Magnetic Influence of Solar Light, 373 



being covered with a non-conducting material, the other 

 freely exposed to the heat of the sun ; it is well known that 

 such differences do give rise to thermo-electric currents, 

 and these have been rendered perceptible under such cir- 

 cumstances as to make it far from improbable that in ex- 

 periments like the present, their influence may have been 

 felt. It has been found that even with small pieces of 

 common clay, unequally moistened and exposed to the heat 

 of the sun, electric currents have set in directions from the 

 hotter to the colder portions, and the mutual re-action of 

 electricity in motion and magnetic forces, is too well known 

 as the source of some of the most interesting and important 

 phenomena in the entire range of science, to require to be 

 more that merely adverted to. 



These different considerations on the causes that may 

 interfere with the results of observations made after the 

 preceding method, have so completely destroyed my confi- 

 dence in it, that I consider it unnecessary to detail the 

 remaining experiments, in which it had been employed. I 

 may however state that their results were uncertain and 

 capricious, occasionally tending to prove, and occasionally to 

 disprove, the theory of the magnetic influence of light. 



In the subsequent tables of experiments, the same method 

 by oscillations has been employed as in the preceding 

 sections, and as the sources of error are almost wholly elimi- 

 nated by it, the indications afforded are decisive and satis- 

 factory. The influence of unequal temperature in the 

 portions of the cylinders exposed to the sun's action still 

 operates, and perhaps to this source may be attributed the 

 very minute indications of variation of intensity some of the 

 tables exhibit. As this however never amounts to more 

 than a single second, and is exhibited in but few instances, 

 the inference is undecisive, and the observed difference may 

 be owing merely to some trifling alteration in the conditions 

 of experiment, or to a minute error of observation. 



