96 



SCIENCE. 



[N. S. Vol. XIX. No. 472. 



elements has never been obtained from the coin- 

 iSidfence of a line dr lines exclusively belongi'fig to 

 the spectrum of one element with a line or lines 

 in the spectrum exclusively belonging to another 

 element. This view is based upon the following 

 grounds: (1) Because the coincidences have gen- 

 erally been shown to be only apparent, and have 

 never been proved to be real ; ( 2 ) because the great 

 difficulty of obtaining one kind of matter entirely 

 free from every other kind of matter is so great 

 that where coincident lines occur in the spectra of 

 what have been believed to be elementary sub- 

 stances, they have been shown from time to time 

 to be caused by traces of foreign matter, such as 

 by chemists are commonly termed impurities; 

 (3) no instance has ever been recorded of any 

 homologous group of lines belonging to one ele- 

 ment occurring in the spectrum of another, ex- 

 cept and alone where the one has been shown 

 to constitute an impurity in the other; as, for 

 instance, where the triplet of zinc is found in 

 cadmium and the triplet of cadmium in zinc the 

 three strongest lines in the quintuple group of 

 magnesium is graphite, and so on. The latest 

 elucidation of the cause of coincidences of this 

 kind arises out of a tabulated record from the 

 wave-length measurements of about three thou- 

 sand lines in the spectra of sixteen elements made 

 by Adeney and myself. The instances where 

 lines appeared to coincide were extremely rare; 

 but there was one remarkable case of a group of 

 lines in the spectrum of copper which appeared 

 to be common to tellurium; also lines in indium, 

 tin, antimony and bismuth which seemed to have 

 an origin in common with those of tellurium." 



The last sentence presents the point I 

 wish to emphasize. Tellurium has long ob- 

 truded itself before a satisfactory vision 

 of the natural system. The table alone 

 recites not a few efforts to obtain the con- 

 taminating constituent of tellurium which 

 a priori is present from Hartley's observa- 

 tions (see also Griinwald 1889 table). The 

 fractionation of a rubidium-csesium mix- 

 ture, perhaps, is a simp^ler problem than 

 that confronting Pellini,* Avho reports a 

 definite amount of an element with a high 

 atomic weight (about 214), similar to and 

 associated with tellurium. 



* Gae. Ohim. ital, 33, 11, 35. 



What has been said applies especially 

 to the elements of the rare earth class 

 —'asteroids of the terrestrial family'— as 

 phrased by Crookes. Many of them have 

 not been secured with sufficient purity to 

 claim an inherent spectrum; further, the 

 spectra attributed have not been obtained 

 under uniform conditions. 



I have referred* somewhat in detail 

 elsewhere to the factors producing varia- 

 tions in the absorption, as well as the ad- 

 vantages and disadvantages of the phos- 

 phorescent and reversal, spectra. 



Without doubt the spectroscopic criteria 

 are the most valuable Ave have in judging 

 finally the elements, and mayhap will re- 

 main so, but in my humble opinion, such 

 have not alone sufficient authority, as yet, 

 to usher the aspirant to a place among the 

 elect. The contention frames itself, how- 

 ever, in an expression of the need for uni- 

 formity. 



Whether we follow the most advanced 

 metaphysico-chemical teachings or no, if 

 there be any one concept upon which mod- 

 ern practical chemical thought depends, it 

 is the law of definiteness of composition. 

 There may be, and doubtless are, definite, 

 perhaps invariable, pi'operties of our ele- 

 ments other than their combining propor- 

 tions, the atomic weights, if you please, 

 yet, as far as we know, they approximate 

 more closely than any fixed, if not perma- 

 nent, ratios. Many of these values, by 

 which Ave lay such store, are dependent 

 upon dataf in which, I venture the asser- 

 tion, too great confidence has been be- 

 stowed, or opinions to which sufficient at- 

 tention has not been given. 



Although in this connection we shall 

 give little heed to the suggested variability 

 of the relative values, it may be remarked 

 that Boutlerow, noting the variations ob- 



* ' The Rare Earth Crusade,' loc. cit. 

 t Others haA'e been referred to in the address to 

 which this is a sequel. Loc. cit. 



