96 DRS. J. A. BARKER AND P. CHAPPUIS ON A 



day for the experiments with thermometer K.9. For the coefficient of expansion we 

 have taken the mean of the determinations given above, viz. : 



= 0-003 668 11. 



The comparisons are numbered 33 to 53 in the table for K.9 at the end. 



For the calculation of all the comparisons with thermometer K.8 we have taken as 

 the initial pressure the value 528747 millims. These comparisons are numbered 

 73 to 91 in the table for K.8 at the end. 



Comparisons of TJiermometer K.2. 



We noticed that during the experiments with K.8 and K.9 above the sulphur 

 point the glass tubes of the platinum thermometers were seriously attacked and 

 showed signs of softening. Some time previously porcelain tubes had been ordered to 

 replace the glass ones, but owing to the delivery of these being inordinately delayed, 

 we were obliged to relinquish the comparisons we had intended to make with K.8 

 and K.9 at higher temperatures and take in their place a low resistance platinum 

 thermometer, K.2, already provided with a porcelain tube. The initial pressure in 

 the gas thermometer was reduced to 391 '88 millims. and a series of 12 comparisons 

 made, the results of which are shown in the table at the end.* 



The constants of the gas thermometer were determined in the usual manner before 

 the comparisons. The value found for the coefficient a was 0'003 667 71. We were 

 prevented by an accident from redetermining this coefficient of dilatation after the 

 measurements. This is to be regretted, as the preceding determinations showed a 

 systematic diminution in its value which we are unable to explain. 



If the coefficient corresponding to an initial pressure of 392 millims. be deduced 



* As a confirmation of the general accuracy of the methods of standardization, &c., adopted in our 

 platinum thermometry, \ve may mention that on the return of the apparatus from France the constants of 

 thermometer K.2 were redetermined at Kew by Dr. CHREE and Mr. HUGO, using the improved Cambridge 

 resistance box, which had just been re-calibrated by them. For the platinum temperature of the sulphur 

 point at 760 millims. pressure they found a value differing only 0- 01 from that got at Sevres, although 

 nearly the whole of the apparatus employed, including the resistance box, leads, and barometer, were of 

 patterns differing materially from those used in France. 



