_5&r U QsA LIBRARY OF THE ^ agricultural Experiment Station, | UNIVERSITY OF ILLINOIS. A Books are not to be taken from the Library Room. IL [Reprinted from American Chemical Journal, Vol. XIV, No. 8.] CRYSTALLISED VEGETABLE PROTEIDS. By Thomas B. Osborne. The existence of crystallised proteids in seeds was pointed out by Hartig in 1855. Four years later, Maschke obtained hexagonal plates of proteid matter by extracting Brazil-nuts with water heated to 40°-50° and evaporating the filtered extract at 40°. Nageli 1 investigated the crystal-like forms from the Brazil-nut as well as the artificially produced crystals of Maschke and concluded that they differed in some respects from true crystals. He there- fore designated them as “ crystalloids.” Sachsse, 2 following Maschke’s method, and also by precipitating the aqueous extract with carbonic acid, obtained several prepara- tions of proteid from the Brazil-nut which he analysed. These preparations were, however, not distinct crystals, but were com- posed of small discs. Schmiedeberg 3 obtained crytallised products by treating the aqueous extract of the Brazil-nut with carbon dioxide, washing the precipitated proteid, digesting the precipitate with magnesia suspended in water at 35 0 , filtering and evaporating at the same temperature. These crystals he considered to be a compound of the proteid with magnesia. Drechsel 4 obtained crystals, presumably hexagonal plates, by submitting the solution containing the “ magnesia compound,” prepared as Schmiedeberg suggested, to dialysis in a vessel con- taining alcohol. Drechsel afterwards obtained finely developed crystals by allowing a warm sodium chloride solution saturated with the pro- teid of the Brazil-nut to cool slowly, when the substance separated in well-formed crystals. 6 At Drechsel’s suggestion Griibler applied this method with some modifications to the proteids of squash-seed, and made a large number of preparations of perfectly formed octahedral crystals whose properties and composition he described at length. 6 1 Botanische Mittheilungen (Miinchen, 1863), Vol. I. 2 Die Farbstoffe, Kohlenhydrate und Proteinsubstance (Leipzig, 1877), p. 315. 3 Ztschr. physiol. Chem. 1 , 205. 4 J. prakt.Chem. 19 , 331. 6 Griibler : cf. ibid. 23 , 100. 6 J. prakt. Chem. 23 , 97. 2 Ritthausen 1 obtained crystallised preparations from the castor- bean, hemp-seed and sesame-seed, which separated from a warm sodium chloride solution on cooling in the form of regular octa- hedrons. The composition and properties of the crystals obtained from the hemp-seed and castor-bean he described in a subsequent paper . 2 The writer next prepared rhombohedral and octahedral crystals from the oat-kernel by cooling a warm dilute sodium chloride solution saturated with the proteid . 3 The writer has also obtained octahedral crystals from flax-seed by extracting with a solution of sodium chloride and dialysing the filtered extract ; the proteid separating in well-formed crystals as the salts were removed . 4 The fact that these proteid substances can be artificially crystal- lised is not only interesting in itself, but is important as presumably furnishing a means for making preparations of undoubted purity which will afford a sure basis for further study of their properties. The contradictory statements made by the various investigators, not only in regard to properties and composition of these bodies but also in respect to the value of the methods of solution and separ- ation which have been employed hitherto, render an exact knowl- edge of all the facts relating to these substances a matter of the highest scientific and practical importance. An examination of the literature indicates that definitely crys- tallised preparations from the Brazil-nut have never been analysed, and also shows that the published analyses do not agree suffi- ciently to fix the composition of the substance. Ritthausen’s observations on the properties of the crystallised proteids of the hemp-seed and castor-bean indicate that they are closely related to, if not identical with, the body obtained by Griibler from the squash-seed ; yet the results of Ritthausen’s analyses differ widely from those first made by Griibler. The composition of the crystallised proteid of the squash-seed has been fixed within narrower limits by the analyses of Chit- tenden and Hartwell , 6 Ritthausen , 6 and Griibler . 7 It seemed desirable, therefore, that the composition and properties of the *J. prakt. Chem. 23 , 481. 2 Il>id. 25 , 130. 3 Report Conn. Agri. Expt. Station 1890, and Am. Chem. J. 13 , 408; 14 , 212. 4 Am. Chem. J. 14 , 329, 629. 6 J. of Physiology 11 , 435. ®J. prakt. Chem. 35 , 137. 7 Ibid. 44 , 369. 3 various crystallised proteids should be investigated anew in order to determine the relations of these substances. As the writer had already prepared crystallised products from the oat-kernel and flax-seed, he determined to reinvestigate the similar bodies obtainable from the Brazil-nut, hemp-seed, castor- bean and squash-seed, and the results of this investigation are embodied in the following pages. Crystallised Proteid of the Brazil-nut, Bertholletia . If a thin section of the meat of a Brazil-nut is examined under the microscope, after washing with ether and then with alcohol, a large number of bodies may be seen which more or less resemble crystals in form. If cold water is next applied to the section, an outer coating of these masses appears to be dissolved and a great number of apparently rhombohedral crystals are set free. These crystals are proteid in character, and on adding some sodium chloride solution dissolve instantly. If warm water is used instead of cold, the substance of these crystals dissolves, doubtless in the dilute saline solution which is formed with the alkali-salts of the seed, but which, when cold, has little, if any, solvent power. This view is supported by the fact that the proteid extracted from the seed is wholly insoluble in pure water, either cold or warmed to 40°. Maschke, 1 as already stated, made a solution of this substance by extracting the oil-free meats with water and evaporating at 4° 0_ 5 0 °* He thus obtained thick six-sided plates which he says “belong undoubtedly to the regular system, and the same form would be obtained if an octahedron were cut into three equal parts parallel to one of its octahedral faces ; the middle part would then produce in every respect one of these six-sided plates.” Maschke made an analysis of his preparations, but considered that the substance was a compound of casein with an acid. Sachsse, 2 following Maschke’s directions, failed to obtain crystals, the substance separating in small discs ( scheibchen ). He also obtained small discs by precipitating the aqueous extract with carbonic acid. As a mean of five analyses of the preparations made after drying at ioo°-iio°, he gave the following: prakt. Chem. 74 , 436. 2 Die Farbstoffe, Kohlenhydrate und Proteinsubstance (Leipzig, 1877), p. 315. 4 Brazil-nut Globulin. — Sachsse. Carbon a. discs. 