Intelligence and Miscellaneous Articles. 293 



than the transverse one, yet it increased more rapidly than the 

 latter with increasing strength of the magnetic field (probably in 

 proportion to the square). 



For a further investigation we examined more minutely a 

 bismuth plate which was provided with eight equidistant electrodes 

 arranged on the periphery of a circle; all the electrodes were 

 within the homogeneous magnetic field. 



If we denote their positions byK, S., E., W., N.E., S.W., S.E., 

 N.W., the heat current flowed from west to east, and the thermo- 

 magnetic effect was observed between two opposite electrodes for 

 two different intensities of the magnetic field (2480 and 4320). By 

 careful regulation of the flow of heat the galvanometer showed pretty 

 regular deflections. 



When the thermal stream-lines in the plate were exactly in the 

 W.E. direction, by connecting the N. and S. with the galvanometer 

 the purely transversal, with the E.W. the longitudinal, and with 

 the N.E., S.W., and S.E., N.W., the effect resulting from these 

 components was obtained. As, in reality, the above condition is 

 not accurately fulfilled the result of a transverse and longitudinal 

 effect is always obtained, which is seen in the unequal intensities 

 of the currents observed when the direction of the magnetic field is 

 reversed ; each, individual effect may, as is obvious, be separately 

 calculated. We find thus for the transverse (r) and loDgitudinal 

 (X) effects the following values : — 



Strength of Eield 2480. 



N.S. E.W. N.E., S.W. N.W., S.E. 



r 142 3 90 96 



A 3 13 13 11 



Strength of Eield 4320. 



r 245 5 153 163 



X 7-5 42 38 38 



While the ratio of the strengths of field is 1'74, the ratio of the 

 transverse effects is respectively 1*73, 1*70, 1*70, 1*70 ; that of 

 the longitudinal on the contrary 2*50, 3-23, 2-92, 3*45 — in the 

 mean verv near the ratio of the square (3*03) of the strength of 

 field. " • 



We endeavoured to obtain actions of this kind in plates of other 

 metals. Hitherto this has only succeeded distinctly with antimony, 

 nickel (two specimens) cobalt, iron (two specimens) and steel. No, 

 or at any rate uncertain effects are observed with copper, zinc, 

 aluminium and palladium. The direction of the transverse current is 

 the same with antimony, nickel, and cobalt as with bismuth ; with 

 iron and steel, on the contrary, it is reversed, but in all the action is 

 considerably weaker. As to the longitudinal effect which must in 

 every case exist, we are wanting in certain materials of observation. 



We compared equal large plates of antimony, nickel, and 



