1869.] 



of Thermometers in a Vacuum, 



321 



has charge of the pendulum-experiments in India, communicated to me 

 that the results of some experiments made by him strengthened Professor 

 Meyer's conclusion, and caused him grave doubts as to the necessity of 

 applying the vacuum-correction in pendulum-experiments, one swing often 

 lasting in such experiments from five to eight hours. 



4. It appeared to me that there were various sources of error in the 

 experiments previously made. The only experiments which seem con- 

 clusive are those made by General Sabine with thermometers placed in ice ; 

 but we are not informed in the account of these experiments how long 

 each of them lasted, probably because there was no reason to regard the 

 element of time as of importance. 



In the experiments made by comparing the thermometers with one 

 enclosed in a glass tube and surrounded by air, it is obvious that the ther- 

 mometers under comparison are throughout under different circumstances 

 as regards their sensitiveness, and that this difficulty cannot be entirely 

 overcome by allowing some time for the equalization of the original effect 

 of the exhaustion. Again, it is questionable whether the glass tube which 

 surrounds the thermometer which must be considered the standard of 

 comparison has not, during the process of being closed up before the blow- 

 pipe, been so heated that the remaining air, instead of representing the 

 pressure of a whole atmosphere, is really of a much less density. Further, 

 there is the question of time, raised by Professor Meyer and Captain Basevi. 



In Professor Meyer's experiments one thermometer was placed within 

 the receiver, and another suspended outside, on the exterior of the receiver 

 itself. He found that exhaustion lowered the reading of the former by 

 from one-half to one degree of the centesimal scale, but that after about 

 half an hour both thermometers agreed again : the readmission of air 

 caused the thermometer in the receiver to rise by the same quantity by 

 which it had previously fallen ; but after the lapse of some time the two 

 thermometers read again alike. This lowering of the mercury on evacu- 

 ation, and rising on readmission of the air, ceased almost entirely when the 

 thermometer was introduced into the receiver immersed in dehydrated 

 glycerine. From these observations Professor Meyer concludes that it is 

 solely the mechanical labour of the air during expansion or compression 

 which produces these fluctuations, and that they do not depend on the vary- 

 ing pressure upon the bulb of the thermometer. This conclusion may be 

 correct as far as the particular thermometer is concerned which Professor 

 Meyer employed, for it will be seen in the sequel that certain thermometers 

 really behave exceptionally ; but it will also appear on examining the ex- 

 periments given in Table II. that two thermometers, under precisely the 

 same external circumstances, and in close juxtaposition, often differ in 

 their readings by half a degree of Fahrenheit's scale, and even more, with- 

 out any assignable cause. "We may obviously infer from this that two 

 thermometers, arranged as in Professor Meyer's experiments, are not strictly 

 comparable when small differences of temperature have to be ascertained. 



