July 1, 1898.] 



SCIENCE. 



tions as far as possible identical with those 

 anticipated in the experiments with liquid 

 oxygen. 



By this means it was ascertained approxi- 

 mately what degree of accuracy could be 

 expected in the results when liquid oxygen 

 was employed. 



The manipulation in these preliminary 

 •experiments was as follows : 



A piece of copper, with a silk thread at- 

 tached (the identical piece to be used in 

 ihe low temperature experiments), was im- 

 mersed in boiling water and allowed to re- 

 main submerged for several minutes. 



It was then lifted out of the boiling water 

 hy the thread, and conveyed as quickly as 

 possible to a calorimeter two-thirds filled 

 with water at 17°C. The rise in the tem- 

 perature of the water was noted, and the 

 specific heat calculated in the usual way. 

 When the copper was transferred from the 

 vessel containing the boiling water to the 

 calorimeter, it was allowed to strike the 

 edge of the vessel a sharp blow to remove 

 ■drops of water remaining on the surface of 

 the piece of copper. 



It was found, however, that a small 

 :amount of hot water, .03 to .04 of a gram, 

 was alwaj'S carried over to the calorimeter. 

 The quantity was determined experiment- 

 ally and the proper correction applied in the 

 specific heat calculations. 



Five consecutive determinations of the 

 specific heat of copper between 23° C. and 

 the boiling point of water (100° C), by this 

 method, were as follows : 



Table 1. 



Average variation of the five determina- 

 tions from the mean .0006, or .7 per cent. 



Greatest variation of any one determina- 

 tion from the mean, .00138, or 1.5 per 

 cent. 



The weight of water used in each deter- 

 mination was 70.00 grams, and the water 

 equivalent of the calorimeter and thermom- 

 eter, 5.87 grams. 



The weight of copper employed in each 

 determination was 63.493 grams. 



The temperature of the boiling water was 

 determined from the atmospheric pressure, 

 the barometric height being measured at 

 intervals during the experiments. 



. The mean value obtained for the specific 

 heat of copper, .0940, is in agreement with 

 the value generally accepted for that metal 

 for the same range of temperature, the best 

 values ranging from .0933 to .0949. 



Also the percentage error of each deter- 

 mination as compared with the mean value, 

 .0940, is small, being less than 1 per cent. 



These facts seemed to warrant proceed- 

 ing with a series of determinations of the 

 specific heat of copper and other metals at 

 low temperatures by the same method, em- 

 ploying liquid oxygen. 



Five determinations of the specific heat 

 of copper between the boiling point of 

 liquid oxygen, under atmospheric pressure 

 ( — 181.4°C.), and 11 C. were as follows : 



Table 2. 



Copper. 



Determination. 

 Number. 



Specific Heat 

 (-181.4°... 11°C.) 



0867 

 0854 



Mean .0940 



Average variation of the five determina- 

 tions from the mean, .0007, or .8 per cent. 



Greatest variation of any one determina- 

 tion from the mean, .0014, or 1.6 per cent. 



