152-3 13-5m TEXAS AGRICULTURAL EXPERIMENT STATIONS BULLETIN NO. 156 MARCH, 1913 DIVISION OF CHEMISTRY TECHNICAL BULLETIN THE PHOSPHORUS COMPOUNDS OF COTTON SEED MEAL AND WHEAT BRAN BY J. B. RATHER, Assistant Chemist POSTOFFICE COLLEGE STATION, BRAZOS COUNTY, TEXAS AUSTIN, TEXAS Von BoEcKMANN-JoNEs Co. I913 TEXAS AGRICULTURAL EXPERIMENT STATIONS. GOVERNING BOARD. (Board of Directors, A, and M. College.) E, B. (jusnrNo, President . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..Houston JOnN I. GOION, Vice President . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..-Ballinger E. R. Astin . - - . - . . . . . - - . . - . - - . - - . . - . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. Bryan WALTON Prrrlcrrr . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . ..Fort Worth L. J. ITART . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . San Antonio J. ALLEN KYLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..Houston R. L. BENNETT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..Paris En R. KONE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..Austin rREsmENT or THE COLLEGE. R. T. NHLNER _ . . . . . . . . . . . . . . . _ . _ _ , _ . . . . . . . . . . . . . . . . . . . . . ..Col1ege Station sTATION STAFF. ll. YOONOBLOOO, U. S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Director M. FRANOIs, D. V. S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Veterinarian G. S. FRAPs, P11, D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..Chemist H. NEss, M. S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Horticulturist J. C. BURNS, B. S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Animal Husbandman WILMON NEWELL, M. S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Entomologist ' A. B. CONNER, B. S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Agronomist F. H. BLODOETT, P11. D . . . . . . . . . . . . . . . . . . . ..Plant Pathologist and Physiologist W. L, BOYETT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..State Feed Inspector HARPER DEAN, B. S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Assistant Entomologist J. B. RATHER, S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Assistant Chemist F. B. PADDOCR, B. S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Assistant Entomologist H. H. JoRsON, B. S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Assistant Agronomist ‘VILLIAM LEVIN, A. l; . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..Assistant Chemist H. G. SPAULDINO, B. S. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Assistant Chemist CHAS A, FELKER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..Chief Clerk A, S. “TARE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..Secretary J. M. SonAEoEl . . . . . . . . . . . . . . . . . . . . . . . . .~ . . . . . . . . . . . . . . . . . . . Stenographer C. A. CASE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . f . . . . . . . . . . . . . . . Stenographer R. L. SPILLEP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . uHailing Clerk STATE AGRICULTURAL EXPERIMENT STATIONS. GOVERNING BOARD. Hrs EXOELLENOY, GOVERNOR O. B. COLQUITT . . . . . . . . . . . . . . . . . . . . . . . . . ..Austin LIEUTENANT-GOVERNOR WILL ,H. MAYES . . . . . . . . . . . . . . . . . . . . . . . . . . Brownwood COMMIssIONER OF AGRICULTURE ED R. KONE . . . . . . . . . . . . . . . . . . . . . . . . . . .Austin DIRECTOR OF EXPERIMENT STATIONS. B. YoUNcRLoon, M. S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . College Station SUPERIN'1‘l:)r1s'r.* In a previous publication (Bulletin 14G, ’l‘cxas Experiment Station) we shmvcd that the phosphorus compounds of cottonseed meal were nearly all organic in nature and not inorganic, as has been claimed, and especially that cottonseed meal does not contain meta- or pyrophos- phoric acid. The work presented in this bulletin is a continuation of the above study of the forms of phosphorus compounds in cottonseed meal, together with an investigation of the acid-soluble phosphorus con - pounds of wheat bran. * REVIEW’ OF LITERATURE. Hardin states (Bulletin 8, new series, South Carolina Experiment Sta- tion) that cottonseed meal consists largely of 1neta- and pyrophosphoric acids. Crawford (J our. Phar. 8t Expt. Therapeutics 1,151) claims that “the chief poisonous principle of cottonseed meal is a salt of pyrophos- phoric acid.” It has been shown by the writer (Bulletin 146) that the product which these writers thought was inorganic phosphoric acid was of organic nature and with very similar properties to pyrophosphoric acid. In Technical Bulletin No. 25 of the New York (Geneva) Experiment Station, B. J. Anderson reports some work on meta- and pyropliosphoric acid in cottonseed meal and concludes that “The aqueous solution of the free acid gives all those reactions which have been attributed previously‘ to the presence of meta- and pyro-phosphoric acid in cottonseed meal.” Bulletin 146, of the Texas Experiment Station, published ten months before the New York Bulletin, deals primarily with the question of the presence of the above named inorganic forms of phosphorus in cottonseed meal and the conclusions were as follows: “The principal compounds containing phosphorus were separated which gave the same reactions as those relied upon forproving the presence of meta- and pyrophospho-Ijic acid- We have no evidence that the samples of cottonseed meal examined contain either pyrophosphoric acid or metaphosphoric acid.” The only difference between our work and Anderson’s on the subject of rneta- and pyrophosphoric acid was a difference in the method of prepara- tion of the free acid tested and a much more complete duplication on our part of those tests on which were based the claims that cottonseed meal contains the above named inorganic acids. Anderson further concludes that “cottonseed meal contains an organic phosphorus acid which is verv similar to phytic acid. Whether the sub- stance is identical with phvtic acid could not be determined.” Posternak (Rev. Gen. Bot. 12, 5 and 65; Oompt. Bend. 137, 3, 5, 8) *Under the general direction of G. S. Fraps, Chemist. 4 '1‘£x.~\.s AG'I{I(;‘I'L'lL‘R.