August 24, 1888.] 



SCIENCE. 



91 



the determination of atomic weights is of the greatest importance 

 for the validity of the modern tlieories of chemistry, many experi- 

 menters of the greatest sl<ill devoted themselves to researches on 

 this subject, and, by means of improved methods, results of great 

 accuracy were obtained. All these researches are of great intrin- 

 sic value and interest ; but nevertheless they show, that even yet, 

 with all the advantages of purity of material, perfection of appara- 

 tus, and precision of methods, united to great skill and extensive 

 attainments on the part of the experimenter, the attractive hypoth- 

 esis of Prout yet remains experimentally unproved. Many hold 

 that the failure in the proof has been due to constant errors in the 

 experimental processes; but Meyer and Seubert, from an elaborate 

 discussion of the determinations of the atomic weight of silver and 

 of those of the other more important elements calculated by its 

 means, declare that they all contradict Prout's hypothesis in its 

 characteristic original conception, and that it must therefore be 

 looked upon as having been disproved by experiments. 



Crookes suggests a hypothesis which may account for certain of 

 the discrepancies in the atomic-weight determinations without re- 

 sorting to the supposition of constant errors. He supposes that 

 elements, instead of being composed of parts of matter which are 

 identical throughout, are really composed of groups of particles 

 which are only approximately alike, and whose weights only ap- 

 proximate to that average which we call the atomic weight. 

 Hence it is possible that in different portions of such congeries dif- 

 ferent average values within small limits may obtain. Still it is 

 remarkable that such close coincidences should result as have re- 

 sulted from the observations made on material obtained from widely 

 separated sources. 



The determination of molecular weights is of nearly equal im- 

 portance with that of the weights of the atoms. Thanks to Avo- 

 gadro's law, we are able, when the substance can be obtained in 

 the gaseous state, to determine its true molecular formula. When, 

 however, the body cannot be completely volatilized unchanged, we 

 have until recently been dependent upon isomorphism and the laws 

 of molecular volumes and of specific heats, and upon analogical com- 

 parisons, to furnish us with estimates of the molecular weights. A 

 new method of determining these weights was discovered by 

 Raoult, who deduced a formula from the depression of the freezing- 

 point of solutions. He showed, that by knowing the weights of the 

 substance dissolved and of the solvent, and by knowing the depres- 

 sion of the freezing-point, the molecular weight may be calculated. 

 He has examined a large number of substances whose molecular 

 weights had previously been determined by their vapor densities, 

 and the results obtained illustrate in a reriiarkable manner the ac- 

 curacy and general application of this new method. 



There has long existed a conviction in the minds of chemists that 

 the molecular constitution of bodies in the solid state was much 

 more complex than in the gaseous, owing to polymerization ; and 

 the opinion finds support in the diminishing density and increasing 

 molecular simplicity of such bodies as acetic acid and sulphur when 

 subjected to high temperatures. By analogy this aggregation of 

 molecules should proceed as we pass from the gaseous through the 

 liquid to the solid state. Is it not, then, singular that the molecular 

 weights derived from Raoult's method for bodies in a state of solu- 

 tion should be identical, or approximately so, with those deduced 

 from their densities in the state of a gas ? This method fails to 

 afford any indication whatever of this molecular complexity in sol- 

 ids and liquids. Must it not, then, be assumed that the solvent has 

 effected the complete dissociation of the complex molecules present 

 in it.' If so, this probably extends to all cases of true solution 

 without chemical action, if such there be ; and this is assumed in 

 this method, for, although the solvent used has been varied, it has 

 g^ven similar results. 



Until recently we have known little precisely about the nature of 

 solution. It has been held by some to dilTer essentially from chem- 

 ical combination, but no satisfactory solution was offered until 

 Mendelejeff made his important researches on this subject. He 

 says, solutions may be regarded as strictly definite, atomic, chemical 

 combinations at temperatures higher than their dissociation tem- 

 perature. Definite chemical substances may be either formed or 

 decomposed at temperatures which are higher than those at which 

 dissociation commences. The same phenomenon occurs in solu- 



tions : at ordinary temperatures they can be either formed or de- 

 composed. In addition, the equilibrium between the quantity of 

 the definite compound and of its products of dissociation is defined 

 by the laws of chemical equilibrium, which laws require a relation 

 between equal volumes and their dependence on the mass of the 

 active component parts. Therefore, if the above hypothesis of so- 

 lution be correct, comparisons must be made- of equal volumes. 