51.OO Hydrogen 7*25 Nitrogen 18.06 Sulphur I.36 Phosphoric acid (P 2 0 6 ) O.82 Oxygen 21.51 Ash 100.00 0.76 The ash was determined by burning the substance mixed with calcium phosphate, and was found to equal in amount the phos- phoric acid obtained by fusing with soda and potassium nitrate and precipitating with uranium acetate. Sachsse therefore con- cluded that the substance belonged to the class of vegetable caseins. Th. Weyl, 1 in a paper on animal and vegetable proteids, pub- lished an analysis of the proteid of the Brazil-nut obtained by extracting the oil-free meats with io-per cent, sodium chloride solution, precipitating by diluting and passing through the solu- tion carbonic acid gas, washing the precipitate with water and then boiling with alcohol and drying. He found the preparation thus made, after deducting the ash, to have the following composition: Brazil-nut Globulin. — Weyl. b. Carbon 5 2 -43 Hydrogen 7.12 Nitrogen 18.10 Sulphur o -55 Oxygen 21.80 100.00 Weyl considered the substance to be a globulin resembling the vitellin obtained from the yolk of eggs. Ritthausen, 2 replying to the criticism^ of Weyl on the methods employed by him which involved the use of weak alkali solutions in preparing proteids from seeds, published the results obtained by extracting the oil-free meats of the Brazil-nut with a dilute 1 Ztschr. physiol. Chem. 1, 72. 4 Pfliiger's Archiv 16 , 301. CRYSTALLISED GLOBULIN OF BRAZIL-NUT. Digitized by the Internet Archive in 2017 with funding from University of Illinois Urbana-Champaign Alternates https://archive.org/details/crystallisedvegeOOosbo 5 solution of potash, precipitating the filtered extract with acetic acid, and washing the precipitate so produced with water, alcohol and ether. After drying at 130°, Ritthausen found his preparation to have the following composition: Brazil-nut Globulin. — Ritthausen. c. Carbon 52.29 Hydrogen 7*24 Nitrogen 18.09 Sulphur 1.32 Oxygen 21.06 100.00 . The considerable discrepancies among the foregoing analyses and the fact that none of the preparations were distinctly crystall- ised make further investigations desirable. For this purpose a large quantity of fresh Brazil-nuts was procured, the outer shells removed, the perfectly sound meats were carefully selected, chopped fine and treated with benzine (petroleum-ether) in order to remove the oil. The finer portions of the material were then washed through a straining cloth by the aid of benzine, and a quantity of fine white meal was thus obtained. 100 grams of this meal were then treated with 10-per cent, sodium chloride solution and the very slightly colored extract filtered perfectly clear. The residue of the meal was treated twice more in the same way with sodium chloride solution, and the extracts obtained were united and saturated with pure ammonium sulphate. The snow-white precipitate resulting was thrown on a filter and, when freed from the solution, was removed from the paper and suspended in distilled water. The dilute solution of ammonium sulphate thus produced dissolved but a part of the precipitate. Complete solution resulted immediately on adding a little sodium chloride, showing that this latter salt is a far more energetic solvent for this substance than ammonium sulphate. The solution, after filtering perfectly clear, was placed in a dialyser and suspended in running water for three days, when it was removed and filtered from the large amount of separated substance. The solid product was washed thoroughly with water, alcohol and ether, and dried over sulphuric acid. It weighed nearly 25 grams. Under the microscope it was seen to consist wholly of crystals in the form of hexagonal plates, as shown in the accompanying plate. Dried at 1 io° these crystals had the following composition : 6 Brazil-nut Globulin: Crystals— Preparation i. I. II Average. Ash-free. Carbon 52.11 52.05 52.08 52.18 Hydrogen 6.91 6.92 6.92 6.92 Nitroge n 18.28 18.25 18.27 18.30 Sulphur 1.03 1.06 I.05 I.05 Oxygen 21.55 Ash 0.20 100.00 Ash. — 0.8475 g ram substance, dried at no°, gave 0.0017 gram ash~o.20 per cent. Carbon and hydrogen, I. — 0.3702 gram dried substance gave 0.7072 gram COuzr52.11 per cent. C, and 0.2303 gram H 2 Orz6.9i per cent. H. II. — 0.3736 gram dried substance gave 0.7130 gram C 0 2 r= 52 .o 5 per cent. C, and 0.2326 gram H 2 0n6.92 per cent. H. Nitrogen, I. — 0.3887 gram dried substance gave 57.14 cc. N at 5 0 (barometer 769.5 mm. at 22 0 ) 18.28 per cent. II. — 0.3953 g ram dried substance gave 57.74 cc. N at 5 0 (barometer 773.0 mm. at 22 0 )— 18.25 P er cent. Sulphur, I. — 0.7940 gram dried substance gave 0.0596 gram BaS 0 4 = 1.03 per cent. S. II. — 0.6020 gram dried substance gave 0.0465 gram BaS 0 4 zr 1.06 per cent. S. A second preparation of this substance was made by extracting 50 grams of the oil-free meats with 10-per cent, sodium chloride solution and at once dialysing the clear filtered extract. During two days nothing solid appeared in the dialyser, but soon after a deposit began to form, and at the end of seven days a considerable precipitate had separated which was filtered off, washed with water, alcohol and ether, and when dried over sulphuric acid weighed 6 grams. Under the microscope this substance appeared to consist of small spheroids with a more or less irregular outline, strongly suggesting crystals, but presenting no definite crystalline form. Dried at 1 io°, the substance yielded the following results on analysis : Brazil-nut Globulin: Spheroids — Preparation 2. Carbon 51.87 Hydrogen 6.90 Nitrogen 18.00 Sulphur 1. 11 Oxygen Ash 0.92 Ash-free. 52-35 6.96 18.16 1 . 1 2 21.4I 100.00 7 Ash. — 0.681 1 gram substance, dried at no°, gave 0.0063 g ram ash = 0.92 per cent. Carbon and hydrogen. — 0.2919 gram dried substance gave 0.5515 gram C0 2 zr 51.87 per cent. C, and 0.1823 gram H a O 31 6.90 per cent. H. Nitrogen. — 0.3963 gram dried substance gave 57.14 cc. N at 5 0 (barome- ter 772.7 mm. at 23 0 ) = 18.00 per cent. Sulphur.— 0.8100 gram dried substance gave 0.0648 gram BaS0 4 =;i.ii per cent. S. A third preparation was made by extracting 25 grams of the oil-free meal with distilled water heated to 6o°. The solution was then allowed to stand several days in a cold room at a tem- perature of about 5 0 , thymol being freely used to prevent decom- position. After the separated substance had settled, the supernatant solution was decanted, the sediment collected on a filter, washed with water, alcohol and ether, and dried over sulphuric acid. It weighed 7 grams. The particles composing this precipitate were similar to those of the preparation last described, but smaller. Dried at 1 io°, it was found by analysis to have the following composition: Brazil-nut Globulin : Spheroids — Preparation 3. Ash-free. Carbon 51.78 52.16 Hydrogen 6-93 6.98 Nitrogen 18.19 18.32 Sulphur 1.06 1.07 Oxygen . . 21.47 Ash 0.72 100.00 Ash. — 0.7667 gram substance, dried at no°, gave 0.0055 gram ash = 0.72 per cent. Carbon and hydrogen. — 0.3252 gram dried substance gave 0.6174 gram C0 2 51.78 per cent. C, and 0.2030 gram H a O zz: 6.93 per cent. H. Nitrogen. — 0.4062 gram dried substance gave 59.42 cc. N at 5 0 (barometer 769.5 mm. at 22 0 ) — 18.19 P er cent. Sulphur. — 0.8210 gram dried substance gave 0.0630 gram BaS0 4 = 1.06 per cent. S. These analyses show that this proteid has the same composition whether crystallised or spheroidal, and that the different methods of extraction give the same results. To facilitate comparison, all the analyses are here tabulated. 