30 



Starkweather — Regnaultfs Calorie and Our 



lie near Battelli's with his. Below is given a table of the frac- 

 tional deviation, d, of 7=- according to Battelli from that accord- 

 ing to Ramsay and Young ; in each case the experiments com- 

 pared had very nearly the same values of p and T. Above 

 each value of d is given a number for reference. 



No. 

 d 



1 



+ •009 



2 

 + •006 



3 



+ •020 



4 

 + •021 



5 



— •006 



6 

 + •003 



7 

 + •005 



No. 

 d 



8 



+ •010 



9 



+ '027 



10 

 + •090 



11 



— •017 



12 

 — •013 



13 

 — •016 



14 

 — •015 



No. 

 d 



15 

 — •014 



16 



-•007 



17 

 + •013 



18 

 — •021 



19 

 -•019 



20 

 — •020 



21 

 — •017 



No. 

 d 



22 

 —•012 



23 

 -•020 



24 

 — •023 



25 

 — •030 



26 

 — •012 



27 

 



28 

 + •009 



Nos. 2, 3, 4 were all at one temperature, about 150° C, while 

 5 to 11 were at 180°, 11 to 18 at 200°, and the remainder at 

 230°. . 



The sudden changes in the values of d in 9 and 10 from 

 those in 5 .... 8, in 16, IT from those in 11 . . . 15, and in 

 26, 27, 28 from those in 18 ... . 25, can all be accounted for 

 by condensation, since Ramsay and Young's figures, as will be 

 shown, give marked evidence of condensation as the saturation 

 line is approached, while those of Battelli do not. But the 

 remarkable fact is that in the first eight, with the exception of 

 the fifth, the deviations of Battelli from Ramsay and Young 

 are all positive, the reverse of what is due to the error in the 

 latter already mentioned. Suppose we ascribe this to large 

 condensation in the case of Ramsay and Young. Looking at 

 11 .... 28 we see that the constant error in their experiments 

 must amount to 1*5 per cent, hence in the first eight there must 

 have been condensation to the amount of 2*5 per cent. 



But there are a few experiments below 100° which can be 

 compared, and in that region Ramsay and Young's experiments 

 do not have the constant error already mentioned, the weights 

 being measured directly. These follow : 



Ramsay and Young. Battelli. 



p 45-47 39-95 35'3 67'57 85-32 100*27 



v 26-06 30*15 33-57 17*851 14-1326 12*0335 



T 348-7 348-7 348*7 352-24 352'24 352-24 



pv 

 T 



~ 3-415 



3-454 



3-399 



3 424 



3-423 



3-422 



Now Ramsay and Young observed that there was far more 

 condensation below 100° than above, so much so that they 

 themselves admit that the larger part of their low temperature 