 The specific gravities are the weights of equal volumes; and, more- 

 over, we must expect the specific gravities of solutions to depend 

 on the extent to which the active substances are produced : there- 

 fore the expression for specific gravity, .t, as a function of the per- 

 centage composition,/, must be a parabola of the second order, 

 while between two definite compounds which exist in solutions we 

 must expect that the differential co-efficient ^ will be a rectilinear 

 function of/. 



This theory has been proved by experiment, and not a single case 

 was found in which it did not hold good. Later on, Crompton and 

 Mendelejeff extended this theory to the discussion of electric con- 

 ductivity of aqueous solutions, and the results have been very en- 

 couraging, being entirely in favor of Mendelejeff's theory of solu- 

 tion. Thus it is shown that even this seemingly simple process is 

 very complex, and it is in the study of such processes that probably 

 the most important progress in the theory of chemical processes will 

 be made. This study will lead us to a clearer understanding of 

 the properties of matter. 



The evidence supplied by the various branches of chemistry 

 has forced the conviction in the minds of many of the ablest 

 chemists, that all matter is one, and varies only as it is acted upon 

 by force ; while, on the other hand, the transformations of energy 

 which are continually to be seen occurring in nature and in art, as 

 continually prove the truth of that glorious conception, the doctrine 

 of the conservation of energy, and equally force the conviction that 

 all energy is one, and varies only in its manifestations. 



The belief in the unity of matter is as old as philosophy, and, as 

 has been said, this belief has in recent times been strengthened to 

 conviction by the development of such facts as I have alluded to- 

 above ; and this conviction has been supported by the more recently 

 discovered evidence that the properties of the elements are func- 

 tions of their atomic weights, and that the elements, when arranged 

 according to their atomic weights, fall into natural and periodic 

 groups ; for it is a fundamental deduction from the law of perio- 

 dicity, that the various elementary atoms must be aggregations or 

 condensations of one and the same primordial substance. Strong 

 as the conviction resting upon this evidence may be, there is yet 

 lacking the crucial proof ; for we have as yet failed to observe the 

 passage of matter from the form of one elementary substance to 

 that of another, or the resolution of any element into or its creation 

 from prirnordial matter. 



The case for the evolution of the elements from periodical matter 

 has been very ably summed up by Crookes, while, in addition, he 

 has brought forward experimental proof of the possible existence of 

 bodies, which, though neither compounds nor mixtures, are not ele- 

 ments in the strictest sense of the word. These bodies, which he 

 styles ' meta-elements,' consist of different groups, which shade off 

 so imperceptibly the one into the other, that it is impossible to erect 

 a definite boundary between any two adjacent bodies, and to say- 

 that the body on this side of the line is an element, while the one 

 on the other side is non-elementary. Yet by means of fractiona- 

 tion these bodies may be separated one from the other, and then 

 they exhibit slight spectral differences. 



Finally Griinwald has announced that during a mathematical in- 

 vestigation of the changes which the properties, and especially the 

 spectra, of two bodies undergo when they unite to form a new sub- 

 stance, he discovered a simple and important proposition of a fu- 

 ture chemico-ma'henr.atical theory of perturbations; and by its 

 means he has shown the compound nature of hydrogen and 

 oxygen, and has demonstrated the dissociation of hydrogen in the 

 sun. The method employed is a spectral one, and requires condi- 

 tions which cannot be reproduced at the will of man ; so that if it 

 stands the tests of criticism, which is doubtful, it will not then en- 

 able us to witness the evolutionary process in actual operation. 



Hence we find for the doctrine of evolution in the domain of 

 chemistry, that the tests yield absolute results when applied to- 



