114 



<»1' t'orinula ( I l.O) ) in licinid wnlcr (.-i dilixdiol of fdriniila 

 (II-())j): ;i1 tlic solution would he s;il iiralcd and ice 

 would prccipilnlc 'Plic nice) lan ism of inil ialion ol' crxslal- 

 li/alion assumed ahoN'c is consistcnf willi this llicoi'v oi' 

 water Ix'ini;' a solution as well as with llic usually accepted 

 view tliat water is a definite coinijound. 



A new lii>hl lias Ix'cii tlirown on tlie sn])ject !)>■ the inves- 

 tii>ati()ns of Barnes ("Tec Eng-iiieerini;', " Montreal, lf)28, 

 ]). 7), who attempted to observe under the microscope the 

 formation of ice in water. He states that : "The particles 

 of ice as soon as they can be seen are devoid of crystal 

 form and a])peai' as a true colloid in small disc like particles. 

 These tioccnlale and t>i'ow, passing through a crystal col- 

 loidal form to a true ice crystal." {(\f. F\g. 6). More 



Fig. f). Disc-like jiarticlcs of (•olloidnl ivv. (From Barnes, 1928.) 



experimental investigations in that direction might prove 

 highly significant for the knowledge of the mechanism of 

 initiation of crystallization. 



Bernal and Fowler (1933) have suggested that liquid 

 water might possess a quasi-crystalline structure. Five 

 molecules would hook up together to foi-m a tetrahedral 

 figure (centered tetrahedron) ; these tetrahedra would ag- 

 gregate, according to a definite pattern, in a mass of whaler. 

 It is evident that, if the structure of liquid water is such, 

 our notions of the ])assage from the liquid to the crystal- 

 line state, and in particular of initiation of crystallization, 

 should be radicallv modified. 



B. THE FREEZING POINTS 



Most of the freezing point determinations of biological 

 material have been undertaken for the information that 



