Maximum Density of Water. 391 



the curve for cooling should correspond with the temperature 

 of maximum density of water. 



We see then that, allowing for the effect of temperature, 

 ordinary molecules of water have the greatest density, aggre- 

 gated molecules have a lesser density, and ice, which must 

 contain molecules in a still greater state of aggregation, has 

 the least density. 



Raoult, from his experiments on the molecular lowering of 

 freezing-points (Ann. Cliim. Phys. (6) ii. p. 66), concluded that 

 while solvents such as acetic acid and benzene had molecular 

 weights expressed by their ordinarily received formulae, water 

 had a molecular weight four times as great; that is, its mole- 

 cular formula was (H 2 0) 4 . From analogy with other liquids, 

 as acetic acid and hydrofluoric acid, and from the fact that at 

 ordinary temperatures water is liquid, while the similarly 

 constituted hydrides of sulphur, selenium, and tellurium are 

 gaseous, it is probable that when water-vapour condenses 

 it undergoes molecular aggregation. Water molecules from 

 100° C. to about 4° C. will probably, therefore, have the 

 formula (H 2 0) 2 ; and below 4° 0. these will condense to (H 2 0) 4 

 molecules. In constitution these condensed molecules may 

 be considered merely as molecular compounds, or, what is 

 really more simple, as atomic compounds linked thus : — 



H\ Q _ /H , H'Y Y V H 

 H /U-<-\ H and | \ , 



H /J-o XH 



with oxygen in the tetravalent state. 



We must now examine whether the evolution of heat 

 taking place when water is cooled to 4° C, that is to say, 

 the increased specific heat of water at this temperature, shows 

 itself in the determinations of observers on the latent heat of 

 fusion of ice and the specific heat of water. Regnaiilt has 

 determined the specific heat of water from 0°-40° C, 0°-80° C., 

 0°-120° C, &c , but not for smaller temperature-intervals than 

 these. If, however, these results are calculated out for 20° C, 

 60° C, 100° C, &c, and a curve plotted out, it will be found 

 that the specific heat of water for the interval 0°-40° C. is 

 appreciably greater than we should be led to suppose from 

 the specific heats at higher temperatures. 



A. W. Velten (Annalen, 1884, p. 61) has made two series 

 of determinations of the specific heat of water. By the 

 method of mixing, the specific heat was determined for dif- 

 ferent intervals from 4°-7 C. to 100° C. For the interval 



