392 On the Maximum Density of Water. 



4 o. 7 _ 42 o. 9 q the spec ifi c heat was found to be *9943; for the 

 interval 7°*38-41°-l C. it was *9863 : again, for the interval 

 8°-24-93°*6 C. it was -9873, and for the interval 1 l°*78-93°*0 C. 

 it was "9837. Several other determinations show r ed that for 

 temperatures near to 4° C. the specific heat was greater than 

 the normal. By means of the ice-calorimeter the specific heat 

 w^as determined for various intervals from 0° C. to 100° G. 

 These results show that the specific heat is about 3 per cent, 

 greater from 0°-7°*31 C. than from 7°*31-10°*81 C., and at 

 higher temperatures. There seemed also to be another slight 

 increase in the specific heat at about 18° 0. ; but the mixing 

 experiments, in which it is possible to attain greater accuracy, 

 do not sIioav this, so it is probably not the case. If it were 

 so, the curves for the cooling of water would also show an 

 irregularity; they w T ere, however, quite regular at this point. 



As the latent heat of fusion of ice is determined by mixing 

 it with water of known temperature, an appreciable difference 

 should be observed in the values obtained, according as the 

 temperature of the mixture is near 4° C. or considerably re- 

 moved from it. In both Regnault's experiments and those of 

 Provostaye and Desains (Ann. Chim. Pliys. 1843, pp. 18 and 

 27), the temperature of the mixture varied from 7°*2 C. to 

 18° C; but no regular variations in the values for the latent 

 heat appear. The values vary among themselves from 78*74 

 to 79*26 units in Regnault's experiments, and from 78*75 to 

 79*46 in those of Provostaye and Desains; so the variation 

 might be masked by experimental error. In Person's deter- 

 minations the temperature of the mixture was about 4°*8 C, 

 and the value 80 was obtained. This seems to show that the 

 specific heat of water is increased at about 4° C, as though this 

 heat is evolved in all the determinations ; yet in those in which 

 the mixture had a higher temperature, this heat was spread 

 over a greater interval, and so would not produce so appre- 

 ciably great an effect. 



We see, then, that both the specific heat and the heat of 

 fusion of ice determinations confirm the results obtained by 

 cooling. There can be, therefore, no doubt as to the molecular 

 change taking place at 4° C. The method of cooling affords 

 a simple and accurate means for confirming any supposed irre- 

 gularities in the specific heat of a liquid ; and, as the above 

 results show, in the density of a liquid as well, provided it is 

 not stirred during the cooling. 



The University Laboratory, Oxford, 