-\L EXPERIMENT Srxrmxs obtained an organic [ihosphorus acid from the seeds of peas, beans, pump- kins, lupine and other plants, which gave inosite as a cleavage product when heated in a closed tube with acid. He called this acid phytic acid, or {lHi1_V(l1’0—0X}'1Tl@-il1ylGDB-(ll]')l1OS]'ll1OI‘lC 21('i(l, and PI‘OPOSG(1 the formula (f‘i,,l~I,,li’,O‘,. Winter-stein (Ber. Deut. Chem. Ges. 30, .2299) separated a phos- phorizcd motluct. from the seeds of a. black mustard, which also gave iliosite as a cleavage product. Patten and Hart (New York [Geneva] Experiment Station, Bulletin 250) studied the acid-soluble phosphorus compounds of wheat bran, and state that the principal compo-und corresponds to the formula G,H8P,O,,, and that it is probably identical with Posternak’s “phytic acid.” Hart and 'l‘ottingha1n (Wisconsin Experiment Station Research Bulle- tin 9) claim that “phytin” is present in the seeds of maize, oats and barley. Suzuki and Yoshimura (Bulletin College Agriculture [Tokyo] '7, 498) N euberg (Biochem. Ztschr. 6, e105), and Forbes (Ohio Experiment Sta-- tion Bulletin 215:) regarded the above compound as phytin, but do not report complete analyses of the product. ‘Levine (Bioc-hem. Ztscher. 16, 390) describes a semiciystalline barium salt of an organic phosphorus acid. Vorbrodt (Anzeiger. Akad. Wiss. Krakau, 1910. Series A. 414) separated an organic phosphorus compound, to the barium salt of which he ascribed the formula 012H26P11Ba17046l A number of constitutional formulas have been proposed by various investigators and several attempts have been made to synthesize “phytic acic ” in the laboratory. Rising (Svensk Kemisk Tidskrift 22, 143 [1910]) prepared a silver salt of “phytic acid” for which he proposes the formula GsHnPfiAgsOgz. Anderson (New York [Geneva] Agricultural Experiment Station Technical Bulletins 19 and 21) studied the salts of “phytic acid” ob- tained from commercial phytin. Among others, he reports the analyses of two silver salts of the acid, one of which he termed o-cta-silver phytate and the other hexa-silver phytate. Carbon and hydrogen in the silver salts were not reported. ' More recently, Anderson (New York [Geneva] Station Technical Bul- letin 22) reported a study of the acid-soluble organic phosphorus acid of wheat bran, in which he states he prepared salts which corresponded to the formulas G._,_.,H55P9Ba._,O_.,, and G,,,H,,_P9Ba5O4,,. He claims that the latter compound is formed from the former by the splitting oit of one pcntose group. He concludes that wheat bran does not contain “phytin.” PRELIMINARY ST UDIES. The first step in the identification of an unknown compound is to estab- lish its purity. Unless this is done the analvsis is of doubtful value. In the purification of the organic phosphorus acids from feeding ma- terials. it is necessary to remove, besides other substances. the inorganic bases iron and aluminum. calcium, magnesium, potassium and sodium, and the inorganic phosphoric acid which is always present to a greater or less extent. The method of purification used bv Posternak (loc. cit.) and modified by Patten and Hart deserves especial attention, because of the fact that PHOSPHORUS COMPOUNDS or COTTONSEED M I-IAL AND \V1§IEAT BRAN 5 upon analyses of products prepared by this method rests the theory that phytic acid has the formula C2HSP2O9, or a multiple thereof. The 1nodi- fication of Patten and Hart consists in the substitution of barium chloride, for calcium chloride. They offer no evidence that inorganic bases have been removed, nor do they appear to have taken into consideration the possibility of the presence of inorganic phosphorus in the final product. A product prepared from Wheat bran by the method of Patten and Hart was dried at 110° and analyzed. The results were as follows: Iron and aluminum . . . . . . . . . . . . . . . . . . . .. 3.33 per cent Calcium . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 0.30 per cent Magnesium . . . . . . . . . . . . . . . . . . . . . . . . . .. 0.18 per cent The method was followed as carefully as possible. In our previous work on this subject, we concluded from qualitative tests that lime and magnesium were practically eliminated by the method of Patten and Hart, but quantitative examinations proved that, although the amount of these bases was relatively small, the total basic impurities was sufiicient to render any formula erroneous if it were based on the analysis of the product prepared in this Way. The mixed iron and aluminum phosphates were assumed to be half and half; calculated to iron, it would be 3.13 per cent, and calculated to aluminum, 3.53 per cent. Inorganic phosphorus was determined according to the method of Forbes et al.; .32 per cent was found. This would correspond to 1.01 per cent ortho-phosphoric acid. The total impurities in this product are, therefore, about 11.7 per cent.* The oxygen in the p-roduct would accord- ingly appear 4.7 per cent too high and the carbon too- low ; the apparent percentage of phosphorus would also be too high. It is obvious that a formula. based on the analysis of such a product would be erroneous. Another product prepared in a similar manner gave the following results: ' Iron and aluminum . . . . . . . . . . . . . . . . . . .. 3.26 per cent Calcium . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 0.38 per cent Another product was prepared in a similar manner, but made alkaline with ammonia after the first hydrogen sulphide treatment to remove inorganic bases and the resulting precipitates filtered off, the excess of ammonia removed by evaporation, and the purification completed as usual. Found: Iron and aluminum . . . . . . . . . . . . . . . . . . . .. 1.85 per cent Calcium . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 0.51 per cent Another product was analyzed after a treatment similar to the above, involving the addition of ammonia in the proportion of about_10 to 1 and filtration from the resulting precipitate, the removal of ammonia and completion as usual. Found: Iron and aluminum . . . . . . . . . . . . . . . . . . . . . 1.93 per cent Calcium . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 0.50 per cent *We have assumed that the iron and aluminum were present as phosphates in the ignited product. Should this not be the case, the total in-organic impurities would be about 10 per cent. 6 "l‘I~l.\'.-\S Aunzcrixrricixr. EXPEIUAIENJ.