8 4 CORRELATION OF PHYSICAL FORCES. 



body influences indeed, I may say determines its con- 

 ducting power, is by no means explained by the theory of a 

 fluid ; but if electricity be only a transmission of force or 

 motion, the influence of the molecular state is just what would 

 be expected. Carbon, in a transparent crystalline state, as 

 diamond, is as perfect a non-conductor as we know ; while in 

 an opaque amorphous state, as graphite or charcoal, it is one 

 of the best conductors : thus, in the one state, it transmits 

 light and stops electricity, 'in the other it transmits electricity 

 and stops light. 



It is a circumstance worthy of remark, that the arrange- 

 ment of molecules, which renders a solid body capable of 

 transmitting light, is most unfavourable to its transmission of 

 electricity, transparent solids being very imperfect conductors 

 of electricity ; so gases readily transmit light, but are amongst 

 the worst conductors of electricity, if indeed, properly speak- 

 ing, they can be said to conduct at all. 



The conduction of electricity by different classes of bodies 

 has been generally regarded as a question of degree : thus 

 metals were viewed as perfect conductors, charcoal less so, 

 water and other liquids as imperfect conductors, &c. But, in 

 fact, though between one metal and another the mode of 

 transmission may be the same and the difference one of 

 degree, a different molecular effect obtains, when we contrast 



o * 



metals with electrolytic liquids and these with gases. 



Attenuated gases may be, in one sense, regarded as non- 

 conductors, in another as conductors ; ' thus, if gold-leaves be 

 made to diverge, by electrical repulsion, in air at ordinary 

 pressure, they in a short time collapse ; while in highly-rarefied 

 air, or what is commonly termed a vacuum, they remain 

 divergent for days ; and yet electricity of a certain degree of 

 tension passes readily across attenuated air, and with difficulty 

 across air of ordinary density. 



Again, where the electrical terminals are brought to a 

 state of visible ignition, there are symptoms of the transmission 

 of electricity of low tension across gases ; but no such effects 

 have been detected at lower temperatures. All this presents 

 a strong argument in favour of the transmission of electricity 

 across gases being effected by convection in the disruptive 



