OP ARTS AND SCIENCES. 45 



an idea of the value of these results in determining the variation of the 

 viscosity with the temperature. We cannot say from them, wliether this 

 variation is proportional to the first or second power of the absolute 

 temperature. Even the results published in the tifth paper, which was 

 to determine this law, are insutHcient. In the first series of these re- 

 sults, shown upon the curve by the extreme lines NS and TS, we see 

 that the exponent representing the law of variation with the tempera- 

 ture varies from x= 0.2 1 for line NS to x=.[).i)9 for line TS, a 

 variation even greater than in the results previously discussed. All 

 the other observations give points intermediate between xVund T, The 

 second series furnishes little better data; and the tliird series, from 

 determinations witii oscillating plates, are not sufficiently complete for 

 discussion in tliis way. They, however, afford no greater satisfaction. 



Puluj has used the method of transpiration for some measurements 

 of this law, and his results appear in the Sitz'ber. Wien. Acad, of 

 1874, Ixix., 287. The results which he has obtained appear rather 

 more concordant than those of Meyer, but still siiow considerable 

 disagreement. Upon the above cut, the lines OP and Q show the 

 extremes of these results as obtained by a discussion of his experi- 

 ments. These lines do not represent the greatest variations between 

 successive results in the same series, but the extreme variation between 

 the mean results of various series. For OP, x= O.Go ; for OQ, 

 x = 0A7. It will thus be seen that these results are more concordant 

 than the different series of Meyer : they are not, however, completely 

 satisfactory. 



Later than these we have a brief notice of some experiments by von 

 Obermayer, in the Phil. Mag., xlix., 332, 1875, in which he states 

 that he has ol)tained results " which confirm those of JNIeyer's experi- 

 ments in a perfectly satisfactory manner." He states Meyer's results 

 as furnishing the exponent | for the variation of q with the absolute 

 temperature ; whence we must conclude that this number expresses the 

 result at which he has arrived. 



What value now are we to place upon these results, and which is 

 the true one ? Maxwell has given a: = 1 ; Meyer, x =z ^ ; Puluj, 

 a; = f ; von Obennnyer, x = |. The first two values, x =1 and 

 a:=:|, we can hardly accept as certain, from the consilerations pre- 

 viously shown. Tlie value given by Puluj of x=5 is undoubtedly 

 somewhat greater than is warranted by his results. Of the remaining 

 experiments we cannot judge, since they have not yet appeared in full, 

 so far as I have been able to ascertain. 



The importance of this question in its bearing upon the kinetic 



