on Chlorine and Bromine. 291 



action of light on chloride of silver we observe a direct severing 

 of the connexion between Ag and CI, — while the latter, a priori 

 at least, does not provoke any objections. Adopting it for the 

 moment, there remains the open question, whether we are to 

 imagine that the rays of the light actually split up the molecules 

 CI 2 , or only that they loosen their bonds of union, so as to faci- 

 litate a complete separation by the affinity to CI of the foreign 

 atom k. Theoretically speaking, the one kind of action would 

 appear as probable as the other ; in fact I see no reason against 

 their coexistence. But the former one, if it be true, must 

 have some remarkable consequences, which may easily be tried 

 by experiment. Hence I was induced to pursue it more closely. 



I assume that light in general diminishes the force uniting CI 

 with CI, and that occasionally, through the cooperation of light 

 and internal motion (heat), a molecule, CI 2 , is actually split up 

 into its constituents CP + Cl 1 . 



If the chlorine is mixed, say, with hydrogen, it is easily seen 

 that these free atoms will readily combine with the atoms of the 

 latter, and may even (if their number rapidly increase) induce 

 an explosion. But if the chlorine be pure, then the isolated 

 atoms CI will fly about like independent molecules between the 

 undecomposed molecules CI 2 ; occasionally two of them will 

 meet and reunite, so that finally, in case of a constant intensity 

 of light and constant temperature, there will be reached a state 

 of dynamical equilibrium, where in any given moment the num- 

 ber of molecules split up will exactly equal the number of mo- 

 lecules reformed. In other words, insolated chlorine always 

 contains a certain proportion of free atoms (increasing, no doubt, 

 with the intensity of the light) ; and this, together with Avo- 

 gadro's theorem, leads to the conclusion that free chlorine, 

 through insolation, increases in specific volume, the more so the 

 more intense the active portion of the rays falling upon it. 



Moreover it is extremely probable, that the reunion of isolated 

 chlorine-atoms involves a production of heat ; if so, the rays of 

 high refrangibility would do a kind of work which ultimately 

 leads to a (be it ever so small) stationary increase of tempera- 

 ture. Hence the final conclusion to be drawn from our assump- 

 tion is, that chlorine when exposed to " chemical" rays must ex- 

 pand, but when brought back into the dark recontract to its original 

 normal volume. 



This proposition I have tested by experiment and found correct. 



The apparatus used was a Leslie's differential thermometer, 

 which was filled with chlorine and illuminated with various parts 

 of a solar spectrum produced by means of a glass prism. The 

 bulbs were of 5-6 centims. diameter; the connecting tube had a 

 bore of about 1 millimetre. In the first experiments the bulbs 



U2 



