of the Solar Spectrum. 237 



In order that the reader may be enabled to compare the action 

 of my instrument with that of others, I may here describe some 

 experiments performed with this aim in view. 



The flame (3 inches in height) of a Bunsen gas-jet, without 

 the employment of a reflector, at a distance of 3 decimetres 

 caused a deflection of 32° when radiating upon the quadratic 

 thermo-battery, and a deflection of 11° when shining upon the 

 lineal one — these deflections being shown by the multiplier before 

 described. 



A blackened copper-plate*, heated by a spirit-lamp at a di- 

 stance of 2-5 decimetres, caused the multiplier to show deflections 

 of 43° and 14° when radiating upon the quadratic and lineal 

 thermo-batteries respectively. The solar rays with which I ex- 

 perimented were reflected into the darkened room, through a 

 hole in the shutter of the window, from the metallic mirror of 

 one of Silbermann's heliostats, which was set up outside. 



Concerning the exactness of the following experiments, I have 

 only further to add a few remarks. 



The scale of the multiplier is only divided into intervals of 

 five degrees ; so that the readings off are only exact to i a de- 

 gree at furthest. A greater exactness in the reading off would, 

 moreover, be wholly unreal ; for, in the first place, the position 

 of the needle of the multiplier was never a very definite one, 

 after the source of radiant heat was removed, the variations 

 being about i a degree; secondly, the placing of the lineal 

 thermo-battery in the spectrum was not capable of the same 

 degree of accuracy as the reading off of the multiplier, because 

 the spectram was not pure enough to show Fraunhofer's lines. 



It seems at present in vain to think of showing the existence 

 of cold bands in the spectrum corresponding to Fraunhofer's 

 dark lines ; indeed, our knowledge of the thermal properties of 

 the spectrum is still far behind that of its optical ones in respect 

 to accuracy. 



In the foUowiug pages the results of the observations are so 

 given that the reader may judge of the limits of the errors of 

 observation. I have in all cases avoided giving the results the 

 appearance of an accuracy which they neither do, nor as yet 

 probably can, possess. 



Experiments on the thei'mal rays transm.itted by coloured liquids. 



When I first decided upon experimenting on the distribution 

 of heat in the solar spectrum, I was not in possession of a lineal 

 thermo-battery. I imagined it possible to dispense with one, 

 and to arrive at the conclusions sought by endeavouring to de- 

 termine the action of the thermal rays which had passed through 

 * Sue my ' Physics,' vol. ii. p. fil 1. 5th cd. 



