136 Drs. Ramsay and Young on 



all pressures ; and it appeared likely that if the variations 

 from constancy of the reduced values of ~ . t for other sub- 

 stances were real, they would be related in some way to the 

 variations from constancy of the ratios of the absolute tem- 

 perature. This is indeed the case; for if we compare the 



values of -Jj . t for any two substances, and make the tempe- 

 ratures of one of them ordinates, and the ratios of the products 

 -^ . t abscissas, we again find that the points representing the 



dp 



dt 



relation of the ratios of the product ^ . t at definite pressures 



to the temperatures of one of the bodies corresponding to 

 those pressures, fall in a straight line. We therefore obtain 

 a similar equation, 



R'=R+c (*'-*), 



R/ and R in this case representing the ratios of the products 



The cause of the apparently greater deviation from 



constancy at low and at high pressures is simply this — that 

 at low pressures a small change of pressure corresponds to a 

 large change of temperature, while at high pressures the 

 intervals of pressure taken are very great, and therefore the 

 intervals of temperature are also great. This is clearly seen 

 by referring to the table of absolute temperatures corre- 

 sponding to the pressures taken. Taking water as an instance, 

 we have the following numbers : — 



dp 



di' 1 



P. 



t absolute. 



Interval 

 of t. 



P. 



t absolute. 



Interval 

 of t. 



millim. 

 10 

 50 

 100 

 150 

 200 

 300 

 400 

 500 

 600 



284-3 

 311-3 

 324-7 

 3331 

 339-6 

 3489 

 356-0 

 361-7 

 3665 



o 



27-0 

 13-4 

 8-4 

 6-5 

 9-3 

 71 

 5-7 

 4-8 



millim. 



600 



700 



800 



900 



1000 



1500 



2000 



3000 



5000 



366-5 



370-7 



374-45 



377-8 



380-85 



393-2 



402-5 



416-5 



435-85 



o 



4-2 



3-75 



3-35 



305 



12-35 



9-3 



14-0 



19-35 



The determinations of the products ~- . t are, however, 



necessarily less accurate than those of the temperatures, espe- 

 cially at low pressures, for the sources of error are more 

 numerous and have greater influences on the results ; yet the 



