390 FKAGMENTS OF SCIENCE. 



ether and those of the molecules on which they im- 

 pinge. 



According to this view transparent and colourless 

 substances owe their transparency to the dissonance 

 existing between the oscillating periods of their atoms 

 and those of the waves of the whole visible spec- 

 trum. From the prevalence of transparency in com- 

 pound bodies, the general discord of the vibrating 

 periods of their atoms with the light-giving waves 

 of the spectrum, may be inferred ; while their synchro- 

 nism with the ultra-red periods is to be inferred from 

 their opacity to the ultra-red rays. Water illustrates 

 this in a most striking manner. It is highly trans- 

 parent to the luminous rays, which proves that its 

 atoms do not readily oscillate in the periods which 

 excite vision. It is highly opaque to the ultra-red 

 undulations, which proves the synchronism of its vibra- 

 ting periods with those of the longer waves. 



If, then, to the radiation from any source water 

 shows itself eminently or perfectly opaque, we may 

 infer that the atoms whence the radiation emanates 

 oscillate in ultra-red periods. Let us apply this test 

 to the radiation from a flame of hydrogen. This 

 flame consists mainly of incandescent aqueous vapour, 

 the temperature of which, as calculated by Bunsen, 

 is 3259 C., so that, if the penetrative power of 

 radiant heat, as generally supposed, augment with the 

 temperature of its source, we may expect the radia- 

 tion from this flame to be copiously transmitted by 

 water. While, however, a layer of the bisulphide 

 of carbon 0*07 of an inch in thickness transmits 72 per 

 cent, of the incident radiation, and while every other 

 liquid examined transmits more or less of the heat, a 

 layer of water of the above thickness is entirely opaque to 

 the radiation from the hydrogen flame. Thus we establish 



