136 O. U. VONWILLER. 



being, according to Miss Frances Wick, (Phys. Rev. 1907), 

 about 400 micro- volts at ordinary temperatures; the only 

 substances for which it is numerically greater are tellurium 

 and selenium. The result of the passage of a current across 

 a metal-silicon junction is a cooling or heating at that point, 

 and this in turn gives rise to an e.m.f. opposing the flow of 

 current — an e.m.f. proportional to the current. On account 

 of the differences in area the temperature change, and 

 therefore the potential difference, must be greater at the 

 surface of smaller area. The rectifying property apparently 

 can be explained by the action of the Thomson effect if we 

 assume that for small currents the main bulk of the silicon 

 is not appreciably altered in temperature, but that the 

 heated or cooled portions are limited to thin layers in close 

 proximity to the junctions. In such a case the additional 

 absorption or evolution of heat due to the flow between 

 points at different temperatures is limited to the boundary 

 layers, and the Thomson effect being the same at both 

 junctions, while the Peltier effect is opposite, we see that 

 at one junction a summational effect, and at the other a 

 differential heating effect, should ensue, so that at either 

 junction the opposing e.m.f. should depend on the direction 

 of the current. 



An investigation on these lines shows that the relation 

 between the back e.m.f. and current should be of the form 

 e = ai ±bi 2 . The actual observations show that this is 

 very approximately the form of the relation obtained for 

 small currents. 



Great, however, as are the thermo-electric effects with 

 the materials here dealt with, the results obtained are such 

 that differences in temperature of very many degrees must 

 occur in order satisfactorily to explain the results obtained 

 and it seems that the above explanation is not correct. 



