226 Royal Society : — Messrs. Stewart and Tait on the 



But in cases where the gas-effect was considerable, this peculiarity 

 disappeared. Thus we have for an atmosphere of -fo hyd. + ^% air : — 



F ^^ Ck ^ S V m Tl E ^ct at first: Effect one minute after :: 1-05 : 1. 

 of 2 observations) J 



For thin disk (mean 1 , i • • 1 -45 • 1 



of 2 observations) J 



Also, for an air-vacuum of ^-, we have 



For thin disk (mean | Efeectatfirgt . E ff ecto ne minute after :: 1-49 : 1. 

 or 3 observations) J 



In all the experiments with aluminium disks the needle of the 

 galvanometer vibrated 20 times in 20 seconds, and the rotation 

 consisted of 30 turns of the handle in 40 seconds. 



Without pretending to explain all these experiments *, it must, 

 we think, be concluded from them that there are two effects which 

 are differently distributed over the particles of the disk. 



45. The relation between the effect for -fa hyd. and that for ^ 3 - 

 hyd. + -$y air for the thin aluminium disk (galv. 20 in 20 seconds) 

 is given as follows : — 



Effect for -^ hyd. (mean 1 # f Effect for -$$ hyd. + ¥ V air 1 . . aq . o98 

 of 22 observations) \ (mean of 10 observations) J 



In the preceding experiments the centre of the pile was at a dis- 

 tance (along the radius) of 1*5 in. from the rim of the disk. In 

 the following experiment it was adjusted so as to be at a radial 

 distance of 2*9 in. from the rim of the same disk. With this alte- 

 ration and with the thin aluminium disk we obtained 



Effect for ^j- hyd. (mean 1 # f Effect for -^ hyd.-f -^ air \ . . 07 . 1 75 

 of 7 observations) \ (mean of 9 observations) J 



We thus see that the effect for -^ hyd. (which may be supposed 

 to represent the residual effect) and that for -^ hyd. +^5- air (which 

 may be supposed to represent the gas-effect) are both diminished 

 in very nearly the same proportion, namely 100 : 77, by the above 

 transference of the pile to a position nearer the centre of the disk. 

 It would thus appear that the difference in distribution among the 

 particles of the two effects brought to light in art. 44 cannot be 

 explained by a difference from centre to rim, but would rather seem 

 to be due to a difference in depth. If this appear to be improbable, 

 it must be remembered that these experiments were not made on a 

 naked metallic disk. These experiments would therefore appear 

 to show that in an aluminium disk covered with varnish, as well as 

 in a disk of ebonite, we may imagine the residual effect to be more 

 deeply seated than the gas-effect. 



46. We venture on the following as what appears to us to be the 

 most probable explanation of the whole body of experiments, inclu- 

 ding those with radiation. 



(1) There is a temporary heat or cold effect (art. 26) which 



* May not this peculiarity of the aluminium disk iu the hydrogen vacuum 

 be due to some hygrometric surface-effect resident in the varnish? In one 

 experiment, but only one, the peculiarity Avas absent. 



