April i, 1880] 



NATURE 



5i7 



The explanation of this phenomenon put forward in 

 the paper alluded to was founded on the molecular 

 hypothesis; but I think that a fuller explanation is 

 afforded by the results of Prof Gibbs's investigations. 



When the water was added on the surface, a small 

 quantity of the matter in the vessel instantly passed into 

 another phase ; this being in contact with matter in the 

 original phase, induced therein a phase of relative insta- 

 bility, and this succession of phases proceeded until a 

 new condition of stable equilibrium was attained. The 

 entropy of the system was altered by the production of 

 small quantities of BiOCl ; K would almost certainly be 

 negative for the new phase with regard to the original 

 phase, and therefore the original phase would become 

 unstable by contact with it of a small portion of matter in 

 the new phase. If, however, the formation of matter in 

 the new phase were prevented by the special contrivance 

 of adding the water in a peculiar way, then the original 

 phase would be stable ; if, however, a somewhat large 

 quantity of water were added, the whole system might be 

 carried much beyond the limits of absolute stability with- 

 out overthrow of equilibrium, but this equilibrium would 

 necessarily eventually be overthrown, as was indeed 

 always found to be the case. 



In a paper recently published in the Journal of the 

 Chemical Society in conjunction with Mr. Slater, of St. 

 John's College, I detailed the results of an examination 

 of the influence exerted by variations in the amount of 

 water of dilution on the chemical change formulated — 



BaCl s + K,C,0 4 = BaC 2 4 + 2KCI. 



It is there shown that the progress of this change is 

 retarded to a proportionately greater extent by a large, 

 than by a small quantity of water of dilution, and that 

 this retardation is especially marked when the action pro- 

 ceeds at low temperatures. In order to explain this 

 result we suggested the hypothesis that when much water 

 is present and a low temperature is maintained various 

 hydrates of barium chloride, especially the cryohydrate 

 (liaCl 2 .37H 2 solidifying at — 8 D ), are produced, and that 

 these, being formed in presence of a large mass of one of 

 the products of their own dissociation, are comparatively 

 stable. We discuss and illustrate this hypothesis in the 

 paper, and, in our opinion, establish for it a fair degree 

 of probability. 



Now, if this hypothesis be granted, I think we have in 

 these experiments another illustration of the general 

 principle laid down by Prof. Gibbs. 



If BaCL-f K a C 2 4 be called phase A of the system, then 

 undoubtedly K is negative n ith regard to BaC 2 4 + 2KCI, 

 i.e., with regard to phase B. Phase A is absolutely un- 

 stable, and tends to pass into phase B. But during this 

 passage a phase, or phases, is reached which is only 

 relatively unstable. Could the matter in this phase (which 

 may be called the cryohydrate phase) be separated from 

 that portion already in phase 15, the intermediate phase 

 might become absolutely stable ; this, however, is not 

 done, and hence the whole system tends to pass into 

 phase B. But it is evident that those conditions which 

 favour the formation of matter in the cryohydrate phase 

 must also retard the passage of the system into phase B. 

 Moreover, while the system is in the cryohydrate phase, it 

 is carried to a certain extent beyond the limits of absolute 

 stability, and if this phase were abnormally extended we 

 should expect to obtain a condition of unstable equi- 

 librium liable to complete overthrow by small exciting 

 causes. In the paper referred to we show that such an 

 expectation can be realised. 



A class of reactions in chemistry, hitherto treated for 

 the most part as isolated facts, seems to find its 

 explanation in the generalisation now established by the 

 Yale Professor, viz., that the equilibrium of matter in a 

 relatively stable phase may be overthrown by contact 

 With even very small portions of matter in another phase. 



As examples of the reactions referred to may be cited/ 

 the decomposition of ozone by silver, and of barium 

 peroxide by platinum ; the production of diphenylmetlnne 

 from benzene and benzylic chloride only in presence of a 

 small quantity of zinc or copper ; the action of hydro- 

 chloric acid on bismuthous oxide in presence of a small 

 quantity of water ; ' and in general those numerous reac- 

 tions which are modified by the presence of traces of 

 foreign substances. 



In other cases contact with small quantities of matter 

 in another phase appears to retard the passage of the 

 main system from its initial phase to a phase of greater 

 stability. Thus Bunsen and Roscoe showed 2 that the 

 resistance to combination of a mixture of hydrogen and 

 chlorine when exposed to sunlight is increased by the 

 presence even of traces of oxygen." 



Every chemical system thus appears to tend towards a 

 phase of maximum stability. If the entropy of the 

 system be decreased, K will be increased in the general 

 equation of Gibbs, and hence the stability of the system 

 will be increased. 



All chemical systems not in phases of absolute stability 

 will therefore tend to lose entropy. 



This is probably a better method of stating Berthelot's 

 so-called law of maximum work than that generally 

 employed. 



Moreover, it may be possible to convert a phase of 

 absolute stability into a phase of relative stability, and 

 thence into a phase of absolute instability, by contact 

 with matter in another phase, i.e., in ordinary chemical 

 language, by the action of a reagent. 



The readiness with which so many chemical systems 

 undergo change leads one to ask whether chemical systems 

 are not generally in one or other of tho;e phases of 

 relative instability which are so easily overthrown by 

 contact with small quantities of matter in other phases. 

 If this be so, and if it be granted as extremely probable, 

 that even apparently stable systems are passing through 

 cycles of change, then we should expect that slight 

 changes in the values of the 'magnitudes" or " intensi- 

 ties " of chemical systems would in many cases induce the 

 overthrow of the stability of these systems. 



The considerable differences in the properties of many 

 carbon compounds, as described by different experi- 

 menters, may not improbably be due to the changes 

 induced in these bodies by contact with small traces of 

 impurities, i.e., of matter in phases other than that of the 

 main pcrtion of the system. 



The more complex the possible actions and reactions in 

 any given system of heterogeneous substances, the more 

 probable will be the occurrence of unstable phases, and 

 the more will the course of what we call the chemical 

 change be turned aside by small variations of the "magni- 

 tudes" or "intensities" of the system. 



In chemical changes involving few intermediate phases 

 — or to put it in another way, in chemical ch inges wherein 

 the action of " secondary forces '' is small— the course of 

 the change may be followed, and generalisations made 

 concerning it, as was done by Harcourt and Esson for a 

 special case.' When the action becomes a little more 

 complex, we appear to gain the conditions under which 

 so-called " chemical induction" becomes possible; while 

 from the study of exceedingly complex actions no genera- 

 lisations can be safely deduced. 



This view of chemical systems as readily undergoing 

 change when in contact with other systems, and of these 

 changes as being dependent on the energy of the systems 



x Examined by me in Chem. Soc./ouni., 1E79: PfW., p. 33«- 



3 Phil. Trans., . 



3 The iii'id fying influence exerted on a pr cess of chemical change by 



the presence of .a foreign substance was c nsidered in 1050 by Brodie. in an 



able paper, in which, speaking of the decomp s t. a f h.inun peroxide by 



platinum, he says: "•The platinum causes that chemical relation betwee 



the particles which renders the decomposition possible." 

 * Chem. Soc. Jourt:., xx. 460. 



