TRANSACTIONS OF SECTION B. 481 



that the change of vis vioa of a system, in passing from an initial to a final configura- 

 tion, is independent of the intermediate stages through which it may have passed 

 Srovided the external conditions are unaltered ; and on the principle of the 

 issipation of energy, that is to say, on the condition that the state of the system, 

 if it be a stable one, must be such that the energy run down in reaching it is a 

 maximum. These principles have been applied successfully to the solution of some 

 particular cases of the equilibrium between a mixture of chemicals by Willard 

 Gibbs, Berthelot, and others. By the first-mentioned principle all consideration of 

 the intermediate stages by which the final residt is reached is avoided. Quite 

 recently Lemoinne has attacked the same problem on another principle. His 

 principle is that of an equilibrium of antagonistic reactions in a mixture of 

 materials, a mobile equilibrium, such as we are now familiar with, dependent on 

 compensating effects •, but he does not seem able to solve the problem in any great 

 number of cases. In fact, the difficulty does not now lie so much in expressing 

 mathematically the conditions of the problem, as in the defect of knowledge which 

 depends upon experiment. And it is just in this that I think the outlook most 

 hopeful. In some cases the patient work of weighing, measuring, and comparing, 

 which is necessary to make our theoretic speculations of any substantial value, has 

 been already done for us. The publication, three years since, of Berthelot 's ' Essay 

 on Chemical Mechanics,' has given us, in a collected form, a large quantity of data 

 of the first importance, and now I am glad to say that the long labours of another 

 worker in the same field, Thomsen of Copenhagen, are in course of publication in a 

 handy form. I think these two investigators have done more than anyone else of 

 late years towards making it possible to give chemistry the rank of an exact science. 

 But besides the data which they have supplied to us, there are others which are 

 yet wanting. For instance, almost every equation of chemical equilibrium involves 

 an expression depending on the specific heat of the materials. At present we do 

 not know enough of the law of specific heats to be able to give in most cases a 

 probable value to these expressions, but these and other data of the kind do not 

 seem out of our reach, and we may hope that the same ingenuity and patience 

 which have gained for us so much firm ground in thermal chemistry will extend it to 

 the imcertain spots where we have as yet no solid foundation. 



Further, the laws of dissociation, so ably investigated by Deville, have taught 

 us that the force called chemical affinity, by which we suppose the atoms of unlike 

 matters are held together in a compound molecule, follows precisely the same laws 

 as the force of cohesion by which particles of a similar kind are iniited in molecides. 

 We have long known that the molecules of sidphur vapour are broken up into simpler 

 molecules by elevation of temperature, and condense again when the temperature is 

 reduced. Other elementary substances behave in a similar way. We have within the 

 last two or three years learnt that iodine is in part dissociated by a high temperature 

 from molecules consisting of two chemical atoms into molecides consisting of 

 only one such atom, and the same is true of chlorine and bromine. That some such 

 change must occur in iodine was inferred as long ago as 1804 by the younger 

 Mitscherlich. He argued that iodine is a compound body from the fact that it 

 shows two spectra, one similar in character to tliose of metallic oxides and the 

 other similar to the spectra of metals, and from the analogy in the behaviour of 

 iodine to a metallic oxide in giving the one spectrum at one temperature and the other 

 at a higher temperature : ' from this,' he says, ' it would follow that iodine at ordinary 

 temperatures and iodine at the temperature of a hydrogen flame must be considered 

 as two different compounds, because the spectrimi of iodine formed at ordinary 

 temperatures ' (that is, the absorption spectrum of iodine vapour) ' is different from 

 that produced in a hydrogen flame.' Also ' that bromine, though it gives no flame 

 spectrum, gives one spectrum by absorption and another by the electric spark, and 

 must therefore in its ordinary state be regarded as a compomid.' Also that ' the 

 spectra formed by the flames of seleniiun, tellurium, and phosphorus, and those of 

 sulphur and nitrogen given by feeble electric discharges, all have the character of 

 the iodine flame spectrum, and these metalloids would therefore, if the above ex- 

 pressed supposition with regard to iodine be confirmed, also be compound bodies.' 

 <' Phil. Mag.' 1864, p. 188.) Since the paper from which the foregoing extract is 

 1882. I I 