8 Brazil-nut Globulin. Sachsse. 1 Ritthausen. Weyl. Osborne. a b c 1 2 3 discs. crystals. spheroids. spheroids. Carbon 5 L 42 52.29 52.43 52.18 52.35 52.l6 Hydrogen 7.31 7.24 7.12 6.92 6.96 6.98 Nitrogen l8.2I 18.09 18.IO 18.30 18.16 18.32 Sulphur 1-37 I.32 °*55 I.06 1.12 I.07 Oxygen 21.69 21.06 21.80 21.54 2I.4I 21.47 • 100.00 100.00 100.00 100.00 100.00 100.00 With the exception of Sachsse’s carbon determination, the figures for this element and for nitrogen are in fair agreement. The figures for hydrogen obtained by the writer are somewhat lower than the others, but as special precautions were taken to keep the reagents absolutely free from moisture, and as these agree with each other, it is probable that they are more nearly correct. Weyl’s sulphur determination is unquestionably incorrect. Sachsseand Ritthausen obtained about 0.2 per cent, more sulphur than the writer, but as it does not appear that either of these investigators examined the ash of their preparations for sulphate, it is possible that the cause of this difference may be due to sulphur present in such form. The writer’s reagents were abso- lutely free from any trace of sulphur. In view of these facts, and also since preparation I was well crystallised and in all probability pure, we may fairly take its analysis as closely representing the composition of the crystalline proteid of the Brazil-nut. There can be little doubt that this artificially crystallised proteid is identical with that naturally occurring in the Brazil-nut, with which it agrees in its properties. The fact that it crystallises during dialysis is strong evidence that it is the same substance, entirely unchanged, for the “ crystalloids ” of the seed almost certainly were deposited in its cells by a similar process. It is true that the artificial crystals are hexagonal plates, while the majority of the natural crystals resemble rhombohedrons, but the writer has shown that oat-globulin under a slight change of conditions may yield either octahedral or rhombohedral crystals. So the proteid of hemp-seed sometimes gives hexagonal plates, although as a rule octahedral crystals are obtained. 1 Sachsse’s analysis is here calculated free from phosphoric acid. 9 The reactions of the artificially crystallised proteid of the Brazil- nut are as follows : Insoluble in pure distilled water, even on warming to 50° C. Soluble in solutions of sodium chloride, from which it is partly precipitated on dilution with water, but not thrown down by saturation with sodium chloride. Precipitated completely by saturation with ammonium sulphate, and partially by saturation with magnesium sulphate. Very soluble in distilled water containing a minute amount of acid. Water containing per cent, of hydrochloric acid dissolves the proteid at once; the amount of acid required being propor- tional to the quantity of substance. In other words, sufficient acid must be present to convert all the substance into an acid compound before complete solution results. Addition of a very small amount of sodium chloride, or other salt, precipitates the solution. This precipitate is insoluble in a large amount of sodium chloride, denoting the formation of an acid-compound. These acid solutions do not coagulate on boiling and are not pre- cipitated by alcohol. Citric acid dissolves the proteid less readily than hydrochloric acid, solutions of per cent, having but little solvent power. £ per cent, solution of citric acid gives a turbid solution, which becomes clear at once on addition of a little sodium chloride, but is precipitated on adding more of the salt, the precipitate being insoluble in an excess. Dilute sodium carbonate solutions of £ to i per cent, dissolve the proteid instantly and completely, as do also solutions con- taining a very little caustic alkali. On careful neutralisation the proteid is thrown down unchanged, being readily and completely soluble in sodium chloride solution. If acid is added in excess the proteid dissolves, and on again neutralising is precipitated in a form soluble in sodium chloride. Dissolved in 10-per cent, sodium chloride solution a turbidity is produced on gradually heating to 70°, and on raising the heat to 86° flocks separate, which increase in amount if the temperature is slowly raised to ioo°. This substance responds to all the usual tests for proteids, such as the biuret, Millon’s, xanthoproteic, etc. The same reactions were given by the preparations made by cooling the warm aqueous extract. These reactions show this proteid to be a true globulin of the vitellin class. IO Crystallised Proteid of Hemp-seed. As already stated, Ritthausen has described 1 a crystalline prepar- ation obtained by extracting powdered hemp-cake with 5-per cent, sodium chloride solution heated to 40°, and allowing the filtered extract to cool. The proteid separated in isometric crystals. The filtrate on diluting gave a further precipitate, which was united with the first and recrystallised according to the method used by Griibler in his work on the squash-seed. A perfectly crystallised product was thus obtained, consisting of “ octahedra, rhombodedecahedra and other forms of the regular system.” In a subsequent paper 2 Ritthausen published the results of his analyses of this proteid. Four preparations were made and ana- lysed, only one of which, however, was crystallised. These were obtained : d, by diluting the sodium-chloride extract with water ; e, by extracting d with 20-per cent, sodium chloride solu- tion and preparing the insoluble portion for analysis ; f, by crys- tallising d by Drechsel’s method ; g, by extracting the seeds with ^-per cent, potash solution. The composition of these products was as follows : Ritthausen. Carbon d. 50*94 e. 51*33 Crystals. 50.98 51.19 Hydrogen 6.85 6.84 6.92 6.97 Nitrogen 18.62 18.80 18.73 18.06 Sulphur 0.8 0.82 O.82 0.92 Oxygen 22.74 22.21 22.55 22.86 100.00 100.00 100.00 100.00 These analyses agreed with the results first obtained by Griibler for the proteid of the squash-seed, with the exception of carbon, of which Griibler found 53.30 per cent. Ritthausen 3 then pro- cured some of Griibler’s preparation and redetermined the carbon, finding as an average of four analyses 51.61 per cent. This result has since been confirmed by Griibler himself, 4 who now gives the percentage of carbon as 51.48 per cent. It will be noticed that the carbon of Ritthausen’s crystallised preparation is 0.5 per cent, lower than these figures and throws some doubt on the identity of these substances. In order to clear J. prakt. Chem. 23 , 481. *Ibid. 25 , 130. 3 Ibid. 25 , 137. * Ibid. 44 , 369. 