388 HENRY A. EOWLAND 



found the mean specific heat between 100 C. and 16 C. to be 1-00709 

 and 1-00890 times that at about 14. 



But the principal experiments on the subject were published by 

 Eegnault in 1850, 23 and these have been accepted to the present time. 

 It is unfortunate that these experiments were all made by mixing water 

 above 100 with water at ordinary temperatures, it being assumed that 

 water at ordinary temperatures changes little, if any. An interpolation 

 formula was then found to represent the results; and it was assumed 

 that the same formula held at ordinary temperature, or even as low 

 as C. It is true that Eegnault experimented on the subject at 

 points around 4 C. by determining the specific heat of lead in water 

 at various temperatures; but the results were not of sufficient accuracy 

 to warrant any conclusions except that the variation was not great. 



Boscha has attempted to correct Eegnault's results so as to reduce 

 them to the air thermometer; but Eegnault, in Comptes Rendus, has 

 not accepted the correction, as the results were already reduced to the 

 air thermometer. 



Him (Comptes Rendus, Ixx, 592, 831) has given the results of some 

 experiments on the specific heat of water at low temperatures, which 

 give the absurd result that the specific heat of water increases about 

 six or seven per cent between zero and 13! The method of experi- 

 ment was to immerse the bulb of a water thermometer in the water 

 of the calorimeter, until the water had contracted just so much, when 

 it was withdrawn. The idea of thus giving equal quantities of heat 

 to the water was excellent, but could not be carried into execution 

 without a great amount of error. Indeed, experiments so full of error 

 only confuse the physicist, and are worse than useless. 



The experiments of Jamin and Amaury, by the heating of water by 

 electricity, were better in principle, and, if carried out with care, would 

 doubtless give good results. But no particular care seems to have 

 been taken to determine the variation of the resistance of the wire 

 with accuracy, and the measurement of the temperature is passed over 

 as if it were a very simple, instead of an immensely difficult matter. 

 Their results are thus to be rejected; and, indeed, Eegnault does not 

 accept them, but believes there is very little change between 5 and 25. 



In PoggendorfFs Annalen for 1870 a paper by Pfaundler and Platter 

 appeared, giving the results of experiments around 4 C., and deducing 

 the remarkable result that water from to 10 C. varied as much as 



"Pogg. Ann., Ixxix, 241; also, Rel. d. Exp., i, 729. 



