226 Proceedings of the Royal Society of Victoria. 



If the temperature correction should be determined directly as 

 a whole, i.e. effect on metal and on air together, by swinging the 

 pendulum in air at two pretty widely separated temperatures, it 

 is to be remembered that as it is made up of two different parts 

 (effect on metal and effect on air) following different laws, the 

 result will not be available unless some element (say the pressure) 

 be kept constant. The experiment would involve the use of an 

 apartment artificially heated in an equable manner,* unless we 

 were content to wait all the time from one season to another, say 

 summer to winter. The temperature correction so determined for 

 a pressure of say 28 in. would not apply (on account of that part 

 of it which depends on the air) to a pressure of say 3 in. It 

 would seem to be best to correct as best may be for that part 

 which is due to the air so as to get the part which is due to the 

 expansion of the metal. I think the effect of the air can be got 

 well by using a wooden model, and altering the observed effect in 

 the ratio of Mh to M'h', and Mj'M' can be got by weighing, and 

 hj/i by balancing separately the model and actual pendulum on 

 their edges. 



I shall be happy to reply to further enquiries. 



Yours very truly, 



G. G. Stokes. 



* P.S.— With a wooden model the effect of the air is so much larger, the time of swinging 

 EO much shorter, and the expansion of the material by heat so much smaller, that there 

 ■would be little difficulty in rigging up an apartment which would serve quite well enough 

 for that. 



