Molecular Constitution of Water. 463 



The value 1*08942 for p t the density at zero of one of the 

 pure ingredients agrees with the value indicated by the 

 asymptote in the figure, and the law of dissociation of our 

 standard mixture S at different temperatures is given by (6), 

 to which we shall return when we have determined the com- 

 position of our standard mixture. We shall neglect in it the 

 small terms and write it 



?=rai77- 3207 < 7) 



In this form it gives for </ = l/2, t=— '06° instead of 4° 

 as the temperature at which the percentage of 1 added to S 

 is ; this is the result of neglecting the small terms, and is 

 of no importance in the study of q at present. For q=0, 

 that is for a mixture of equal parts of 1 and S, £ = 147. The 

 largest possible negative value of q is when t = <x) and is 

 — "32, so that obviously the formula does not include in its 

 range the case of q= — 1/2, that is of the complete dissocia- 

 tion of water into the pure substance 1. Nor should the 

 formula be expected to do so, as its empirical range is only 

 to the boiling-point, though it may be used for purposes of 

 approximation up to nearly 200 degrees. 



To determine the composition of S we shall now consider 

 another remarkable exceptional property of water, namely, 

 its large expansion upon solidification. According to Bunsen 

 unit volume of water at 0° becomes 1*09082 of ice at 0°, and 

 notwithstanding this considerable forcing apart of the mole- 

 cules of water against their mutual attractions, for which we 

 should expect that energy would have to be put into water 

 to solidify it, we find actually that energy or heat has to be 

 taken out of water to change it to ice. For comparison we 

 can consider the expansion of metals and phosphorus when 

 they melt. According to Vicentini and Omodei (Wied. Beibl. 

 xii.) a volume 1000 of each of the following increase by 



Pb. 



Cd. 



Sn. 



Na. 



K. 



P. 



Ho- 



34 



47 



28 



25 



26 



35 



37 



Therefore there is on the average an expansion of such 

 solids on melting by 3*3 per cent., whereas with water there 

 is a contraction of 8*3 per cent. These facts alone are enough 

 to suggest that the solidifying of water into ice is no mere 

 physical change of state, as with most melting bodies, but is 

 chiefly a profound chemical change. This view of the melt- 

 ing of ice will be gone into fully in sections 5 and 8, and in 

 the meanwhile we shall assume that water, in solidifying 

 changes entirely into its other pure ingredient, which we 



2K2 