‘ Sr.»x'r1oi\‘s Still another product contained 1.75 per cent iron and aluminum and 0.30 per cent calcilnn. It is evident, therefore, that tl1e assumption that the method of Patten and Hart removes inorganic bases completely is not justified by the facts. 'l‘hc amount present is enough in most cases to seriously inter- fere with the accuracy of the work and anv formula based on these analyses is undoubtedly xvrong. The oxygen content would appear too high by several per cent and the carbon, hvdrogen, and phosphorus con- tents too low. This effect on the twhosphorus would be offset by the in- crease due to the presence of inorganic phosphorus. It is noteworthy that although the calculated percentage of carbon in the compound CQHgPZC), or “phytic acic” of Posternak, Patten and Hart, Hart and 'l‘otting‘l1an1, and Anderson is 10.08, the American investigators invari- ably secured higher results 011 analysis, ranging from 10.22 to 10.76 per cent and averaging 10.50 per cent. Patten and Hart state that their reason for using barium chloride instead of the calcium chloride used by Posternak was that the lime was much more difficult to remove. If this is correct, then the method of Posternak is of still less value than that of Patten and Hart, which we have already shown left in the product, in one case, 4.7 per cent inor- ganic impurities. The Patten and Hart method has been used by Hart and Totting- , ham (loc. cit.) on corn and other grains. They make no mention of tests for the completeness of the removal of inorganic bases and their work is open to the same objections as that of Posternak, and Patten and Hart. Anderson (loc.»cit.), who studied a large number of the salts of phytic acid, used the method of Patten and Hart to prepare free phytic acid and based the calculated composition of these salts on the theory that phytic acid has the formula CSHMPSOW, a multiple of the O2H8P209 of Posternak. In a large number of cases he made no carbon and hydrogen determinations in these salts, and as far as we have been able to find has made no tests to see if inorganic bases were absent. We will show that the acid does not correspond to the above formula, and it is obvious that all of the above work of Anderson is open to_ the same objections as that of Posternak, Patten and Hart, and Hart and Tottingham. From a review of the literature it is evident that the theory that phy- tic acid has the formula C2HSP20£,, or a multiple thereof, rests entirely on the work of ‘Posternak, Patten and Hart, Hart and Tottingham,land Anderson. The methods used were very much alike. Rising, Verbrodt. and the writer (this bulletin) used different methods and obtained dif- ferent results. As far as we have been able to find, few of the investigators of the so-called “phytin” have‘ made any attempt whatever» to prove that the products analyzed were homogeneous. This was apparently assumed after the completion of the various methods of purification used. SEPARATION AND PURIFICATION OF THE PRINCIPAL ORGANIC PI-IOSPHORUS ACID OF COTTONSEED MEAL AND WHEAT BRAN. In our previous work on cottonseed meal we found that more than twice as much phosphorus was soluble in water as in 0.2 per cent hydro- PHOSPHORUS COMPOUNDS or COTTONSEED MEAL AND WHEAT BEAN '7 chloric acid, and that after extraction with acid the remaining phosphorus was no longer soluble in water, but could be dissolved in 0.2 per cent ammonia. The following method was designed to separate the water-soluble phos- phorus in two fractions, the acid-soluble and the ammonia-soluble after acid extraction. M eth-odr-The metho-d finally adopted for this Work was a modification of the modified Patten and Hart method used by us in previous work on the phosphorus compounds on cottonseed meal. The purpose of the precipitation with ammonia and with alcohol was to remove inorganic phosphorus and inorganic bases. It was found by analysis (see below) that this result was accomplished. The method is as follows: Two kilograms of cottonseed meal were digested with 8000 c.c. 0.2 per cent hydrochloric acid for three hours with frequent shaking. The extract was strained through cheesecloth and the residue washed well with water. The residue was digested with 8000 c.c. 0.2 per cent am- monia for three hours with frequent shaking and allowed to settle, and washed with water by decantation. Copper acetate was added to the acid extract of the cottonseed meal in sufficient quantity to precipitate most of the phosphorus compounds, the precipitate was washed well with water, decomposed with hydrogen sulphide, filtered, and evaporated to a sirupy consistency. The product was dissolved in a small amount of water and about ten times the volume of ammonia added. The mixture was allowed to stand over night. The precipitate was filtered off. The filtrate from the pre- cipitation with ammonia was evaporated on the water-bath to remove the ammonia and taken up with water. Barium chloride was added and the resulting precipitate was filtered and washed with water. The barium salts were decomposed with sulphuric acid and filtered. The filtrate was again made alkaline and precipitated with barium chloride. This process was repeated two or three times and the acid finally precipitated with copper acetate in acid solution, the copper salt decomposed with hydro- gen sulphide, filtered and evaporated to a small volume. A large volume of alcohol was then added and the resulting precip- itate filtered ofi. The alcohol was evaporated from the filtrate and the product take11 up in a small volume of water. The ammonia extract of the cottonseed meal was made acid With hydro- chloric acid. the precipitate allowed to settle and the liquid decanted through a filter. This extract was precipitated with copper acetate, ammonia, barium chloride and alcohol exactly as described for the acid extract. In the examination of the wheat bran, the ammonia extraction of the feed was omitted. Analysis of the wheat bran product, dried at 110° C, gave the follow- ing results: 0.53 gram gave .0011 gram combined iron, aluminum, lime and magnesium phosphates. i 0.10 per cent inorganic phosphorus. The above amount of phosphates would correspond to 0.08 per cent iron. It is evident that this amount of inorganic phosphorus. calcu- 8 TEXAS AGRICULTURAL EXPERIMENT STATIONS ‘ lated as silver