara. 2). Clayey sub-soil „ „ „ 3). Alluvial sand soil from Arakawa. The cylinders of the diametre of 4 cm. and of the height of 55 cm. open at both ends, were filled with each soil in the air dry state, upper layer of 25 cm. deep in loose condition and lower layer of 30 cm. deep in compact condition. The soils of the first series were mixed with a solution of sodium nitrate and those for the second series with the same nitrate solution and starch, the ratio of both the nitrate and starch to the soil being i : 100. The soils were kept moderately moist and the lower open end of the cylinder was covered with filter paper and linnen and kept air tight with melted paraffin. The sides of the cylinders were covered with black paper. The experiment was begun Oct. 5, 1905, and was finished Nov. 24, during the first period of which every two days, and later on every 10 days, a portion of the upper and lower layers of the soil was taken out by boring and tested for nitrate, nitrite and ammonia with the following result : (i). In the humy top-soil to which both the nitrate and starch were applied, much nitrate was present in the upper layer during the whole period, while in the lower layer the nitrate decreased gradually and after 20 days only a trace of it was observed. The soil to which nitrate alone was applied showed a strong reaction for nitrate in both upper and lower layers until the end of the experiment. As to nitrite a trace of it was found in the lower layer during the first period in all cylinders which received nitrate and the reaction disappeared after 12 days in the cylinder to which the nitrate solution alone was applied, while in the cylinder to which both nitrate and 22 ^- Dalkuliara niid T. Iinaseki. starch were applied the reaction became evident gradually, reaching the maximum after lo days and then it gradually diminished to a trace after 20 days from start. (2). In the clayey sub-soil a trace of nitrite was observed in the lower la5''er of all cylinders except that of the control case, and no further increase of nitrite was noticed in the cylinder to which nitrate alone was applied, while in the cylinder with nitrate and starch the reaction increased gradually reaching the maximum after three weeks from the start and afterwards decreasing gradually but showing the reaction until the end of the experiment. (3). In the sandy soil with nitrate and starch the nitrate decreased remarkably after 10 days from the start in both upper and lower layers of the soil and after 20 days all the nitrate was reduced, while in the soil with nitrate alone, the nitrate reaction remained intense through the whole period. With nitrite there was no reaction in the control soil and the soil with nitrate alone, while in the soil with nitrate and starch a trace in the upper layer and much of it in the lower layer of soil were observed after 2-3 days from the start, reaching the maximum after 10 days and then decreasing gradual- ly again. The above results show that in the soil of the dry land state there is hardly any nitrite formation if the nitrate alone is applied. When, however, some organic matter is applied together with the nitrate some reduction will take place, especially in the sub-soil, where the access of air is insufficient, even all the nitrate here will eventually be reduced. (D). Comparison of the effect of some organic manures upon the formation of nitrite and denitrification. In order further to compare the effects of different organic matters, such as starch, rape cake, compost and straw, upon the degree of denitri- fication in the dry state of land, the following experiment was carried out : Ten glass flasks of ca. I litre capacity with greased stoppers were filled with 700 g of fine humy soil from Nishigahara containing 19,15% of hygroscopic water. 