1 1 up this point, ioo grams of ground hemp-seeds — from which the oil had been removed by extracting with benzine, and the greater part of the hulls separated by passing through a fine sieve — were extracted with io-per cent, sodium chloride solution and the filtered extract placed in a dialyser set in running water. After five days a large deposit was found in the dialyser, consisting almost wholly of octahedral crystals mixed with a few spheroids. This substance was filtered off, washed with water, alcohol and ether and dried over sulphuric acid. 7.8 grams were so obtained, which, when dried at no°, had the following composition : Hemp-seed Globulin : Crystals — Preparation 4 . Ash-free. Carbon 51.00 51.26 Hydrogen 6.83 6.86 Nitrogen 18.59 18.68 Sulphur 0.94 0.94 Oxygen . . 22.26 Ash 0.50 100.00 Ash. — 0.4042 gram substance, dried at no°, gave 0.0020 gram ash = 0.50 per cent. Carbon and hydrogen. — 0.2977 gram dried substance gave 0.5567 gram C 0 2 =51.00 per cent, C, and 0.1840 gram H 2 0 = 6.83 per cent. H. Nitrogen. — 0.3464 gram dried substance gave 53.96 cc. N at 6° (barome- ter 741.7 mm. at 22 0 ) = 18.59 P er c ent * Sulphur. — 0.9930 gram dried substance gave 0.0670 gram BaS0 4 = o.92 per cent. S. Another preparation was made by extracting 200 grams of ground hemp-seeds, freed from oil, with two liters of 5-per cent, sodium chloride solution heated to 6o°, and filtering while hot. On cooling gradually to 5 0 a large amount of substance separated, which consisted wholly of octahedral crystals. This substance was then filtered off and washed with water. After most of the salt had been washed out, the preparation began to dissolve, the washings being precipitated on falling into the filtrate containing sodium chloride. Dilute alcohol was then substituted for water and the washing continued, the alcohol being gradually increased in strength to absolute. After washing next with ether the prepar- ation was dried over sulphuric acid, and weighed 12.25 grams. Dried at no° it had the following composition : 2 Hemp-seed Globulin: Crystals — Preparation 5. Ash-free. Carbon 51.24 51.28 Hydrogen 6.84 6.84 Nitrogen 00 bo OJ 18.84 Sulphur 0.87 O.87 Oxygen 22.17 Ash 0.0 7 100.00 Ash. — 0.7188 gram substance, dried at no°, gave 0.0005 gram ash — 0.07 per cent. Carbon and hydrogen. — 0.4022 gram dried substance gave 0.7556 gram C 0 2 51.24 per cent. C, and 0.2477 gram FI 2 0 ~ 6.84 per cent. H. Nitrogen. — 0.3495 gram dried substance gave 54.16 cc. N at 6° (baro- meter 754.7 mm. at 23 0 ) ~ 18.83 P er cent. Sulphur. — 0.9355 g ram dried substance gave 0.0597 gram BaS 0 4 n 0.87 per cent. S. A small amount of substance similarly prepared was found to contain, dried at no°, 18.72 per cent N and 0.30 of ash, or, ash- free, 18.77 P er cent. N. All these analyses are here brought together. Hemp-seed Globulin. Ritthausen. Osborne. d. e. /. Crystals. s- 4 . Crystals. 5 . Crystals. Carbon 50-94 5 r *33 50.98 5 LI 9 51.26 ' 5 t -28 Hydrogen 6.85 6.84 6.92 6.97 6.86 6.84 Nitrogen 18.62 18.80 00 -b GJ 18.06 18.68 18.84 Sulphur 0.85 0.82 0.82 0.92 0.94 O.87 Oxygen 22.74 22.21 22-55 22.86 22.26 22.17 100.00 100.00 100.00 100.00 100 00 100.00 The crystallised preparations, f f 4 and 5 are in close agreement in all respects, excepting the carbon of f, which is somewhat low. Ritthausen’s analyses e and g agree with the writer’s and it is therefore almost certain that these are very nearly correct. That Ritthausen’s carbon was a little too low was probably due to the fact that these crystalline preparations absorb moisture with great rapidity. Preparations which had been dried at 1 io° and preserved in closely-corked bottles in a desiccator were found to have absorbed from 0.5 to 1.0 percent, of moisture in consequence 3 of the brief exposure to the air incurred in removing the portions required for nitrogen determinations. The figures obtained for nitrogen agree as closely as could be expected, with the exception of that for g, which is unquestionably too low. As 5 was perfectly crystallised and undoubtedly pure, its analysis may safely be taken as closely representing the true composition of this proteid. The reactions of this body were found to be as follows : In pure water 4 was wholly insoluble, even on heating to 50° ; on the other hand, 5 yielded a turbid solution with distilled water, becoming very nearly clear on warming to 50°. 4 was prepared by diatysis ; 5 by cooling a warm sodium chloride solution. A similar difference for preparations made in these two ways was noticed in the globulin of the oat-kernel, 1 and also in that of the castor-bean, next to be described. The solution in water is pre- cipitated on adding a little sodium chloride; the precipitate so produced being soluble on a further addition of this salt. 4 is wholly insoluble in glycerin, either concentrated or dilute, while 5 dis- solves to a perfectly clear liquid, which is not precipitated by dilution, but is precipitated by adding a little salt, the substance redissolving on adding more salt. Both preparations dissolve in solutions of sodium chloride at 20°. The solubility increases with the temperature up to 6o° or over. Near o° very little, if any, dissolves in 10-per cent, sodium chloride solution. Saturation of the sodium chloride solution of this globulin with salt produces a trifling precipitate; saturation with ammonium sulphate gives a complete precipitation, as does also saturation with magnesium sulphate. It is very soluble in distilled water containing a minute amount of acid. Preparation 5 dissolves instantly and com- pletely in Tofoo'P er cent - hydrochloric acid, provided enough is taken to combine with all the substance present. 4, on the other hand, requires a somewhat stronger acid, but is readily soluble in y^-y-per cent, hydrochloric acid. Very dilute solutions of sodium carbonate (about y 1 -^ per cent.) dissolve both prepara- tions readily and completely, as do also solutions of caustic alkalies. Dissolved in 10-per cent, sodium chloride solution, a turbidity occurs on heating to 75 0 , and flocks form at 86°. If the solution is then filtered and' again heated, turbidity results a second time at 88°, flocks forming at 95 0 . The amount of substance separated Am. Chem. J. 13, 408. 14 at 86° is hardly more than a trace and undoubtedly indicates the presence of an extremely small amount of some other proteid, for if the sodium chloride solution is saturated with salt and the trifling precipitate produced is filtered off and the filtrate diluted until it contains io-per cent, of sodium chloride, it first becomes turbid when heated to 87°, and no flocks appear until heated to 94 0 . If the solution is then raised to 98°, kept at that temperature for some time, filtered from the coagulum and boiled, no more precipi- tate is produced. If acid be added a very heavy precipitate falls, showing either that the globulin is only partly precipitated at ioo°, or that it undergoes a change, a portion being converted into a proteid unaffected by boiling. This substance gives all the reactions common to proteids. The crystals in which this proteid usually separates are simple octahedrons, although modifications of this form are frequently to be seen. Occasionally hexagonal plates may be found among the octahedrons. In two cases where this body was crystallised in small quantity, the entire crop of crystals consisted of very thin hexagonal plates. The facility with which crystals of this proteid can be obtained is very remarkable, it separating in this form whenever precipi- tated by dialysing or by cooling its solutions. The other seeds from which crystallised proteids have thus far been obtained yield well-crystallised products only with considerable difficulty, as the proteid usually separates in the spheroidal form. Crystalline Proteid of the Castor-bean. Ritthausen 1 first obtained the globulin of the castor-bean in a well-crystallised condition by following Grubler’s method. Sub- sequently 2 he gave the results of his analyses of this substance (^), which were as follows : As the mean of several analyses of two preparations, previously described, 3 obtained by extracting the seeds with sodium chloride solution and precipitating the extracts by diluting and .passing through them carbonic acid gas, Ritthausen also gave the figures i and j. 1 J. prakt. Chem. 23 , 481. 