phosphate in the silver salt subsequently analyzed, Would not change the composition of the organic silver salt beyond the limit of analytical error. ' Analysis of the products from the extractions of cottonseed meal gave the following results: .6675 gram of the product from the acid extraction gave .0020 gram combined iron, aluminum, lime and magne- sium phosphates. .3830 gra.m of the product from the ammonia extraction gave .0015 gram combined inorganic phosphates. This would correspond to 0.11 per cent and 0.111 per cent iron, respect" ively. Inorganic phosphorus was found to be 0.12 and 0.10 per cent in the products. , The zibove percentages of inorganic phosphorus was calculated from analysis of silver salts hereinafter described. The method for the prepar- ation of the acid [rroduct is therefore satisfactory, as far as the removal of inorganic impurities is concerned. IIOMOGENEITY OF THE PRODUCTS. In order to see if the products Were the same, the following experi- ments were made: ' Portions ‘of the acid products prepared as described above were made up to a volume of 200 c.c. with water, and aliquots titrated with standard potassium hydroxide and phenolphthalein. The rest of the solution was evaporated to about 50 c.c. and cooled to room temperature, an amount of standard ammonium hydroxide calculated to make a neutral ammonium salt was added to each. They were then diluted to make the concentra- tion of the salts the same, and the theoretical amount of standard silver nitrate solution required to form a neutral salt was added. The re- sulting‘ precipitates were filtered and washed with equivalent amounts of water. The washed precipitateswere then dried in air, powdered and dried in a vacuum desiccator over sulphuric acid. The filtrates from the silver precipitates were made slightly acid with nitric acid (two or three drops) and precipitated with copper acetate. The copper was removed from the washed. precipitate with hydrogen sulphide and the resulting filtrate was evaporated to asmall volume. The acids were then precipitated with silvernitrate under exactly the same conditions as to concentration, temperature, alkalinity’, etc., as before. Unfortunately the amounts of the second fractions of the silver salts secured were too small to obtain good results with, and the analyses were not considered. The silver was determined in 0.2000 grams of the dried first fractions described above, with the following results: (“ottonseed meal : Grams silver. Acid extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1224. Ammonia extraction . . . . . . . . . . . . . . . . . . . . . . . . . .123’? Wheat bran: Acid extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . .. .1206 PHosPHoaUs COMPOUNDS or Corroxsnnl) hIIZAL irxn WHEAT BRAN 9 The results agree closely and it appears probable that the products were the same, since the fact that the salts contain the same amount of silver when precipitated under the same conditions is evidence that they are identical. One-half gram of each salt, dried in a vacuum over sulphuric acid, was digested at room temperature in 200 c.c. of water for five hours with fre- quent shaking. The solutions were then filtered and the residues washed with equal volumes of water. Silver was determined in the filtrates. The results are as follows: Cottonseed meal: l Grams silver. Acid extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .0178 Ammonia extraction . . . . . . . . . . . . . . . . . . . . . . .. .0172 Wheat b-ran: Acid extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .0162 These results agree very closely, all of them lacing xvell within the limit of error. It is evident, therefore, that "the compounds are the same. The above is evidence that the compounds are pure, for it would be very improbable for them to be impure and each have the same amount of impurities and of thc sanie solubility. Further evidence that the compounds are pure, will be mentioned later. In our previous work on cottonseed meal (loc. cit.) we made a large number of tests on the products of the acid and ammonia extracts and in every case the reactions were the same. This is further proof that- the cottonseed meal products are the same. PREPARATION OF THE SILVER SALTS OF THE PRODYCTS. Aqueous solutions of the acid products prepared as already described were made alkaline with ammonia and boiled until the excess of am- monia was expelled. Silver nitrate solution was added in slight excess and the resulting precipitates washed well with water. (Fractions l.) The filtrates from the above precipitation were made slightly acid with nitric acid (two or three drops) and precipitated with copper acetate. The precipitates were washed well with water, suspended in water, decom- posed with hydrogen sulphide and the crapper sulphide filtered off. The acid filtrates were then evaporated to remove excess of hydrogen sulphide and taken up with water. The products WQTQ then precipitated with sil- ver nitrate in neutral solution exactly as described above. (Fractions 2.) The silver salts were precipitated in warm solution and an excess of silver nitrate was generally used. The portion which went into solution was thereore more than in the homogeneity tests described before, and no difiiculty was found in getting sufficient amounts of the second frac- tion to analyze. The silver salts were dried in air, powdered, dried in a vacuum desic- cator over sulphuric acid and analyzed. . Products were also prepared from the hydrochloric acid extract of cot- tonseed meal by a. method similar to the above, but no ammonia was used to precipitate inorganic phosphates, etc. The crude free acid was pre- cipitated with magnesia mixture as suggested by Forbes et al. (Ohio 1O 'I‘nx.1s Ac-ruoLrLrUnnL EXPERIMENT Srarrons Plxperinug-nt Station Bulletin 215), and washed with 0.2 per cent nitric acid in alcohol. The magnesium was removed by precipitation of the acid with copper acetate in acid solution and the subsequent purification and the preparation and the fractionation of the silver salts was the same as described above. (Mg). The silver salts prepared from the acid and ammonia extracts of cot- tonseed meal and from the acid extract of Wheat bran were found on analysis to be free from nitrogen and pentosans. We have already shown that the amount of inorganic impurities is negligible. The analyses of the silver salts and the formulas corresponding to them are shoyvn in Table 1, together with those of the salts reported in our previous work on cottonseed meal. TABLE 1. Percentage Composition of the Silver Salts of the Phosphorus Compounds. Carbon Hydro- Silver Phos- gen phorus T, q f Fraction 1, found ............ .. 4.60 0.91 61.79 9.54 '9 g 12H2‘Ag17p9042’ C 4.88 0.81 62.10 9.45 E": t Fraction 2, found.............. 4.87 0.74 59.86 10.06 1s h? I C,,H,5Ag,6P,,O4,, Cal 5.06 0.88 60.66 9.81 $5 y Fraction 1, (Mg), found. 5.46 0.94 56.15 10.73 VJ °< 1 C12H2QAgl3P9O42, Cal 5.70 1 . 11 55.57 11.04 n“ | Fraction 2, (Mg), found . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 60.37 10.00 8?, , C ,H,5Ag,8P9O,,,, Cale. 5.06 0.88 60.66 9.81 ‘as’ ' P’E‘t“fi“’1“3‘il‘"d4‘”c“f““d 26% 8'33 366% 3'3? U l 12 2s‘ E16 9 42» 3 c - - - - a m = f Fraction 1, found ............... .. 4.78 0.78 62.68 9.40 Fag t **~H““g""1?"' ca‘ $38 9%‘ 2312 1368 o ~ g5 l Eitiirééféi-cql-saiscsi c. @112 111i as 121a Q” t ( . 4 ), n o. . . . 13%? 1 rg,3I§I25Ag,Gli’9O“, C825... 5.06 0.88 60.66 9.81 *5,‘ c, F t Fraction 1, found .................. .. 5.21 0.82 61.86 9.67 ME a»; } c,.H..Ag.,P,,o_,,,ca1c. 4.88 0.81 62.10 9.45 “@361 g Fraction 2, found ............ .. 4.15 0.91 59.27 10.12 3 ~ 1 cnrifingwego“, Calc.. ......... .. 5.25 0.95 59.10 10.18 The products correspond to the formulas given in the table. It is evi- dent that the products from the acid and ammonia extracts of cottonseed meal and the acid extract of wheat bran are all salts of the same acid, but containing different quantities of silver. The latter is to be expected as no attempt was made to precipitate the salts under identical conditions, for it has already been shown that the salts have the same amount of silver when they are so precipitated. The free acid corresponds to the formula 012H41P9O42n These results are not in accord with the conclusions of Patten and Hart, who claim that wheat bran contains an inosite-phosp-horic acid cor- responding to the formula C,H8P,O,,, nor with those of Anderson, who claims that cottonseed meal contains an acid of similar composition, nor with those of Anderson, claiming that wheat ‘bran contains an inosite— phosphoric acid of the formula C2OH,5P,,O,5. In our previous work on cottonseed meal it was shown that more than twice as much phosphorus was soluble in water as in 0.2 per cent hydro- chloric acid. and that after extraction with acid the remaining phos- PHosPHoRUs CoirPorrNns 0F CoTToNsunn MEAL AND WHEAT BRAN 11 phorus was no longer soluble in water, but soluble in 0.2 per cent am- monia. The principal acid-soluble and ammonia-soluble compound have been shown to be the same acid, an inosite-phosphoric acid corresponding to the formula CI2H41PQO42Q The amount of ammonia. soluble inosite-phos- phoric acid. is then“ about twice as much as that soluble in 0.2 per cent acid. The difference in solubility of the salts as they occur i11 the meal is a point which is as yet unexplained, but on which we hope to work in the future. v _ It appears from the work on cottonseed meal that ‘the amount of inosite-phosphoric acid. present is not necessarily limited to that in the acid extract as has apparently been assumed. The behavior of cottonseed meal in regard to acid and water is similar to that of oats, malt sprouts, and wheat ‘bran, for Patten and Hart found nearly twice as much water- solublc as acid-soluble phosphorus in oats and appreciably more in the case of the last named materials. It is probable, therefore, that phytin is the principal constituent not only of that part o-f the material soluble in 0.2 per cent acid, but is the principal phosphorus compound in the entire feeding material. DECOMPOSITION OF THE PHOSPHIORUS COMPOUNDS OF‘ COTTONSEED MEAL INTO INOSITE AND PHOSPHORTC ACID. Unweighed portions of the crude free acids extracted from cottonseed meal by the acid and ammonia treatment were heated in sealed tubes for three hours at 160-65° C. After cooling, the tube-s were opened, the products removed, the sulphuric and phosphoric acids precipitated by barium hydroxide and carbon dioxide added to remove excess of barium hydroxide. The filtrates were evaporated nearly to dryness, taken up with hot water and the remaining barium carbonate filtered oif. The filtrate was dissolved in purified alcohol, and ether was added until a cloudiness was produced. The liquid was then cooled in ice water and allowed to stand. A crystalline precipitate separated in both cages. The amount of the precipitates iras small and they were not recrystallized. Dried in a vacuum over sulphuric acid, the. substances gave the fol- lmving results on combustion: Carbon. Hydrogen. Product of acid extraction . . . . . . . . . . .. 38.96 6.86 Product of ammonia extraction . . . . . . . . 40.55 7.18 Calculated for inosite . . . . . . . . . . . . . . . . .. 40.00 6.66 The products gave the reaction of Sherer and Gallois. The product of the acid extraction melted at 220° C (uncorrected) and the product of the ammonia extraction melted at 225° C (uncorrected). Inc-site melts at 218° C (uncorrected). Although the above products are impure they are undoubtedly inosite. The product from what bran was not decomposed into inosite and phosphoric acid, this already having been done by Patten and Hart on a sample prepared in a similar manner to ours. 12 'l‘|~:.\'.\s Aumouixrtraai; EXPERIMENT STATIONS IJXABIINATWON OF 'l‘.HE AMMONIA AND ALCOHOL PRECIPITATES OBTAINED IN PURIFYING THE PRODUCTS. The ammonia precipitates from the acid extract of cottonseed meal were powdered, dried in a vacuum desiccator over sulphuric acid, and analyzed. Found: Carbon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 8.24 per cent Ilydrogen . . . . . . . . . . . . . . . . . . . . . . . . . . .. 3.71 per cent The substance was rich in phosphorus and inorganic bases. It is evi- dently composed of inorganic phosphates, salts of inosite-phosphoric acid and unknown substances. Its carbon and hydrogen content agrees roughly with that of a calcium-magnesium-potassium salt of inosite-phos- phoric acid. The substance was not examined further. The alcohol precipitate obtained in the purification of the product from wheat bran was powdered, dried as above and analyzed. Found: Carbon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 9.24 per cent lluvdrogcn . . . . . . . . . . . . . . . . . . . . . . . . . . .. 