ESSAY-REVIEWS 133 



the phosphorus content of the egg-albumen (E. G. Willcock and W. B. Hardy, 

 Proc. Chem. Phil. Soc. vol. 14, p. 119, 1907). Further purification was effected by 

 a number of recrystallisations, generally six. The ammonium sulphate used in the 

 crystallisation is removed by dialysis. Egg-albumen as an amphoteric compound 

 binds both acid and basic compounds, particularly the latter, as egg-albumen is 

 more acid than basic. Consequently it is impossible by simple dialysis to remove 

 every trace of the compounds of ammonium sulphate, particularly the ammonia. 

 Sorensen therefore allows the dialysis to progress so that all sulphuric acid is 

 removed, which can be done by addition of ammonia at intervals whereby the 

 sulphuric acid bound to the egg-albumen is transformed into ammonium sulphate, 

 which dialyses away. In this way all sulphuric acid can be completely removed, 

 but it is impossible to remove the ammonia completely by dialysis, because the 

 albumen combines with the ammonia and only the ammonia in excess dialyses 

 and later the ammonia freed by hydrolysis of the ammonium salt. But as it is 

 possible to determine small quantities of ammonia with great accuracy the 

 albumen solution can nevertheless be well defined, and if a quantity of acid 

 equivalent to the remaining ammonia is added to the system there results a 

 system containing albumen and a known small quantity of ammonium sulphate, 

 but neither ammonia nor acid in excess. By addition of acids, salts, or ammonia 

 one can of course obtain systems having very different but nevertheless quite 

 definite composition. 



Egg-albumen in crystallising out contains water, so that the factor by which 

 the nitrogen has to be multiplied is not 6*45 but 7 '86, the quantity of water in 

 a gram of egg-albumen being o"22 gm. 



The most striking and probably the most important part of the results of 

 Sorensen's researches with the egg-albumen prepared in this manner is that 

 dealing with the application of Gibbs' phase-rule to the equilibrium condition 

 between the crystalline egg-albumen and its mother liquid. The generally 

 accepted axiom in colloid chemistry is that the phase-rule does not hold for 

 colloid substances, and probably this is true for suspensoids where the conditions 

 are very complex, but in this respect emulsoids can be regarded in the same way 

 as true solutions. If an egg-albumen solution is precipitated with ammonium 

 sulphate and left until equilibrium is established, one has a heterogeneous system 

 with three components : (1) water, (2) egg-hydrate, and (3) ammonium sulphate, 

 and three phases: (1) solid crystalline egg-albumen, (2) the watery phase con- 

 taining ammonium sulphate and egg-albumen, (3) water vapour. The system 

 should then, according to the phase-rule, have two degrees of freedom — that is, 

 if, for example, temperature and ammonium sulphate concentration were chosen, 

 the system should be completely determined. Or, in other words, with crys- 

 tallisation at a definite temperature and a definite ammonium sulphate concen- 

 tration, there should be a definite concentration of egg-albumen in solution. As 

 this is not the case, as, for instance, is seen from Galeotti's work, it is generally 

 assumed that the phase-rule does not hold for colloidal substances. This false 

 conclusion is due to the fact that the egg-albumen used by earlier workers did not 

 have always the same quantity of acid or ammonia in excess, and the same holds 

 for the content of ammonium sulphate which very seldom contained equivalent 

 quantities of ammonia and sulphuric acid. Consequently, the number of com- 

 ponents is not three, but four — namely, water, egg-albumen, ammonium sulphate 

 and sulphuric acid (or ammonia). With this increase in the number of com- 

 ponents the number of degrees of freedom becomes three, and this is confirmed 

 by all Sorensen's experiments, for, at a definite temperature, definite ammonium 