3 Pfliiger’s Archiv 19 , 15. 3 Ibid. 25 , 130. 15 Castor-bean Globulin. — Ritthausen. h. Crystals. i. J. Carbon 50.88 52.05 S l - 3 l Hydrogen 6.98 6.83 6.90 Nitrogen 18.58 18.57 18.43 Sulphur O.77 0.96 0.97 Oxygen 22.79 21.59 22.39 100.00 100.00 100.00 As these analyses leave the composition of this proteid in doubt, and since many of the reactions now considered important were not given by Ritthausen, this substance has been reinvestigated by the writer. The material for the preparation of this proteid was obtained by freeing the castor-beans from their shells, bruising the meats in a mortar, extracting the oil with ether, and pulverising the residue. 60 grams of this meal were extracted with io-per cent, sodium chloride brine and the filtered extract placed in a dialyser. After removal of the chlorides, the separated proteid was filtered off, washed with water, alcohol and ether, and dried over sulphuric acid. 4.25 grams of material were so obtained, consisting of spheroids and .crystals mixed with intermediate forms. After drying at no°, this preparation was analysed with the following result: Castor-bean Globulin : Crystals and Spheroids . — Preparation 6 . Ash-free. Carbon 51-57 51.62 Hydrogen 6.92 6.92 Nitrogen 18.79 18.81 Sulphur 0.76 O.76 Oxygen 21.89 Ash 0. 10 100.00 Ash. — 0.5615 gram substance, dried at 1 io°, gave 0.0006 gram ash — 0.10 per cent. Carbon and hydrogen. — 0.3069 gram dried substance gave 0.5803 gram CO a ~ 51.57 per cent. C, and 0.1914 gram H a O — 6.92 per cent. H. Nitrogen. — 0.2855 gram dried substance gave 43.64 cc. N at 6° (baro- meter 763.3 mm. at 23 0 ) zz. 18.79 P er cent. Sulphur. — 0.9173 gram dried substance gave 0.0503 gram BaS 0 4 — 0.76 per cent. S. i6 Another extraction was made by treating 75 grams of the fine oil-free meal with 1000 cc. of a 5-per cent, sodium chloride solu- tion heated to 6o°, filtering perfectly clear and allowing the extract to cool to 5 0 . A considerable amount of substance separated, which was filtered off and washed with water, in which it was found to dissolve when the salt was nearly all removed. The entire precipitate was then dissolved in 10-per cent, sodium chloride brine and the resulting clear solution saturated with salt. This produced a considerable precipitate, which was filtered off and washed with a saturated salt solution. The filtrate and wash- ings were then put in a dialyser. The portion precipitated by saturation with sodium chloride was dissolved in a dilute solution of this salt and likewise sub- mitted to dialysis. After the contents of the dialysers had become free from chlorides the precipitates were filtered off, washed with alcohol and ether, and dried over sulphuric acid. The prepara- tion obtained from the precipitate produced by saturation with sodium chloride weighed 3.3 grams and consisted of imperfectly formed crystals and spheroids ; while that obtained from the saturated salt solution weighed 2.5 grams and was likewise partly crystalline. These two preparations were analysed, the compo- sition of the former being shown by the analysis of preparation 7 , the latter by that of 8 . Castor-bean Globulin: Crystals and spheroids insoluble in saturated sodium chloride solution — 7 . Ash-free. I. II. Average. Carbon 51.40 51.29 SI -35 Hydrogen 6.81 6.86 6.84 Nitrogen 18.88 18.88 Sulphur 0.79 O.79 Oxygen 22.14 Ash 0.00 100.00 Ash. — 0.6773 gram substance, dried at no 0 , gave no weighable ash. Carbon and hydrogen, I. — 0.3614 gram dried substance gave 0.681 1 gram C 0 2 zz 51.40 per cent. C, and 0.2215 g rar n H s O zr 6.81 per cent. H. II. — 0.3544 gram dried substance gave 0.6664 gram C 0 2 z= 51.29 percent. C, and 0.2187 gram H 2 0 z: 6.86 per cent. H. Nitrogen. — 0.3481 gram dried substance gave 53.55 cc. N at 6° (barometer 762,3 mm. at 23.5 0 ) =z 18.88 per cent. 1 7 Sulphur. — 0.7960 gram dried substance gave 0.0460 gram BaS 0 4 zz 0.79 per cent. S. Castor-bean Globulin: Crystals and spheroids soluble in saturated sodium chloride solution — 8. I. II. Average. Ash-free. Carbon 5 X * I 4 50-99 5 x -o7 5 I-I 9 Hydrogen 6.82 6.81 6.82 6.84 Nitrogen 18.88 18.88 18.92 Sulphur ) Oxygen \ •• 2 3-05 Ash 0. 24 0.24 100.00 Ash. — 0.4480 gram substance, dried at no 0 , gave 0.0011 gram ash=: 0.24 per cent. Carbon and hydrogen, I. — 0.3483 gram dried substance gave 0.6531 gram C 0 2 z: 51.14 per cent. C, and 0.2137 gram H a O 11:6.82 per cent. H. II. — 0.3485 gram dried substance gave 0.6516 gram C 0 2 zr 50.99 per cent. C, and 0.2136 gram II a O =3 6.81 per cent. H. Nitrogen. — 0.2987 gram dried substance gave 46 05 cc. N at 6° (barometer 760.2 mm. at 22 0 ) zz 18.88 per cent. The filtrate from the precipitate, produced by cooling the 5-per cent, sodium chloride extract last described, was heated to 40° in a water-bath, and twice its volume of distilled water heated to 40° was added. A perfectly clear solution resulted, which when slowly cooled to 5 0 yielded a well-crystallised precipitate, the crystals being almost exclusively octahedrons with one angle truncated. This precipitate was filtered off, washed with water, alcohol and ether, and when dried over sulphuric acid weighed 3.5 grams. Dried at no°, this preparation was found to have the following composition : Castor-bean Globulin : Crystals — Preparation 9. I. II. Average. Ash-free. Carbon 5 I- 3 I 51.26 51.29 5 r - 3 [ Hydrogen 6.94 7.00 6.97 6.97 Nitrogen 18.75 18.75 18.75 Sulphur 0.70 0.82 O.76 O.76 Oxygen 22.21 Ash 0.03 O.03 100.00 i8 Ash. — 0.6315 gram substance, dried at no 0 , gave 0.0002 gram ash — 0.03 per cent. Carbon and hydrogen, I. — 0.3157 gram dried substance gave 0.5939 gram C 0 2 = 51.31 per cent. C, and 0.1969 gram H 2 0 — 6.94 per cent. H. II. — 0.3976 gram dried substance gave 0.7473 gram C 0 2 — 51.26 per cent. C, and 0.2.516 gram H 2 0=.7.oo per cent. H. Nitrogen. — 0.3074 gram dried substance gave 48.1 cc. N at 6° (barometer 744.0 mm. at 21 0 ) — 18.75 P er cent. Sulphur, I. — 0.5468 gram dried substance gave 0.0273 gram BaSO.* — 0.70 per cent. S. II. — 0.4588 gram dried substance gave 0.0275 gram BaS 0 4 =10.82 per cent. S. The following table shows the results of the analyses of Ritt- hausen and the writer: Castor-bean Globulin. Ritthausen. h i j Crystals. Carbon 50.88 52.05 5 L 3 i Hydrogen 6.98 6.83 6.90 Nitrogen 18.58 00 18.43 Sulphur 0 77 0.96 0.97 Oxygen 22.79 21.59 22.39 100.00 100.00 100.00 Osborne. 