3.42 per cent This product contained phosphorus and inorganic bases and is appar- ently largely organic in nature. It comprised about one-fourth of the purified wheat bran product. It evidently consists of salts of inosite phosphoric acid and other substances. No further examination was made. innrrriirv 01-‘ THE COMPOUNDS. To summarize the above work, we have the following evidence that the compounds are salts of the same acid: (l) the free acids yield inosite on decomposition with sulphuric acid; (2) the silver salts correspond to salts of the same acid; silver salts having the same amount of sil- ver are formed under identical conditions; (4) the silver salts, prepared under identical conditions, have the same solubility in water; (5) the qualitative reactions of the products from cottonseed meal are the same. The facts that the various fractions of the silver salts correspond to salts of the same acid, and that the silver salts prepared under identical conditions have the same solubility, is evidence that the products are pure. ' We are therefore justified in concluding that the inosite~phosphoric acids of wheat bran and cottonseed meal are the same, and that the free acid corresponds to the formula C,.,2H,,I’9O,2. These results are not in accord with those of Patten and Hart, and of Anderson, whose work has already been discussed on preceding pages. We have already shown that the method of Patten and Hart would give erroneous results, since we found as much as 4.7 per cent inorganic im- purities present in a product prepared by their method. Anderson’s method for the preparation of the organic phosphorus acid from wheat bran, given in detail on a following page, consists in repeated precipitation of the crude salt with alcohol from acid solution and finally with barium hydroxide. It is a. well known fact that calcium and mag‘- nesium phosphates are precipitated under the same conditions, and unless PHOSPHORUS COMPOUNDS or Corroxssun NIEAL AND WHEAT BRAN 13 the inorganic acid is present in a free state, which is not likely, this purification process would remove very little of the inorganic phosphoric acid or inorganic bases. This method of Anderson is open to a still more grave objection, which will be discussed on a following page. The method used by Anderson for the preparation of the inosite-phos- phoric acid of cottonseed meal Was similar to that used by him in his work on wheat bran, the only essential difference being the crystallization of a barium salt from neutral solution and from acid-alcohol. He ob- tained crystals of constant composition, but he apparently assumes that iron, alumnia, lime and magnesium were absent, for he offers no ana- lytical evidence concerning them. His product is therefore liable to con- tain these substances for the reasons we have already stated. THE QUESTION OF THE PRESENCE OF PENTOSE IN THE INOSITE-PHOS- PI-IORIO ACID OF YVHEAT BRAN. Anderson (loc. cit.) claims that wheat bran does not contain phytin, but that it does contain an acid, the barim salt of which corresponds to the formula G25H55O54P,,Ba,, and that this compound is easily trans- formed into a compound, the barium salt of which corresponds to the formula Q2OH1tSO4QPDBa-5, and that it loses one pentose group, CJHWOE, in so doing. Anderson digested the bran with 0.2 per cent hydrochloric acid, pre-- cipitated the strained extract with alcohol, and purified the resulting precipitate by reprecipitating six times with alcohol from 0.2 per cent hydrochloric acid; .3887 gram of the dried substance gave .0367 gram phlorglucid when distilled with 12 per cent hydrocholoric acid. Anderson dissolved portion of the above substance i11 water and pre- cipitated hot with barium hydroxide. The washed precipitate was dis- solved in 0.5 per cent hydrochloric acid and reprecipitated with barium hydroxide, the resulting precipitate dissolved by addition of dilute hydro- chloric acid, and then precipitated by the addition of a like volume of alcohol. The filtered substance, washed i11 dilute alcohol, was again dis- solved in 0.5 per cent hydrochloric acid and reprecipitated in the same manner as before. These operations were repeated four times. It was then dissolved in the same strength hydrochloric acid, pre- cipitated with alcohol, filtered, washed in dilute alcohol, alcohol, and ether, and dried in a vacuum over sulphuric acid; 1.2121 grams of this substance gave 0.0053 grams phlorglucid when distilled with 12 per cent hydrochloric acid. From analyses Anderson concludes that the for- mula. is C2SHJ5OS‘LP5Ba'-JI The above salt was suspended in water, decomposed with dilute sul- phoric acid, filtered from the barium sulphate and precipitated with copper acetate. The copper was removed by hydrogen sulphide and the filtered solutio-n concentrated, dried and analyzed. The formula O2OH5SPS,G4Q is proposed for this product. The following is quoted from the summary, page 2: “The purified bariumsalt of the compo-und corresponded to the fol- lowing formulas: C25HS5OMPDBa5 and C2,,H4,O,_,,P.,Ba5. Attempts to isolate the free acid corresponding to the first salt did not succeed. From both salts the same acid, corresponding to that of the second salt, 14 Texas AemccLrrusir. EXPERIMENT STATIONS CwHfiOwPv was obtained. 