July 6, 1§00.] 



SCIENCE. 



23 



by the platinum. This was the origin of 

 Dobereiner's hydrogen lamp, regarding 

 which he writes under date of August 5, 

 1823 : " You have doubtless guessed before 

 this that I have already utilized this new 

 observation (of the heating effect of the 

 condensation of hydrogen) for the prepara- 

 tion of a new Feuerzeug and a new lamp, 

 and that I shall put it to many other im- 

 portant uses."* The interest which at- 

 tached to this discovery of Dobereiner's 

 can be judged from the fact that in the 

 literature of the decade following are found 

 over fifty references to the subject. Little 

 attention was, however, paid to the similar 

 action of palladium upon combustible gases, 

 although the phenomenon had been noticed, 

 until in 1868, half a century subsequent to 

 Davy's first observation on platinum, Gra- 

 ham presented to the Royal Society his 

 remarkable paper on the occlusion of hy- 

 drogen by metals,t followed the next year 

 by his papers on the relation of hydrogen 

 to palladium, J and additional observations • 

 on hydrogenium.§ 



Graham's view that the hydrogen is pres- 

 ent in solid form as a metal and that the 

 palladium saturated with hydrogen must 

 be looked upon as an alloy, was receive'd 

 with considerable dissent. The work of 

 Troost and Hautefeuille 1 1 tended to the 

 view that the substance is a definite com- 

 pound, Pd,H. Against this is the fact that 

 the conductivity of palladium is but slightly 

 reduced by the occlusion of hydrogen. Cal- 

 culations of the specific gravity of Graham's 

 hydrogenium by Dewar gave the number 

 0.62 and the same figure is obtained for the 

 hydrogen in hydrids of sodium and potas- 

 sium, studied by Troost and Hautefeuille. 

 The recent determinations of the specific 



* Loc. cit. 



■\Proc. Boy. Soc, 16, 422 (1868). 



%lbid., 17, 212 (1869). 



llUd., 17, 300 (1869). 



II Compt. rend., 78, 686, 968 (1874) 80, 788 (1875). 



gravity of liquid hydrogen by Dewar, how- 

 ever, show a figure only about one- ninth of 

 the densitj"- of occluded hydrogen, so that the 

 question as to the nature of the hydrogen 

 condensed by the palladium and platinum 

 remains still unsettled. The other metals 

 of the group possess this property to some 

 considerable degree, but much less than is 

 the case with palladium and platinum. In 

 this connection it is interesting to note that 

 one of the earliest papers of Dewar was on 

 the motion of a palladium plate, during the 

 formation of Graham's hydrogenium.* 



Reference has been made to the natural 

 grouping of the elements of the eighth 

 group into three triplets, iron, ruthenium, 

 osmium ; cobalt, rhodium, iridium ; and 

 nickel, palladium, platinum. That this is 

 a natural grouping is attested by a com- 

 parison of the compounds of these metals. 

 However, in considering now some of these 

 compounds the evidence of this grouping is 

 only incidentally presented ; I desire chiefly 

 to call attention to some of the more un- 

 usual of these compounds, especially with 

 reference to problems which this group 

 presents, and to problems of other groups, 

 suggested by the chemistry of this group. 



The position of an element in the periodic 

 system is, to a very considerable extent 

 determined by its oxids, and that too by its 

 highest oxids, excluding the peroxids of the 

 hydrogen peroxid type; a considerable num- 

 ber of these last have been studied especially 

 by MelikofiF and Pissarjewsky of Odessa, 

 but their character still presents many points 

 of obscurity and cannot be used with refer- 

 ence to the periodic law. The triplet iron, 

 ruthenium, osmium presents the highest 

 oxids of the eighth group, and, as is the 

 case with other divisions of this group, an 

 increasing stability of the higher oxids with 

 increasing molecular weight. The type of 

 salts of the acid-forming oxids Fe03, ^uO,, 

 OsO.„ occurs in this group, as in the posi- 



*Proc. Roy. Soc. Edirib. 6, 504 (1869). 



