MOLECULAR AGGREGATION 257 



Hence, if there is any change in the absorptive 

 coefficient of radiation of any particular wave- 

 length, that change would be due not to any pre- 

 existing periods in the gas, but to others brought 

 into being by new molecules or atomic groupings, 

 due, most probably, to chemical or unstable semi- 

 chemical interactions. 



Thus it is by no means true that a body when 

 cold absorbs the radiation which it emits when hot 

 and it is only for the sa/me temperature that the 

 absorptive and emissive spectra correspond. 



M. Gouy has made a very interesting and im- 

 portant experiment on the spectra of coloured flames, 

 to determine their transparency to the rays which 

 they emit. 



He found by a very simple method doubling the 

 the thickness of the flame and measuring the 

 brightness of the line with a spectral photometer, 

 that the brightness is not doubled, from which it 

 may be inferred that there is an appreciable ab- 

 sorption accompanying the emission. 



He found also that absorption set in as soon as 

 the line had attained a certain brightness, and was 

 careful to note that it was only those parts of the 

 flame which emit the light whose absorption was 

 studied. 



. Iodine vapour heated to redness in a tube be- 

 comes luminous. The spectrum appears at first sight 

 continuous, but on close examination is found to 

 consist of difiuse bands and flutings which corre- 

 spond to those of the absorption observed at the 