'l‘his acid is apparently formed from tl1e first by the splitting; off of the elements of one pentose. . . . The conclu- sion secms justified that wheat bran does not contain phytin.” The following‘ review of the evidence offered by Anderson will show that his conclusions are 11ot justified by the facts: According to theory; if the product to which he assigns the formula C._,5H5,O_,,P9Ba, contained one pento_se group, this salt would contain 6.87 per cent pentose. But by his analysis 1.2124 grams yielded 0.0053 grams phlorglucid, which would correspond to 0.88 per cent pentose. This is 12.8 per cent of 6.87, or only one-eighth as much as his theory requires. From the 1Jhosphorus-pentose ratio in the theoretical barium salt one can get an idea of what the lime-magnesiu1n-potassiun1 salt of the acid should contain. In the pure salts the ratio would be the same regardless of the amount of lime, magnesium, potassium or barium present. The theoretical percentage of pentose in the barium salt; is 6.87 and the per- centage of phosphorus 12.76. The amount of phosphorus found in the crude lime-magnesiumpotassium salt was 14.42. Then if X repre- sents the per cent of pentose theoretically present in this salt, 6.87: 12.76::X:14.42, whence X equals 7.76. The amount of phos- phorus in the crude salt, then, corresponds to 7.7 6 per cent pe11tose ac- cording to theory. . Anderson’s analysis shows that 0.3887 grams of substance yielded 0.0367 grams phlorglucid, vhich would correspond to 10.94 per cent pentose, yet lay theory 7.76 per cent. is present in the salt, a difference of 3.18 per cent. Il-is CTtlClfi salt, than, contains 41 per cent t00 much pen- tose and his barium salt 87 per cent too little. It is evident from Ander- son’s results only one-eighth of the barium salt can be in the form claimed by him. This matter was tested further in the following manner: Twenty-five grams of wheat bran and of cottonseed meal were digested for three hours with 250 c.c. 0.2 per cent hydrochloric acid, filtered, and 200 c.c. precipitated with alcohol, just as described in the method used by Anderson. The precipitate was filtered and washed with alcohol, washed from the filter Wltll hot water, boiled and some insoluble matter filtered off. (Anderson claims that the crude salt is readily soluble in water.) The filtrate was heated nearly to boiling and precipitated with barium hydroxide, filtered hot and washed with hot water. The phos- phorus and pentosans "were determined in the precipitate. The prepara- tion of the barium salt was exactly the same as that used by Anderson except that we precipitated the product only one time with each precip- itant. . The filtrate from the barium precipitate was freed from- barium, con- centrated and precipitated with alcohol. The precipitate was filtered on acid-digested asbestos and washed with alcohol. Phosphorus and pen- tosans were determined. The results are shown in Table 2. PHOSPIIORUS COMPOUNDS OF (lorroxsmzn “EAL .-~..\'n Writwr lhtm 15 TABLE 2. Products of Acid Extracts of Wheat Bran and Phosphorus and Pentoses in Cotton Seed Meal Percentage of Original Substance. Phos- Pentoscs phorus ‘Phosphorus Barium Salt: \Vheat Bran ........................................................................ .. 0.314 0.144 0.459 Cottonseed Meal ................................................................ .. 0 .085 0.111 1.307 Alcohol Precipitate from filtrate from Barium Salt: Wheat Bran ........................................................................ .. 0.001 0.065 65.000 Cottonseed Meal ................................................................ .. 0.001 0.044 44.000 Theoretical: (According to Anderson’s Proposed Formula) ...................................................... .. 0.539 Sum of Pcntoses and of Phosphorus: Wheat Bran ........................................................................ .. 0.315 0.209 .................. .. Cottonseed Meal ................................................................ .. 0.086 0.155 .................. .. Percentage lost by treatment with Barium Hydroxide: Wheat Bran ........................................................................ .. 3. = 31. Cottonseed Meal ................................................................ .. 12. ' 28. From the table it is_ evident that wheat bran and cottonseed meal con- tain soluble pentoses that are associated with very little, if any, phos- phorus, and that these pentoses are insoluble i11 alcohol. The evidence on this point is positive. The barium salt from the wheat bran contains not quite enough pentoses to satisfy the theory, and the cottonseed ineal product contains more than twice as much as is required. The barium hydroxide treatment removed 31 per cent of the pentoses, from the crude salt from wheat bran and. 28 per cent from that of cottonseed meal, "while the amount of phosphorus removed in the same fractions was only 3 per cent and 12 per cent, respectively. Anderson’s barium salt purified by repeated treatment contained only one-eighth as much pentoses as is required by his theory. So it is evident that pentoses are removed in each successive treatment with barium hydroxide. Anderson’s crude salt contained 1.25 per cent nitrogen. He does 11ot report the nitrogen in the barium salt, so We do not know what other non-phosphorised impurities it contained. The above results, and those of Anderson show that we have no ari- (Zence that wheat bran contains‘ an zinosite-phpsphortc acid with pentose in the molecule. There is also no evidence that cottonseed meal contains such an acid. The method of Anderson described above for the preparation of the lime-magnesiumpotassium salt of phytic acid ivas the same as that used by Patten and Hart. The latter workers analyzed a product prepared in this way and which they regarded as an impure;lime-magnesiirm-potas; sium salt of phytic acid. This salt containediby’ analysis 17.30 per cent carbon; the free acid, according to- their theory, contains 10.08 per cent carbon. The lime-magnesium-potassium salt, then, contained 7O per cent more carbon than the free acid, and at least twice as much as their theory would require for such a salt. This abnormal result can be explained by our Work described above. They merely reprecipitated water soluble, alcohol insoluble pentosans, by their method, and obtained a product highly contaminated with those substances. 16 TEXAS AGRICULTURAL EXPERIMENT STATIONS COMPARISON or FORh/IULAS rnorosnn FOR INOSITE-PHOSPHORIC ACID or WHEAT BRAN AND OOTTONSEED MEAL. In Table 3 is shown the percentage composition of inosite-phosphoric acid of wheat bran and cottonseed meal, as calculated from the formulas proposed by different investigators. TABLE 3. Percentage Composition of Inosite-Phosphoric Acid of Cotton Seed Meal and Wheat Bran, Calculated from Formulas Proposed. Proposed Formula ‘ Carbon Hydrogen Phosphorus Patten and Hart ...... .. CQI~IRP2OQ ............................................ .. 10.08 3.36 26.07 Anderson .................... .. C2oH55P9049 ......................................... .. 