1899.J on Measuring Extreme Temperatures. 113 



power n of the temperature, though fairly exact over a moderate range 

 of temperature, fails entirely when tested at higher points. The 

 results of Ohermayer appear to be the most accurate for the different 

 gases between 0° and 100° C, but if the same formula is retained, the 

 value of the index n diminishes as the temperature is raised. Taking 

 the average value between 0° and 100° for air as being 0*76, the value 

 falls to 0*70 between 100° and 450°. A result of this nature was 

 found by Wiedemann, but the rate of diminution which he gives 

 appears to be far too great. He gives, for instance, the value n = • 67 

 for air between 0° and 184°, which implies a rate of diminution of the 

 index many times greater than that which actually occurs. It would 

 be very difficult by the method which he employed to make sure of 

 any deviation whatever from the formula over so small a range, and 

 since the error of his determination is much greater than that of the 

 formula, he can hardly be said to have disproved the index law. 



The problem is seriously complicated by the failure of the simple 

 formula ; but since the measurements are capable of great exactitude, 

 and since it is possible to obtain many independent checks by com- 

 paring the results of the two methods of effusion and transpiration, 

 and also by examining the behaviour of different gases, the author 

 is confident of ultimate success. The method of experiment here 

 described has already led to many promising and interesting results, 

 and it is probable that the complete solution of the problem when 

 attained, besides leading to more accurate determinations of extreme 

 temperatures, may also throw light on dissociation and on many other 

 points which are at present obscure in the theory of gases. 



[H. L. C] 



Vol. XVI. (No. 93.) 