6 7 8 9 Crystals and Spheroids. Crystals. 51.62 51-35 5 I - I 9 5 I - 3 i 6.92 6.84 6.84 6.97 18.81 18.88 18.92 18.75 0.76 21.89 o- 79 \ 22.14 / 23.05 0.76 22.21 100.00 100.00 100.00 100.00 This table shows plainly that Ritthausen’s figure for carbon in the analysis of his crystallised preparation is too low, probably because the sample had taken up a little moisture from the air, as explained before in discussing the analyses of the hemp-seed proteids. Apart from this, Ritthausen’s analysis of his crystallised preparation agrees well with those of the writer. The analysis of the perfectly crystallised preparation 9 may be taken as closely representing the true composition of this proteid. In pure water this substance behaved like that obtained from the hemp-seed — that is, preparations separated by dialysis were insoluble in water, either at 20°or 50°; those separated by cooling warm saline solutions were soluble to opalescent solutions at 20°, becoming clear on warming. Solutions in water are precipitated by a small quantity of sodium chloride, the precipitate dissolving on adding a larger quantity of the salt. Preparations 7 and 8, which dissolved in water when first separated from the warm l 9 solution, after being redissolved and then precipitated by dialysis, were wholly insoluble in water. Toward glycerin this proteid behaves in the same way as the hemp-seed globulin. Some of the preparations made dissolved completely, others partly, in a io-per cent, sodium chloride solu- tion, leaving a considerable insoluble residue (Weyl’s “albumi- nate ”), which, however, dissolved almost wholly on warming to 50°, a part separating in spheroids when the solution was cooled. Saturation of the salt-solution with ammonium sulphate com- pletely precipitates this proteid. Saturation with sodium chloride precipitates a large part of the globulin contained in the extract of the seed, the part precipitated and that remaining in solution being alike in most of their reactions and composition. It is probable that the precipitate produced by saturation with sodium chloride solution is caused by a small amount of some other substance extracted from the seed, for the crystalline prepar- ation 9, which was obtained from the extract after previous precipitation of a part of the globulin contained in it, gave on saturation with this salt only a trifling precipitate. A part of the precipitate is also unquestionably due to a small amount of another globulin coagulating at from 75 0 to 86°, for where this precipitate is dissolved in io-per cent, sodium chloride brine and heated to this temperature, a much larger coagulum is produced than when the substance obtained from the salt-saturated solution is similarly treated. The perfectly crystalline preparation 9, which yields but a very small amount of precipitate on saturating its salt-solution with sodium chloride, likewise gives but a trace of coagulum when heated to 89°. The crystalline globulin is therefore not precipi- tated by saturation with sodium chloride. Dissolved in 10 per cent, sodium chloride solution, the crystall- ine preparation 9 became turbid when heated to 87°, and at 89° a very minute amount of flocks separated. At 95 0 a large coag- ulum developed and, on heating the solution for some time at 98°, after filtering off the coagulum, the filtrate gave only a turbidity when boiled, but on adding acid a very large precipitate fell, showing that the globulin is but partly precipitated by heat. The other preparations behaved, on heating their salt-solutions, in exactly the same way, except that the coagulum separating at the lower temperature was somewhat greater in amount. 20 This proteid can be crystallised only with difficulty, being in this respect very different from the globulin of the hemp-seed. It is not improbable that this is due to other constituents of the seed which interfere to a greater or less extent with the formation of the crystals. The Crystalline Proteid of the Flax-seed. In a former paper 1 the writer has described a crystallised globulin obtained from flax-seed, the appearance of the crystals being shown by the accompanying plate. As the mean of a large number of accordant analyses its composition was given as follows : Crystallised Flax-seed Globulin. Carbon 51.48 Hydrogen 6.94 Nitrogen 18.60 Sulphur 0.81 Oxygen 22.17 100.00 In pure distilled water at 20° this proteid, whether separated from solution by cooling or by dialysis, is wholly insoluble, while in water at 40° it is very slightly soluble. In 10 per cent, sodium chloride solution it is mostly soluble, a part (Weyl’s “ albumi- nate”) generally remaining undissolved, which, however, dissolves readily on warming the solution and partly separates, on cooling, in very finely developed crystals. In glycerin diluted with water the substance separated by dialysis is wholly insoluble, either at 2o°or40°; while the substance separated from a warm sodium chloride solution is largely soluble at 20°. Saturation with ammonium sulphate and also with magnesium sulphate completely precipitates this proteid from its solutions. Saturation with sodium chloride gives a small precipitate which consists partly of other globulins, for the filtrate from this precipi- tate, when diluted so as to contain 10 per cent, of sodium chloride, yields but a trace of coagulum on heating to boiling, while the precipitate dissolved in 10-per cent, sodium chloride coagulated at the various temperatures observed in the solution before saturation. 1 Am. Chem. J. 14 , 629. CRYSTALLISED GLOBULIN OF FLAX-SEED. 21 In very dilute acids — as-^-per cent, hydrochloric acid or -j^-per cent, citric acid — the proteid readily and completely dissolves to solutions which are precipitated by adding a little sodium chloride. In ^-per cent, sodium carbonate solution or y^-per cent, potash solution the substance dissolves at once. On heating a io-per cent, sodium chloride solution of this globu- lin small amounts of substance separate at different temperatures, according to the method employed in the preparation of the pro- teid. The substances coagulating below 90° are undoubtedly traces of other globulins, as already stated. These globulins are also separated from the crystalline globulin if the latter is precipi- tated from solution by cooling, and after washing with water, alcohol and ether and drying over sulphuric acid, is then dis- solved in io-per cent, sodium chloride brine. Thus separated from these coagulable proteids, boiling produces but a trace of a coagulum, while in the case of the globulin obtained by dialysis the coagulum produced by boiling is a little greater. In any case the coagulation produced by boiling is only partial, for a very large precipitate is always obtained on adding acid to these solu- tions, even after boiling them for some time. The crystals of this proteid were always obtained as octahedrons. The Crystallised Proteid of the Oat-kernel. In this Journal 1 and also in the Report of the Connecticut Agri- cultural Experiment Station for 1891. the writer described a globulin from the oat-kernel which was obtained in a perfectly crystallised condition. The form of these crystals is shown in the plate. As an average of three closely agreeing analyses of different preparations, its composition was found to be as follows : Crystallised Oat-globulin. Carbon 52.18 7.05 17.90 °*S 3 22.34 Hydrogen Nitrogen Sulphur Oxygen 100.00 Am. Chem. J. 13 , 408 ; 14 :, 212. 