14 Starkweather — RegnauWs Calorie and Our 



Before making use of Regnault's experiments on heat of the 

 liquid it is well to note the error discovered bj \^elten* in 

 them. In Regnault's table the second column gives the weight 

 of cold water in the calorimeter, P ; the third, the weight of 

 hot water, jp ; the seventh, the temperature, T, of the hot 

 water according to the air-thermometer; the eighth, the initial 

 temperature of the calorimeter, t ; the ninth, the final temper- 

 ature of the calorimeter, t x \ the eleventh (t x — t^) corrected; 

 the last, the mean specific heat between t t and T, x. Then we 

 should have, according to Regnault's assumption that at low 

 temperatures the specific heat is practically constant, 



x - *(*-*•) 



x ~ P(T-^) ' 



Now Yelten has found that out of Regnault's forty experi- 

 ments in only five does his value of x agree with this formula ; 

 in twenty-two cases the difference is insignificant, while in 

 thirteen cases there is complete disagreement. Yelten has 

 ascribed this discrepancy to errors in calculation, but it is very 

 strange that erroneously calculated these values of x should 

 bear a good agreement with the others and as ' correctly ' cal- 

 culated by Yelten should not. Sutherlandf has almost con- 

 clusively shown that the error is one in the copying of Reg- 

 nault's numbers into the tables, namely, that Regnault in these 

 thirteen experiments, which belong to a different series from 

 the others, copied a wrong set of numbers for the weight of 

 hot water, and that his values of x are the correct ones. 

 Sutherland's proof is based on considerations of the cubic 

 capacity of the calorimeter, as these very experiments show an 

 unusually large amount of hot water. In fact, as the num- 

 bers stand, there is indicated a total weight of hot and cold 

 water greater than the capacity of the calorimeter. 



The writer has made corrections to Regnault's experiments 

 on the heat of the liquid necessitated by the variation of the 

 specific heat within the calorimeter range, but has found that 

 Ekholm's formula fits as well as any. Although this was 

 formed without the above mentioned corrections, nevertheless 

 the latter are considerably smaller than the variations in the 

 experiments themselves. Also in developing the formula 

 Ekholm left out certain experiments which are those the cor- 

 rections affect the most. The formula is 



h == £ + 0-00000929 * 2 + 0-000000265 t\ 



and is to hold only above 100°. 



The unit here is the specific heat at 15° according to Reg- 



* Wied. Ann., xxi, p. 45, 1884. 

 f Phil. Mag., xxvi, p. 302, 1888. 



