1896.] 



Problems in Electric Convection. 



343 



merits are not yet finished, but the first, which was made on January 

 22, 1S96, seems to leave no doubt on one point of the investigation. 



A |~|-tube was taken, and at the bend was fixed a plaster of Paris 

 plug about 1*5 cm. thick; in one of the limbs two platinum wires 

 were inserted. The plug was saturated with hydrogen to free it 

 from air; the tube was then plunged into a mercury trough, and 

 fixed upright with the limbs full of mercury. Into the leg (A) with 

 the platinum wires a small quantity of hydrogen was passed, and as 

 soon after as possible another small quantity of a mixture of helium 

 and hydrogen from samarskite was put up the other limb (B) of the 

 fi -tube. 



Immediately after the helium was passed into trie limb (B), spec- 

 troscopic observations were made of the gas in the limb (A) ; D 3 was 

 already visible, and there was no trace of 5015*7. This result seems to 

 clearly indicate that if a true diffusion of one constituent takes place, 

 the component which gives D 3 is lighter than the one which gives the 

 line at wave-length 5015*7. 



Although this result is opposed to the statement made by Range and 

 Paschen, it is entirely in harmony with the solar and stellar results. 

 In support of this I may instance that of the cleveite lines asso- 

 ciated with hydrogen in the chromosphere, and the stars of Group 

 III7, those allied to D 3 are much stronger than those belonging to 

 the series of which 5015*7 forms part. 



IX. " Problems in Electric Convection." By Gr. F. C. SEARLE, 

 M.A. Communicated by Professor J. J. Thomson, F.R.S. 

 Received March 16, 1896. 



(Abstract.) 



The paper contains an investigation into the distribution of electric 

 and magnetic forces which are called into play when some electro- 

 magnetic systems are made to move with uniform velocity through 

 the ether. Maxwell's theory is employed in obtaining the funda- 

 mental equations, and it is found that though the electric and 

 magnetic forces, E and H, have generally no potential, still they can 

 be derived from two functions and O ; the differential equations 

 satisfied by these functions are obtained. The equations are first 

 employed to obtain the solution for a moving point-charge, and the 

 result is identical with that obtained by Heaviside and J. J. Thomson. 

 The mechanical forces F and R experienced by a unit charge and a 

 unit pole respectively when moving forwards with the rest of the 

 system are next investigated, and found to have true potentials, *** 

 and Q, the same functions as those mentioned above. These func- 