22 In distilled water at 20° this proteid in the crystallised condition is not soluble; in the spheroidal condition, separated from a warm sodium chloride solution, by boiling it yields an opalescent solu- tion. On slowly warming, the crystals partly dissolve, while the spheroids quickly give a clear solution. Addition of a little sodium chloride precipitates these solutions, but the addition of a further quantity of this salt redissolves the precipitate. In io-per cent, sodium chloride the crystals dissolve at 20°, with the excep- tion of a slight residue which is nearly completely dissolved by warming. This solution is not precipitated by boiling. Dilution precipitates the proteid. In diluted glycerin most of the substance dissolves. Saturation with ammonium sulphate and with magnesium sulphate com- pletely precipitates solutions of this proteid in sodium chloride brine. Saturation with sodium chloride likewise precipitates such solutions. In very dilute acids, y^-per cent, hydrochloric acid or £-per cent, citric acid, it is readily soluble to a perfectly clear solu- tion. In Tper cent, sodium carbonate solution the crystals are soluble with the exception of a slight residue; the spheroids, on the other hand, with -j^-per cent, sodium carbonate give at once a perfectly clear solution. In y^-g-per cent, potash solution both spheroids and crystals are at once dissolved to a perfectly clear solution. The Crystallised Proteid of the Squash-seed. This proteid was first prepared and analysed by Barbieri , 1 whose preparations, however, were not crystalline ; soon after Griibler , 2 employing a method suggested by Drechsel, succeeded in obtain- ing perfectly crystallised preparations of this substance and published the results of his observations and analyses. The per- centage of carbon found by Griibler was shown by Ritthausen 3 to be too high, owing to the use of an imperfect method for the determination of this element. Later Chittenden and Hartwell 4 prepared this proteid in the crystallised form for use in digestion experiments, and published analyses confirming those of Ritthausen. Griibler 6 has lately made a new analysis of this body nearly according with the results of Ritthausen and with those of Chittenden and Hartwell. prakt. Chem. [2] 18 , 102. *lbid. [2] J 33 , 97. *Ibid. [2] 23 , 481. 4 J. Physiol. 11 , 435. 6 J. prakt. Chem. [2] 44 , 369. CRYSTALLISED GLOBULIN OF THE OAT KERNEL. " I 23 As, however, the analyses of these investigators do not har- monise in all respects, the writer has examined this proteid with the hope of fixing its composition more nearly. A quantity of squash-seeds, freed from hulls and finely chopped, were extracted with ether. The coarse powder was then thrown on a straining cloth and the finest portions washed through together with the ether. The coarse meal remaining in the cloth was then ground in a mortar and again washed on the cloth with ether. After freeing from most of the oil and drying in the air, 50 grams of coarse meal were obtained from 200 grams of freshly- shelled meats. The fine meal passing through the cloth was col- lected on a filter and washed with ether until fat-free, and, after drying in the air, weighed 70 grams, all of which passed through a sieve with holes To -inch in diameter. Fifty grams of the coarse meal were then treated with 800 cc. of 2-per cent, sodium chloride solution heated to 6o°, and at once thrown on to a filter. The clear extract was allowed to cool in a beaker left over night in a room having a temperature of about io°. In the morning a large deposit had separated, entirely com- posed of perfectly formed octahedrons. This was filtered off, washed once with water and then thoroughly with 50-per cent, alcohol, then with absolute alcohol and finally with ether, and dried over sulphuric acid. A snow-white preparation, weighing 9 grams, was thus obtained, which was dried at no ? and analysed with the following results : Squash-seed Proteid : Crystallised — Preparation 10 . I. II. III. Average, Ash-free. Carbon 5 I *49 5 l M 5 I *74 51.62 51.66 Hydrogen 6.88 6.92 6.84 6.88 6.89 Nitrogen 18.50 18.49 18.50 18.51 Sulphur 0.88 0.87 0.88 0.88 Oxygen 22.06 Ash 0.08 100.00 Ash. — 0.4789 gram substance, dried at no°, gave 0.0004 gram ash ~o.o8 per cent. Carbon and hydrogen, I. — 0.3868 gram dried substance gave 0.7303 gram C 0 2 — 51.49 per cent. C, and 0.2394 gram H 2 0 — 6.88 per cent. H. II. — 0.3986 gram dried substance gave 0.7548 gram C 0 2 22: 51.64 per cent. C, and 0.2484 gram H 2 0 = 6.92 per cent. II. 24 III. — 0.3116 gram dried substance gave 0.5917 gram C 0 2 =: 51.74 per cent. C, and o. 1919 gram H a O — 6.84 per cent. H. Nitrogen, I. — 0.2577 gram dried substance gave 39.21 cc. N at io° (barom- eter 765.4 mm. at 21 0 ) = 18.50 per cent. II. — 0.3881 gram dried substance gave 59.32 cc. N at n° (barometer 764.7 mm. at 24°)=: 18.49 P er cent. Sulphur, I. — 0.7086 gram dried substance gave 0.0457 gram BaS 0 4 = 0.88 per cent. S. ▲ II. — 0.7804 gram dried substance gave 0.0493 gram BaS 0 4 = 0.87 per cent. S. Another preparation was made by extracting 50 grams of the fine meal with 10-per cent, sodium chloride solution at 20°, filter- ing from the insoluble residue and dialysing the clear extract. As the chloride was removed the proteid precipitated in the form of irregular granules together with some spheroids and crystals. After dialysing three days the large deposit was filtered off, washed with water, alcohol and ether, and dried over sulphuric acid. This preparation weighed 21.27 grams, and when dried at no° was analysed with the following result : Squash-seed Proteid: Spheroids — Preparation 11 . I. II. Average. Ash-free. Carbon 5M 1 51-23 5I- 2 7 51.42 Hydrogen 6.82 6.80 6.8l 6.83 Nitrogen 18.50 18.68 18.59 18.64 Sulphur 0.90 0.89 O.90 O.9O Oxygen ... 22.21 Ash 0.29 ... 100.00 Ash. — 1.1487 grams substance, dried at 1 io°, gave 0.0033 gram ash = 0.29 per cent. Carbon and hydrogen, I. — 0.3835 gram dried substance gave 0.7215 gram C 0 2 = 51*31 per cent. C, and 0.2356 gram H 2 0 = 6.82 per cent. H. II. — 0.3743 gram dried substance gave 0.7032 gram C 0 2 =151.23 percent. C, and 0.2289 gram H 2 0 ~ 6.80 per cent. H. Nitrogen, I. — 0.3313 gram dried substance gave 50.64 cc. N at 1 1° (barom- eter 765.3 mm. at 23 0 ) = 18.50 per cent. II. — 0.3761 gram dried substance gave 58.13 cc. N at n° (barometer 764.0 mm. at 24°)= 18.68 per cent. Sulphur, I. — 0.6426 gram dried substance gave 0.0424 gram BaS 0 4 =20.90 per cent. S. II. — 0.8344 gram dried substance gave 0.0545 gram BaS 0 4 = o.89 per cent. S. 25 The following table gives the results obtained by all the chemists who have analysed this substance : Barbieri, amorphous Griibler’s ist analyses, Rittbausen Chittenden and Hartwell. crystals Griibler’s 2d analyses, 10 Osborne 11 “ 51.88 53*21 51.61 51.60 spheroids 52.03 amorphous 51.81 crystals 51.48 “ 51*66 spheroids 51.42 Carbon. Hydrogen. Nitrogen. Sulphur. Oxygen. 