236 Royal Society: — Messrs. Lockycr and Roberts 



battery, was placed opposite one end of the tube, and a spectro- 

 scope (G-) opposite the other. This last instrument was by Desaga, 

 of Heidelberg, and its single prism, the angle of which was 6u°, 

 was capable of distinctly separating the D lines, at the same time 

 that it enabled us to see the whole spectrum in a single field of 

 view, an essential point in such inquiries. The magnifying-power 

 of the telescope was 7*5 linear. 



Some preliminary experiments indicated the advisability of in- 

 creasing the length of the column of vapour. To effect this, a tube 

 30 centims.- long was made in a fresh block of lime, the cavity 

 being arranged as before ; in each end a short accurately fitting 

 iron tube, luted with a mixture of graphite and fireclay, was in- 

 serted ; and the total length of the column thus became GO centims. 



The lime block (C) with its fittings was then placed in the 

 charcoal -furnace (E), by means of which the whole could be raised 

 to a high temperature. As soon as the block was heated to bright 

 redness, the metal, the vapour of which was to be examined, was 

 introduced into the cavity (C), and the flame of the oxyhydrogen 

 blowpipe (D) was allowed to play on its upper surface, care being 

 taken to employ an excess of hydrogen. In almost every case the 

 metal experimented on was rapidly volatilized (the exceptions being 

 gold and palladium). The central portion of the lime block was 

 raised to a white heat by the action of the blowpipe. As the glass 

 plates rapidly became clouded by the condensation of the metallic 

 vapours, it was necessary to adopt an arrangement by which they 

 could be easily replaced. We may state that, among the pre- 

 cautions which we adopted in order to assure ourselves that oxides 

 were not present to disturb the accuracy of the results, one of the 

 glass plates was removed at the conclusion of each experiment, 

 and the presence of an excess of hydrogen conclusively proved by 

 igniting it at the open end. 



We were enabled at any time, by modifying the conditions of the 

 gas-supply, to introduce the spectrum of the oxyhydrogen flame. 

 It may further be stated that, with few exceptions, the metals were 

 previously melted in a stream of hydrogen and enclosed, until ex- 

 perimented on, in sealed glass tubes. We ascertained that the 

 effect of oxides, and of the metallic rain due to condensation, was 

 to produce a general absorption obviously different from the special 

 effects of absorption which we now proceed to record. 



Details of the Experiments. 



Silver. — Eifty grammes of pure metal were placed in the cavity 

 (C) ; and this amount produced a continuous supply of vapour for 

 about 10 minutes. 



With the smaller thickness given by the first lime block, and 

 with a less powerful blast, the spectrum of silver consisted of an 

 absorption in the blue which at times extended almost to the green. 



With the elongated tube and a stronger blast an exquisite chan- 

 nelled-space absorption was observed, the channels being far enough 



