The tests

cc. of

red reaction, which in each case appeared with 12 cc. were then repeated by adding to each solution II acid and afterward increasing this quantity by successive additions of 0.2 cc. By testing with tropaeolin after each such addition, it was found that one gram of II had reacted with 11.2 cc., 12 with 11.5 cc., and 13 with 11.3 cc. Calculating these figures for the preparation dried at 110° and ash-free, and adding to this the acid originally contained in them, we find that II had reacted with 13.9 cc., 12 with 13.7 cc., and 13 with 14.1 cc., which corresponds almost exactly to a compound of 1 molecule of edestin with 20 molecules of acid, assuming this protein to have a molecular weight of approximately 14,500, or, in other words, to exactly ten times the quantity of acid required to form a soluble compound with 1 gram of edestin.

Strong evidence of a definite reaction with about 10 molecules of acid was obtained by testing with potassium nitrite and iodide. A series of five 1-gram portions of II were suspended in water in small, glass-stoppered bottles, and to them were respectively added 4 cc., 5 cc., 6 cc., 6.5 cc., and 7.5 cc. of decinormal hydrochloric acid and then to each 7.5 cc. of a solution of soluble starch, containing per cent. of potassium iodide and 1 per cent. of potassium nitrite.

The portion containing 7.5 cc. of acid became blue throughout within five minutes, the color first appearing at the top of the solution; that with 6.5 cc. began to turn blue at the top within a minute and a half, and became wholly blue in twelve minutes; that with 6 cc. showed a trace of blue on the surface after five minutes, which, even after thirty minutes, was very slight and limited to the upper surface; that with 5 cc. showed a trace of blue on the surface after fifteen minutes, which was still slight after an hour and a half; that with 4 cc. behaved like that with 5 cc., except that, on adding the nitrite solution, a large, permanent precipitate formed, whereas all the other solutions remained very nearly clear. On standing over night, in the stoppered bottles, the difference between the various portions was much more pronounced, for from those to which 4, 5, and 6 cc. of acid had been added, an opaque yellowish jelly had separated, above which was a clear blue jelly, whereas the portion with 6.5 cc. formed a thin blue jelly containing but little of the

opaque substance, and presented a wholly different appearance from those with 6 cc. and less.

From this it would appear that this edestin preparation combined with the hydrogen ions contained in 6 cc. of decinormal hydrochloric acid more firmly than with those contained in the larger quantities. If we add the acid originally combined with. the edestin, we may conclude that the hydrogen ions equivalent to 7 cc. of the acid were more firmly combined with the protein than those contained in the larger quantities.

5. Solubility of Edestin in Hydrochloric Acid.

Having found that edestin forms a water-soluble salt with hydrochloric acid, I undertook to determine the amount dissolved by definite quantities of this acid. To do so, it was necessary to make a preparation which should be as neutral as possible to phenolphthalein, free from any of the hydrolytic derivative of edestin, mentioned on page 56, and as free as possible from ash.

This is accomplished by extracting oil-free hemp-seed meal with 3 per cent. sodium chloride brine, previously heated to 60°, to which is added enough saturated baryta solution to render the extract neutral to litmus, the requisite quantity being determined by a preliminary experiment with 100 grams of the meal. It is important to avoid an excess of baryta, since otherwise, compounds of edestin with basic constituents of the seed seem to be forined, which are difficult to get rid of afterwards.

The hot extract is strained on coarse cloth and the residue pressed. The very turbid extract is thrown on large paper filters and allowed to stand at rest for about two hours. During this time a part of the extract filters through and the residue settles in the funnels so that about two-thirds of the solution can be drawn off as a turbid liquid, which, however, contains but little suspended matter. This is filtered by suction on thick felts of filter pulp on perforated porcelain plates placed in large funnels, all being previously washed with 3 per cent. salt solution heated to 70°. By thus filtering, this part of the extract may be readily obtained perfectly clear and the filter be washed with hot dilute salt solution, within two hours, 2 liters being passed through each filter. During this time, most of the residual extract passes through the paper filters, so that what remains can be rejected without serious loss. The clear extracts are united in a large

