6 3 o SCIENCE PROGRESS 



including tin, zinc, bismuth, cadmium, copper, silver, lead, 

 and antimony. There can no longer be any doubt but that 

 many metals as we know them are metastable (as has been 

 insisted on by Cohen) and in these cases are a mixture of two 

 allotropic forms such as are postulated in Smit's theory of 

 allotrophy. 



Quite a different type of investigation, though at the same 

 time one dealing with change of state, has been carried out by 

 Walton and Brann (Journ. Amer. Chem. Soc. 38, 317, 1916), 

 who studied the effect of dissolved substances upon the velocity 

 of crystallisation of water. It is well known that a super-cooled 

 liquid tends to crystallise. Tammann was the first to investi- 

 gate this phenomenon fairly closely. He showed that it 

 involved in general three temperature zones. The speed of 

 crystallisation at first increases with increased super-cooling 

 (zone A) ; it then generally remains constant for several degrees 

 (zone B) ; and finally decreases (zone C). The present investi- 

 gation refers to zone A. The velocity of crystallisation was 

 determined at 9 below zero for forty-five substances. It was 

 found that in all cases the velocity was retarded. Solutions of 

 equi-molecular concentration show different retarding effects. 

 The effect is therefore more or less specific. For substances 

 with more than eight atoms in the molecule there is a rough 

 relation between the number of atoms and the retardation of the 

 velocity ; the greater the number of atoms, the slower the speed. 

 That these effects cannot be explained by capillary effects is 

 evidenced, according to the authors, by the fact that the 

 sugars which are inactive in a capillary sense are very active 

 in retarding the rate of crystallisation. Colloidal substances 

 such as gelatine, ferric hydroxide and certain dyes inhibit the 

 rate. Further, the stability of the super-cooled solutions 

 towards spontaneous crystallisation varies with the solute. 

 Decimormal solutions of hydrochloric acid are characterised 

 by being exceptionally stable. As regards the mechanism of 

 the retarding influence in general, it appears that the effect 

 is connected with a shift in the equilibrium which exists between 

 the different polymers of (H 2 0) which constitute liquid water. 

 The idea is that, ice being mainly trihydrol ((H 2 0) 3 ), any 

 decrease in the quantity of this substance present in the liquid 

 state would correspondingly diminish the rate at which crys- 

 tallisation would take place. This is an interesting hypothesis, 