17.67 4.05 20.54 The writer .................. -|NC1QH4IPQO42 ....................................... .. 12.68 3.61 24.56 We have already shown that the acid would apparently contain less carbon and more phosphorus than it really contained if prepared and analyzed by the method of Patten and Hart. The product of Anderson was obviously not pure, as it lost by simple separation as the free acid enough of some substance to change the calculated formula by OSHMOS. It was not purified further and no homogeneity test was made. Since we have shown that our product was homogeneous and probably pure, we believe that the formula CHHHPOOH is more nearly the truth than the others proposed. THE COMPOSITION OF PHYTIO ACID. The theory that the inosite-phosphoric acid, or so-called “phytic acid” of feeding materials has the composition represented by the formula OQHSROO or a multiple thereof, rest upon the work of Posternak, Pat- ten and Hart, Hart and Tottingham, and Anderson, all of whom used similar methods and analyzed, in most cases, the free acid. We have shown that grave errors are introduced by this method. Patten and Hart, Hart and Tottingham, Anderson, and as far as we have been able to find Posternak made no attempt to prove that their products were homogeneous and free from inorganic substances. The analysis of the free acid is attended with much risk, as gummy substances of this nature are liable to contain water which is imperfectly removed 0-n drying. The method used by Rising provides for the removal of inorganic bases but not for inorganic phosphorus. His method involves precipitation of bases with alcohol and ether, and from our results with alcohol it ap- pears that it accomplished its purpose. Rising analyzed the silver salt of his product and his calculated formula. agrees better with our for- mula than does that of any other investigator. Verbrodt, with the details of whose Work we are unacquainted, pre- pared and analyzed a barium salt of inosite-phospho-ric acid, and assigned the formula O,£,H,,_,P,1O,fi to the free acid. The percentage composition (calculated) of inosite-phosphoric acid, or so-called phytic acid, according to these investigators. is shown in Table 4. Our formula is O,._,H,,P,,O_w, Rising’s multiplied by 2, would be C,,H40P,0O,,,Verb1"odt’s is O12H4OP11046, while the formula. proposed by Posternak, and agreed to by other investigators, multiplied by o‘, would PHOSPHORUS COMPOUNDS or COTTONSJCIEI) ll/[IEAL AND WVIIEAT 131mm 17 be CHH48PI2OM. Both the formula and percentage composition of our acid agree fairly well with that of Rising, and since he made no effort to remove inorganic phosphorus, the slightly lower content of carbon and higher content of phorphorus could be explained in this way. At any rate, the preponderance of evidence is in favor of a carbon content higher than 10.08 per cent, the results of Patten and Hart, Hart and Tottingham, and of Anderson, themselves being always higher than this figure. TABLE 4-. Calculated Percentage Composition of Inosite-Phosphoric Acid According to Proposed Formulas. Proposed Formula Carbon Hydorgen Phosphorus Pasternak, P a t t e n and Hart; Hart and Tottingham; a n d Anderson ................ .. CQHQPQOQ ............................................ .. 10.08 3.36 26.07 Verbrodt .................... .. C12H2oP9042 ....................................... .. 11.42 3.17 27.04 Rising ........................ .. c6H9°P5042 ....................................... .. 12.02 3.34 25.88 Anderson (s y n- CHHHPWOM, calculated .................. .. 12.60 2.80 27.14 thetic).......................... ound .......................................... .. 12.62 3.24 26.51 The writer ................. .. CWHHPSO“ ........................................ .. 12.68 3.61 24.56 We believe our formula, O,,H,,,P9O_,,, more nearly represents the truth and propose it as the empirical formula of i.nosite phosphoric acid, or the so-called “phytic acid” of feeding materials. It will be noticed that the calculated composition of our acid is similar to that of the di-inosite-tripyrophosphoric acid ester Anderson claims to have synthesized from inosite and pyrophosphoric acid, the phosphorus, however, being about 2 per cent higher in the product of Anderson. An- derson states that “the substance was isolated as the barium salt in exactly the same way . . . as described for the dipyrophosphoric ester.” The preparation of this last named ester involved precipitation from phytic acid solution with alcohol. Since the phvtic acid had not been shown to be free from inorganic bases, and since they have been found by us in considerable amounts in products prepared as Anderson prepared the phytic acid, and since these are’ precipitated by alcohol as salts of phytic acid, Anderson’s products are necessarily impure. SUMMARY AND CONCLUSIONS. 1. The method for the preparation of the so-called phytic acid, on which rests the theory that this acid corresponds to the formula C2H,P,O,, or a multiple thereof, have been shown to produce an impure product containing notable amounts of iron and aluminum, calcium, magnesium and inorganic phosphorus. Formulas based on analyses of products prepared by these methods are therefore erroneous. 2. The principal organic phosphoric acids of cottonseed meal and wheat bran were separated and purified and the silver salts fractioned and analyzed. 3. Upon hydrolysis both the acid and the ammonia-soluble organic- phosphorus compounds of cottonseed meal xvere found to yield inosite. 18 'l‘1cx.»\s AGRICUIJEURAJ. ].‘lXI’.l~1l{li\’ll5N'l‘ STATIONS 4. The inosite phosphoric acids of wheat bran and cottonseed meal were found to be identical and correspond to the formula O12H41P9042I 5. Probably two-thirds of the inosite-phosphoric aeid of cottonseed meal is not soluble in 0.2 per cent acid, but is soluble in Water. It is not soluble in water after extracting the material with acid, but may be dis- solved in 0.2 per cent ammonia. The inosite-phosphoric acid of feeding materials is therefore not necessarily confined to the a-eid extract, which has formerly been assumed. ' 6. We have no evidence that wheat bran contains an inosite-phos- phoric acid with pento-se in the molecule. 7'. The formula O12H41P9042 is proposed for inositc-phosphoric acid, or the so-ealled “phytic acid” of feeding materials.