136 hepoet — 1886. 



these find only a small portion of curve representing a nearly constant 

 vapour-density nearly equal to that required by the molecule I — just 

 enough to convince all who wish to be convinced— and until by higher 

 temperatures there can be shown a longer range during which the density 

 of iodine vapour is always (compared with air) half what it is at ordinary 

 temperatures, that is to say shows no tendency to diminish further, the 

 evidence from the experiments mentioned will be accepted as conclusive 

 only by chemists and physicists who have a predisposition to accept the 

 conclusion. But confirmatory evidence of some fundamental change in 

 iodine at high temperatures is given by the fact that it gives a band spec- 

 trum when subjected to electrical discharges of comparatively low tension, 

 and a line spectrum under higher tension.' 



The results obtained for chlorine and bromine were a diminution of 

 density continuing up to the highest furnace temperatures under which 

 the experiments could be performed ; it is impossible to use a much higher 

 temperature than 1700° with platinum vessels, for platinum melts a little 

 above this — at 1775° according to Violle^ — and is very appreciably 

 attacked by chlorine at a white heat. 



At about 970° stannous chloride was found to have vapour density 

 corresponding to SnClz, the density found about 200° lower corresponding 

 to Sn2Cl4 ; and FcaCIj had density at white heat much diminished ; while 

 the formulae AloCl,;, &c., were in concordance with vapour-densities found 

 at the highest temperature ; HgCla was found again to be the molecular 

 weight corresponding to the vapour-density of mercuric chloride at the 

 highest temperatures, the compound thus showing evidence of not having 

 been dissociated at these high temperatures ; and SO2 had vapour-density 

 for this formula at a white heat. CO was '^ partially decomposed at 1690° 

 thus, 2CO=C + COo; on this account the volume is less than it should 

 be, and also on account of a slight difiTusion of the CO through the plati- 

 num at the high temperatui-e ; hence there is found, in place of an ab- 

 normal expansion of CO, a slightly increased density fas compared with 

 air at the same temperature) ; there is almost normal expansion of CQ 

 up to 1200°, but at much higher temperatures decomposition begins to 

 set in. 



N2O is almost entirely split up at 900° into nitrogen and oxygen ; and 

 at 1690° it is split up to just about the same extent. 



NO is entirely decomposed into nitrogen and oxygen at 1090°, but is 

 unaltered at 1200°. 



HCl was considerably decomposed at 1300° and higher temperatures, 

 the hydrogen difiFusing through the platinum, and the chlorine being 

 shown by the amount of iodine liberated from potassium iodide solution. 



CO,, heated in polished platinum, was only very slightly decomposed 

 at the highest temperature, about 1690° ; in presence of fragments of 

 porcelain Deville had found much dissociation at 1300°. 



The Curmdative Evidence for Avogaclro's Lcm — Application of the Law to 

 the Behaviour of the Salogens at High Temperature. 



In a paper on some points of the atomic theory, published in 1826, 

 Dumas'* gave determinations of the vapour densities of iodine and mercurj', 



' C.R. Ixxv. p. 76. = CM. Ixxxix. 702. 



^ Langer and Meyer's Pyrocliemiselw Untersuchungen. 



* Annales de Chimie et de Physique (2), xxxiii. p. 337, 1826. 



