74 



KNOWLEDGE 



[April 1, 1892. 



per ceut. of the Siriau stars had proper motions, amounting 

 to one-tenth of a second annually (for the Orion stars, 

 Spectrum B of the Draper Catalogue, the proportion falls 

 to 11 per cent.), 86 per cent, of the Arcturian and 65 per 

 cent, of the Capellan. The difference seemed to become 

 more marked as the stars became fainter. It is curious, 

 however, that for stars with very large proper motion 

 — say over one second and a half per annum — the 

 Arcturian type largely predominates. The Arcturians 

 are perhaps the real " runaway stars." The steady-going 

 members of the Solar cluster show no tendency to bolt 

 the course. Such, at least, is the appearance which they 

 present on a first inspection. 



I remahi, Sir, truly yours, 



W. H. S. MoNCK. 



THE CHEMICAL RESEARCHES OF JEAN 

 SERVAIS STAS. 



By Vaugh.\n Cornish, B.Sc, F.C.S. 



IN the last month of last year the chemical world 

 received with profound regret the news of M. Stas's 

 death, at the advanced age of seventy-eight. 

 The name of Stas has been a household word 

 among chemists for half a century, and his writings, 

 the celebrated Reclurches stir lea Lois ih's Proportions 

 Chimiques, have come to be regarded as among the 

 canonical books of chemistry. In all that related to the 

 experimental art Stas stood unsurpassed. The marvellous 

 patience with which he matured his methods, and the 

 skilful care with which the final experiments were carried 

 out, stand recorded in his classical memoirs with that 

 clearness and precision of expression characteristic of 

 French scientific writings. Stas's work bore on one sub- 

 ject only, the determination of " atomic weights," with a 

 view more particularly to ascertain if there existed any 

 simple definite relation between the weights of the 

 chemical atoms. In order to explain how this investiga- 

 tion came to be the mission of Stas's life, we must refer 

 to the state of chemical theory in the second decade of the 

 present century. At this time the laws of chemical com- 

 bination had been formulated and accepted — the laws, viz., 

 which may be epitomized by saying that " chemical 

 elements combine together only in the proportion of their 

 equivalent weights, or in simple multiples of those propor- 

 tions." Dalton had propounded an explanation of these 

 laws in his " Atomic Theory," according to which chemical 

 combination was due to the union of chemically indivisible 

 particles, the particle or atom of each element having its 

 own particular fixed weight. 



Dalton's theory, the next great generalization after 

 Lavoisier's explanation of the phenomena of combustion, 

 was the result of the discovery of definite and simple 

 numerical relations between certain chemical quantities. 

 It was natural that other minds, impressed by Dalton's 

 theory, should seek for other such numerical relations in 

 the hope of fresh discoveries of Nature's laws. In 1815 a 

 paper appeared in Thomson's Annals of Philosophij by 

 Dr. Prout, in which he pointed out certain apparent 

 relations between the atomic weights of the elements as 

 then determined. The idea was at once taken up by other 

 chemists, and took shape in the following form, known as 

 Front's Hijiiothesis : — "The weight of the atom of each 

 element is a simple multiple of the weight of the atom of 

 hydrogen." The observed deviations were referred to 

 errors of experiment, just as the apparent deviations from 

 the laws of chemical combination were referred to 

 experimental error. 



It has been the life-work of Stas to investigate both 

 assumptions, and to show that while the laws of chemical 

 combination are rigidly exact the supposition of Prout is 

 unsupported by experimental evidence. 



Prout's hypothesis owes its importance in the history of 

 science to the fact that it seemed to restore the old theory 

 of the unity of matter, which appeared to have received its 

 death-blow with the discovery of the chemical elements. 

 But if the atom of each element be exactly once, twice or 

 thrice the weight of the atom of hydrogen, then it is 

 reasonable to suppose that the atoms of all elements 

 contain only one kind of matter, and that the hydrogen 

 atoms are the one class of ultimate particles of 

 which all matter is built up. As the art of 

 chemical analysis developed under the hands of the 

 great Swedish chemist, Berzehus, it became evident 

 that Prout's hypothesis was not tenable in its original 

 form. It was revived, however, in a modified shape 

 chiefly owing to the influence of Dumas. In the modified 

 form, the hypothetical unit weight was that of the half- 

 atom of hydrogen. Later on, Dumas was compelled to 

 retreat yet further from the original position, and to take 

 the quarter-atom of hydrogen as the greatest common 

 divisor of the atomic weights. In this modified form the 

 idea of Prout loses much of its interest, since the "quarter 

 atom " of hydrogen is itself an unknown thing. Neverthe- 

 less, the idea of the oneness of matter always exerts a 

 certain fascination, and to some minds this unity of matter 

 appears to be almost a logical necessity. Hence the 

 tenacity with which chemists have clung to the belief that 

 apparent discrepancies were due to errors of experiment, 

 rather than to the inaccuracy of Prout's hypothesis. 



Stas began his researches on atomic weights with a 

 strong prepossession in favour of the hypothesis. He 

 chose for his determinations such substances as could be 

 prepared in a high state of chemical purity, and worked 

 with large quantities of substance in order to eliminate the 

 effect of errors in weighing. A large number of experi- 

 ments, which occupied several years, furnished him with 

 extremely accurate values for the relative weights of the 

 atoms of silver and the alkali metals, and of chlorine, 

 bromine, and iodine. Moreover, the variety of methods 

 employed served to eliminate possible systematic errors — 

 errors, that is to say, not due to want of skill in the 

 performing of an experiment, but due to the method itself. 

 Each substance, moreover, was prepared in several difl'erent 

 ways and from different natural sources. Not the least 

 remarkable tribute to Stas's skill is the close accordance 

 between the values he obtained for the atomic weights by 

 different processes of determination. The numbers 

 obtained in this first series of researches were closely 

 accordant among themselves, and wholly at variance with 

 those demanded by Prout's hypothesis. Stas concludes 

 his memoir thus : " Prout's hypothesis must be looked 

 upon as a pure delusion ; the elements must be considered 

 to be distinct entities, with no relation between their 

 atomic weights." 



The accuracy of Stas's work was admitted on all sides, 

 but his conclusions were contested. The criticisms of the 

 Genevese chemist, Marignac, are historically important, 

 having led Stas to his second and more celebrated 

 research. Marignac contended that it was far from being 

 proved that the constituent elements of many chemical 

 compounds were present exactly in the proportion of then- 

 atomic weights. It was possible that many chemical 

 compounds contained normally a very small excess of one 

 or other of their constituents. This criticism strikes at 

 the basis of the Atomic Theory, since that theory is 

 founded on the assumption that the laws of chemical 



