towards a Dynamical Theory of Solutions. 57 



of acetic acid. The exceptional behaviour with formic acid 

 is traced to its dielectric capacity's having a value 60 near 

 to the 80 o£ water. Molecules of water in formic acid are 

 nearly in the same electrical environment as in waiter, 

 whence the less vigorous action of formic acid. On the 

 whole, the available data for propionic and butyric acids 

 corroborate those for acetic acid, though, so far as they go, 

 they indicate for hydrol the same sort of density as those for 

 alcohol. 



As regards refraction the value of (n 2 2 — l)/(n 2 2 + 2)p 2 for 

 water as a chanomo- substance is calculated from the values 

 for mixtures of water and alcohol on the assumption that 

 alcohol is an unchanging substance. As the amount of 

 alcohol increases from to pi = 0'5, the values diminish to a 

 minimum, in qualitative agreement with the fact that 

 (?i 2 — l)/(n 2 + 2)p has a smaller value for dihydrol than for 

 trihydrol. The minimum value 0-20733 is not so small as 

 0*2(JG74 for pure dihydrol, so that the optical results do not 

 fully confirm onr conclusion from contraction that the change 

 from^i — O to /> x = 0*5 may be considered to consist almost 

 wholly of change of trihydrol into dihydrol. Indeed with 

 the contraction it appeared that rather more contraction 

 occurs than is accounted for solely by change of trihydrol 

 into dihydrol, perhaps a little change to hydrol having 

 occurred. With the optical results we can suppose the 

 difference between 0*20733 and 0*20674 to be accounted for 

 by change of a little dihydrol into hydrol having a larger 

 (n 2 — l)/{n 2 + 2)p. Beyond ^ = 0*5 the results show that, if 

 hydrol is produced, it must have (n 2 — l)/(n 2 + 2)p larger than 

 for dihydrol. But in water of crystallization (n 2 — l)/(n 2 + 2)p 

 has a value only J that for dihydrol. On the other hand, in 

 vapour of water this has for hydrol a value 3 per cent. 

 greater than for dihydrol. Here we encounter again the 

 difficulty that hydrol is a substance of variable properties, a 

 fact which I would ascribe to the variable electric moment of 

 the doublet #p latent in its oxygen atom. Hydrol in excess 

 of alcohol has a value between 0*2107 and 0*2110 for 

 (n 2 — 1)/0* 2 -•- 2)yo and in vapour of water the value 0'2092. 

 In excess of acetic acid it is 0*2101. In the case of acetic 

 acid [n 2 — l)/(n 2 2 + 2)p 2 with increasing p 1 reaches a minimum 

 when p 1 = 0"25, a result in harmony with the view that with 

 acetic acid the changes of trihydrol into dihydrol and of 

 dihydrol into hydrol occur concurrently- On the whole then 

 the optical results confirm the theory advanced to account for 

 contraction. 



With (K — V)Jp for mixtures of water and alcohol we 



