472 REPORT— 1888. 



is exhibited in the case of iodine, whose two-atom molecules undergo by- 

 heat dissociation, which is gradually brought to completion after passing 

 through more than 300° under ordinary conditions. 



When bodies in the state of vapour are heated and found to have, not 

 far from their boiling-points, vapoar-densities corresponding to definite 

 molecular weights, we have in by far the majority of cases no dissociation, 

 and these are the molecular weights of the compounds ; the contrast 

 between these cases and the preceding is in the short temperature-range 

 from the boiling-point at which a constant (relative) vapour-density has 

 been arrived at. 



It is the force of general considerations such as these that makes one 

 hesitate, in cases such as that of sulphur, antimony, aluminium chloride, 

 ferric chloride, in which the simpler formula So, Sb., (?), AICI3, FeCl3 are 

 not attained until a temperature has been reached very far above the 

 volatilising or boiling-point, before assuming that we are simply dealing 

 with vapours which do not attain the approximately gaseous state for an 

 extraordinarily long interval above the temperature at which they boil or 

 volatilise. 



Although positive proof may be absent of the existence of more 

 complex molecules, it may be that at intermediate temperatures there 

 exists, mixed with the simpler molecules, a larger proportion of more 

 complex molecules as the temperature sinks from the point at which the 

 simpler molecules exist alone. 



The alternative to some such supposition is that Gay-Lussac's law, 

 as a purely physical law, has a few recently discovered very remarkable 

 exceptions in the case of bodies, some of which have not very high 

 boiling or volatilising points ; that certain bodies, in fact, expand for a 

 very long range of temj^eratures much more rapidly than Gay-Lussac's 

 law requires. 



In the Bakerian Lecture, May 1887, Professor J. J. Thomson gives an 

 account of the effect of passing the electric discharge, by sparks or by 

 the silent discharge, through the vapours of iodine, bi-omine, nitric 

 peroxide, and chlorine. An abstract of this paper is given in the ' Proc. 

 Roy. Soc' 42, 1887, p. 343. Among the results obtained is that in a 

 tube exhausted of air and containing vapour of iodine, maintained at a 

 temperature 214° the vapour-density of the iodine is diminished from 

 130 (H=:l) to 84; this result, which was a permanent change, lasting at 

 least for several hours, was produced by direct heating by Crafts and 

 Meier at a temperature estimated by them at 1400°.' The effect of 

 sparking bromine vapour is transient, but while it lasts produces a con- 

 siderable increase of pressure, due, as Professor Thomson thinks, to 

 a transient dissociation of bromine molecules. The vapour-density 

 determinations in these experiments ' showed that bromine vapour is 

 dissociated if it is heated for a long time at a low pressure, even though 

 the temperature is not very high.' He makes the remark — which seems 

 to be of great importance, in determinations of vapour-densities, in all 

 cases where molecules may be undergoing dissociation — that ' for experi- 

 ments on vapour-density the gas should be maintained at a constant 

 temperature for some time before the experiments are made.' The 

 neglect of this precaution may not unlikely account for a number of 

 discrepancies between different experimenters, or the same acting under 



' C.R. 90, 1880, p. 690. 