150 cc. of a solution containing 3.5 g of NaNOg, 2 g of III. IV. Oil llie Belia.viui- of Nitrate i» Pafldy Sail. 2^- K2HPO4 and 0,3 MgSOi were added to the flasks No. I-V, and the same amount of a solution containing only 2 g K2HPO4 and 0.3 g MgSO^ without nitrate to the flasks No. VI-X, after mixing the soil thoroughly with the organic matters^^ mentioned : I- NaNOs ... 3.5 g. jj fNaNOa 3-5 ., \Starch 5.0 „ fNaNOa 3 5 » (Compost, well rotten ••. 21.0 „ rNaN03 3.5 „ (^ Rape cake 5.0 „ ^ iNaNOg 3-5 - ISiraw ••• lo.o „ VI. Starch (Control (a)) 5.0,, VII. Compost well rotten (Control (b)) ■21.0,, VIII. Rape cake (Control (c)) 5.0 ., IX. Straw (Control (d)) lo.o „ X (Control (e)) — The experiment was begun Sept. 5, 1905, and every 10 days a certain portion was withdrawn from each vessel, the same amount of water added and tested for nitrite in the filtrate. After 10 days much nitrite was observed in the flasks II and IV, the amount of which as potassium salt approximately determined 0.35% resp. 0.088% KNO.j. In flask III only a little nitrite and in all other flasks (except II and IV) only a trace or none of it was observed. After one month in all the flasks only traces of nitrite could be found. The result shows that in top-soil of " dry " land only certain organic matters, such as starch or rape cake, favor nitrite formation to some extent, but well rotten compost and straw do not.-^ Increase of moisture will lead to an increase of nitrite and loss of nitrogen ; water was added after 33 days from the start at the ratio of 10 cc. i). The contents of N in the nitrate and organic manures are the same as in Lab. Expt. V. 2). Stoklasa (Z. Landw, Vers. Wes. in Oesterreich, 1906, p. 844.) observed in Austrian soils serving for culture of sugar beets also denitrification did not take place to such an extent that it could be proved analytically. 24 G. D.tikulvara mid T. Iinaseki. HgO to every too g original soil, thus keeping the soil in just a moderately- moist condition. After lo days nitrate reactions were obtained in the flasks II, IV and V. After two months from start the soil was dried after acidulating with a little HoSO^ and the total N, ammonia N and nitric N determined. No. of Kind of Manures. In loo pts dry soil. N in total dry soil (g). flasks. Total N. Amm. N. Nitric N. Total N. Amm, N. Nitr. N. I. NaNOs 0.452 0.G05 0.102 2.558 0.028 0-577 II. NaNOsH- starch 0-397 0.005 0.018 2-145 0028 0.102 III. NaN03 + compost 0.456 0.022 O.IOO 2.581 0.125 0.566 IV. NaN03 + rape cake 0.477 0.038 0.082 2.683 0.215 0,464 V. NaN03 + straw 0.432 0016 0.082 2.445 o.ogt 0.464 VI. starch 0.372 trace 0.006 2.105 — 0.034 VII. compost 0.382 0.005 O.OII 2.162 0.028 0.062 VIII. rape cake 0.401 0016 0.027 2.269 0,091 0.153 IX. straw 0-354 0.0 1 1 0.027 2.C03 0.062 0.153 X. 0.368 trace 0.004 2.083 — 0.023 Oil tlie Bella vioi- of \itrate iii Paddy Soil. From these fig^ures the loss of nitroo-en was calculated. 25 No. of Kind of manures. Soluble N applied (g). Loss of N by denitrification. flasks. g % I. NaNOg 0-544 o.o6g 12.7 II. NaN03 + starch 0-544 0.504 92.6 UI. NaN03 + compost 0-544 0.125 23.0 IV. NaN03 + rape cake 0,544 0.130 239 V. NaN03 + Straw 0.544 0.102 18.8 These results show that when common dry land top-soil becomes sufficiently moist, as by continuous rains, the presence of an easily putrefying organic matter favors denitrification to a considerable degree.^^ VEGETATION EXPERIMENT 1. 2) In order to ascertain the effect of the formation of nitrite and denitrifica- tion upon vegetation, the following experiments were made : Six zinc pots of ca. 30 cm. in diametre were filled with 1 5 Kg of humy soil from Nishigahara and six with alluvial soil from Arakawa. The following manures were applied per pot (in dupl |(NH,),S04 (A). JNasHFOi iKaSOi XaN03 (B). • Na2HP04 K2SO4 cate) : 5-0 g 5-0 „ 5-0 „ 6.7 „ (N in equivalent to (NH4),S04) 5-0 .. 50 .. i). A considerable degree of denitrification was obseived by Hugo Fisher after moderately liming the soil, but the amount of moisture was not stated. 2). M\ vegetation experiments were carried out by G. Daikuhara. 26 G. Uaiktilini-a nn