glass-stoppered bottle and allowed to stand over night and cool to 5° or less. The edestin separates as a dense deposit of crystals, from which the solution is siphoned and thrown away, since very little more can be obtained from it by further dilution and cooling. The crystallized edestin is next dissolved in 10 per cent. salt solution, best by adding a volume of 20 per cent. solution equal to that of the mother-liquor remaining with the crystals after siphoning off the greater part. Enough 10 per cent. salt solution is then added so that the solution contains about 8 per cent. of edestin, since stronger solutions, under the subsequent treatment, do not, as a rule, yield well crystallized products. This saline. solution of edestin is heated to 50° and gradually diluted with two volumes of water, at the same temperature, whereby a perfectly clear solution results. By again cooling to 5° the edestin is recrystallized. By repeating this process a very pure crystalline product is obtained, which is again dissolved in enough 10 per cent. sodium chloride brine, free from carbonic acid, to make an 8 per cent. solution of edestin. An aliquot part of this solution is neutralized to phenolphthalein with decinormal sodium hydroxide solution, and the quantity of alkali necessary to neutralize the whole is determined. The edestin solution is then heated to 50° in a glass-stoppered jar, and twice its volume of carbonic acidfree water at the same temperature and containing 4 or 5 cc. more than the calculated quantity of the decinormal alkali is gradually added. The mixture, carefully protected from carbonic acid, is allowed to cool during the night to 5°, the nearly clear solution siphoned off, and the crystalline precipitate collected on a piece of Schleicher & Schüll's thick, hardened filterpaper placed on a perforated plate. The precipitate, consisting of crystals, is very quickly sucked almost dry and washed two or three times with 1 per cent. sodium chloride solution, cooled to o°, and free from carbonic acid, then three times with carbonic acid-free water, ten times with 70 per cent. alcohol, and many times with absolute alcohol, all the wash-water and alcohol being at oo. It is necessary that the washing should be complete and the dehydration with absolute alcohol thorough, so that on drying over sulphuric acid no water should be left after the alcohol has gone off, which would convert a part of the edestin into the insoluble edestan.

Owing to the physical state of the crystalline edestin, the fil

tration, washing and dehydration of 50 to 100 grams can be accomplished within twenty minutes. Preparations made in this way were either neutral or very nearly neutral to phenolphthalein, completely soluble in salt solution, contained not more than 0.02 to 0.03 per cent. of ash, and consisted of fine powders, free from lumps, which can be uniformly suspended in water and almost instantly dissolved by the requisite quantity of acid, alkalies, or salts.

It is very difficult to keep the edestin from combining with minute quantities of carbonic acid, since during the final filtration and washing, a brief exposure to the air is unavoidable, without employing elaborate and cumbersome apparatus. As a result, from 1 to 2 cc. of centinormal alkali were required to neutralize one gram of most preparations thus made.

This process is given in detail, as I was unable to prepare edestin suitable for the experiments next to be described, until I had worked out this method of preparation in all its details.

A series of gram portions of preparation 28, made as above described, was suspended in glass-stoppered bottles, in enough water to make a final volume of 20 cc. with the acid subsequently added. To one portion no acid was added, to the next 2 cc. of centinormal hydrochloric acid, to the next 3 cc., and then I cc. more to each succeeding portion, up to 14 cc.

A second, exactly similar series, was also made, commencing with 6 cc.

After frequently shaking the contents of the bottles for about two hours, they were allowed to stand at rest for two hours longer until the suspended matter had practically all deposited. From each solution 10 cc. were drawn out with a pipette, the acidity of each such 10 cc. determined with centinormal potassium hydroxide solution, and then all separately evaporated to dryness on a water-bath and the residues dried to constant weight at 110°. In this way the results given in the following table were obtained :

TABLE XII.-EDESTIN DISSOLVED BY

A CENTINORMAL SOLUTION OF HYDROCHLORIC ACID.

The acidity of the solutions and of the residues of Series I was determined by titration with centinormal alkali and phenolphthalein, with the following results:

TABLE XIII.—DISTRIBUTION OF ACID BETWEen the DisSOLVED AND UNDISSOLVED EDESTIN IN Terms of CeNTINORMAL ACID.

The degree of acidity of the dissolved and undissolved edestin, —that is, the amount of centinormal alkali neutralized by 1 gram of the edestin chloride contained in the solution and residues of each of these portions,—is given in the following table :

TABLE XIV. THE ACIDITY CORRESPONDING TO ONE GRAM OF THE DISSOLVED AND UNDISSOLVED EDESTIN Chloride of TABLE XIII.

From this table it appears that 1 gram of the substance in the solutions to which from 7 to 12 cc. of acid had been added, neutralized nearly the same quantity of centinormal alkali as that calculated for a compound of one molecule of edestin with 2 molecules of hydrochloric acid, assuming edestin to have a molecular weight of about 14,500, namely 13.8 cc.

rect.

Since the amount of edestin in many of these portions was small, most of these determinations are only approximately corAs already stated, one-half of each of the solutions in these experiments was drawn out with a pipette and the acidity and quantity of dissolved matter was determined as shown above. In order to more accurately determine the acidity of the dissolved edestin by using a larger proportion of substance, the remainder of those solutions of Series I to which from 7 to 13 cc. had been added, was decanted from the undissolved edestin, united and