gation. As an illustration of what has just been said, the behaviour of 

 the crystalline globulin, excelsin, from the Brazil-nut may be given. 

 The proteins extracted by dilute ammonium sulphate solution from 

 this seed were subjected to repeated careful fractionation with ammonium 

 sulphate and two fractions obtained I. between 40 and 50 per cent, 

 saturation, II. between 50 and 60 per cent. These fractions were then 

 dissolved in one-tenth saturated ammonium sulphate solution and re- 

 precipitated by dialysis. I. yielded a precipitate consisting wholly of 

 hexagonal crystals which when washed with water and alcohol and 

 dried over sulphuric acid weighed 24-5 grammes ; II. by a similar treat- 

 ment, yielded 5 grammes which consisted of a mixture of imperfectly 

 formed crystals and spheroids. The precipitation limits of these pre- 

 parations were then found to be for I. between 19 and 46 per cent, of 

 saturation, for II. between 21 and 60 per cent, of saturation. Thus 

 preparation II., which before precipitation by dialysis had been entirely 

 freed from all protein precipitable below 50 per cent, saturation, showed 

 a lower limit of precipitation identical with that of preparation I., 

 all of which was originally precipitated below 50 per cent, saturation, 

 and an upper limit of precipitation of 60 per cent, saturation which 

 agreed with its original upper limit. After precipitation by dialysis 

 each of these preparations had, therefore, retained its upper limit of 

 precipitation practically unchanged, but the lower limits were greatly 

 reduced and were the same for each fraction. Although doubt exists 

 as to the extent to which the precipitation limits of proteins can be 

 employed for characterising .them as chemical individuals, nevertheless 

 ammonium sulphate is of great use in effecting the separation of pro- 

 teins from their solutions, as well as from one another. 



Osborne and Harris (351) determined the precipitation limits of 

 preparations of a number of different proteins which had been pre- 

 cipitated by dialysis from sodium chloride solutions and dried over 

 sulphuric acid, after being washed with water and alcohol in the usual 

 way. A weighed quantity of each of the proteins was dissolved in 

 one-tenth saturated solution of ammonium sulphate, the solution 

 filtered clear and 2 c.c. mixed with enough two-tenths saturated 

 sulphate solution to make a final volume of 10 c.c. with the saturated sul- 

 phate solution to be afterwards added. Successively greater quantities 

 of the saturated sulphate solution were used and the points noted at 

 which the solution first became permanently turbid and that at which 

 the protein was all precipitated, as shown by saturating the filtered 

 solution with ammonium sulphate. The results are stated in the 

 following table in cubic centimetres of a saturated solution of am- 



4* 