7 * 5 X 18.08 0.60 21.93 7.22 19.22 1.07 19.10 7.00 6.97 6 *93 6.94 6.76 6.89 6.83 18.80 19.08 18.71 18.14 18.51 18.64 1. 01 1.04 1.01 0.96 0.88 0.90 / 21.62 20.92 2I *53 2X *53 22.06 22.21 If we omit Griibler’s first analysis, which, as already stated, is incorrect, the average of the remaining eight is Carbon 51.69 Hydrogen 6.97 Nitrogen I 8. 57 Sulphur 0.91 Oxygen 2 i.86 100.00 The reactions of the preparations made by the writer were as follows : In distilled water either at 20° or warmed to 6 o° the prepara- tions, whether separated from solution in sodium chloride brine by dialysis or by cooling, were wholly insoluble. In glycerin either strong or diluted, they were likewise insoluble, even when warmed to 6o°. In 10-per cent, sodium chloride solution they both dissolved with the exception of a mere trace, and the result- ing solutions were not precipitated by saturating with sodium chloride, but were completely precipitated by saturation with ammonium sulphate and also with magnesium sulphate. In per cent, hydrochloric acid and x V-per cent, citric acid both yielded perfectly clear solutions when sufficient acid was employed to form an acid compound with the proteid. In T Vper cent, sodium carbonate and rrirper cent, potash solutions the preparations dissolved readily and completely; these solutions were not preci- pitated by sodium chloride. When dissolved in io-per cent, sodium chloride solution and heated to coagulation the two prepa- rations behaved somewhat differently, as is best shown by the following table : A being the preparation obtained by cooling the 26 warm 2-per cent, sodium-chloride extract ; B, that precipitated by dialysis. Heat-coagulation of Squash-seed Globulin. Turbidity at A. B. '73 0 \ Small Flocks at 84 j precipitate. After filtering at 8 * 7 ° Turbidity 87 87 [?] Flocks 95 95 After filtering and boiling no coagulum. no coagulum. On adding a drop of acetic acid both yielded a large precipi- tate. These figures indicate that B contains a small amount of some other globulin which coagulates between 73 0 and 84°. Conclusions. I. The data presented in this paper indicate that the crystallised globulins of the Brazil-nut and of the oat-kernel are distinct sub- stances. To facilitate comparison, analyses of these two globu- lins are here stated as follows : Carbon Brazil-nut. 1 . 52.18 Oat-kernel. Average. 52.18 Hydrogen 6.92 7.05 Nitrogen 18.30 17.99 Sulphur I.06 o -53 Oxygen 21.54 22.34 100.00 100.00 If the differences in nitrogen- and sulphur-content are not perhaps sufficient to distinguish these two proteids, their reactions prove them to be distinct, for when prepared in the same manner they are unlike in many respects. In distilled water heated to 6o° the globulin of the Brazil-nut is wholly insoluble, while that of the oat-kernel dissolves com- pletely. Saturation of a 10-per cent, sodium chloride solution of these substances with salt almost completely precipitates the proteid of the oat-kernel, that of the Brazil-nut being unaffected. Saturation of similar solutions with magnesium sulphate preci- pitates but little of the Brazil-nut, but all of the oat-globulin. When solutions of these bodies in 10-per cent, sodium chloride brine are heated, the Brazil-nut globulin begins to separate at 70°, a flocculent coagulum forming at 84°, which increases on raising the temperature to boiling, the proteid being largely, but not 27 wholly, precipitated. The globulin of the oat-kernel, on the other hand, is not coagulated at all by boiling. II. The crystalline globulins of the hemp-seed, castor-bean, squash-§eed and flax-seed are almost identical in composition, as may be seen by comparing the analyses: * Hemp-seed. Castor-bean. Squash-seed. Flax-seed. 5 . 9 . 10 . Average. Carbon 51.28 5 !- 3 i 51.66 51.48 Hydrogen 6.84 6.97 6.89 6.94 Nitrogen 18.84 18.75 18.51 18.60 Sulphur 0.87 0.76 0.88 0.8l Oxygen 22.1 7 22.21 22.06 22.17 100.00 100.00 100.00 100.00 The carbon content of the hemp- and castor-globulins is less than that of the squash- and flax-globulins by about 0.25 per cent., a difference too slight to have importance, if it were not constant for almost all the preparations analysed. The deportment of these globulins toward reagents is also very similar, but in this respect the hemp- and castor-globulins show slight differences from those of the squash- and flax-seed ; the two former are, however, almost exactly alike, and the two latter likewise closely agree together. The following table plainly exhibits these relations: Heat-coagulations. Hemp. Castor. Squash. Flax. Turbidity, 75 ° 87° 67° Flocks, 86 89 . . 80 Filtered, Turbidity, 88 87 87 90 Flocks, 95 95 95 96 Filtered, Boiling, no coagulunt. ditto. ditto. ditto. Filtered and acetic acid added, heavy pp. ditto. ditto. ditto. Solubility . 1 Hemp. 2 Castor. Squash. Flax. A. B. A. B. A. B. A. B. H 2 0 at 20°, P. I. p. I. I. I. I. H 2 0 at 40°, S. I. S. I. I. P. I. Glycerine at 20°, S. I. S. I. I. P. I. Glycerine at 50°, S. 1. s. 1. 1. P. I. 10-per cent. NaCl sol. diluted, Pp. p P . P P . p p . p P . Pp. PP. Saturated with NaCl, T. pp. T. pp. T. pp. T. pp. No Pp. V. V. Saturated with MgSQ 4 , C. pp. C. pp. C. pp. C. pp. C. pp . C. pp. C. pp, 1 P. = Partly soluble ; I. = Insoluble ; S. = Soluble ; Pp. = Precipitate; T. pp. =Trace of precipitate ; C. pp. = Completely precipitated ; V. = Very little precipitate. 2 A indicates preparations obtained by cooling warm salt solutions; B, those obtained by dialysis. 28 As already stated, the proteids coagulating at the lower tem- perature are traces of other globulins imperfectly separated from the crystalline globulin. The coagulum separating at the higher temperature is undoubtedly a part of the crystalline globulin which is broken up when heated to this temperature. It is seen that the temperature at which this coagulum separates is the same for all four substances. In solubility these four proteids are very nearly alike, the most noticeable difference being that the globulins of the hemp-seed and castor-bean, when separated from a warm sodium-chloride solution, are soluble in water and diluted glycerin, while the other preparations, both of this substance and the globulins of other seeds, are insoluble under the same conditions. Of the flax-seed globulin separated from a warm salt solution, a little dissolves in water at 40°. The small precipitate obtained by saturating a sodium -chloride solution of the globulins with salt undoubtedly consists mostly of traces of other globulins. It is at present impossible to assert that these four globulins are the same, but since differences exist between different preparations of globulin from the same seed as great as those found among the globulins of these different seeds, the writer is disposed to con- sider these four globulins as identical. Laboratory of the Connecticut Agricultural Experiment Station, New Haven, Conn. . w MU. '