12 TIIK AVtSORlTION AND EMISSION OK AIH. 



aiiiuilletl tbe effect of the lays at a wave-length of about 1G2 /'/' tlirougli an 

 extended stretch. Tliis behavior of india-rubber led me, in tbe sequel, to con- 

 nect the aj^iaratus only with glass tubes and a metallic-screw closing device, 

 which depends upon having optically plane surfaces which are pressed together 

 with gieat ivvi'f. I have used this clamp for different [)urposes and with constant 

 success. 



In the preparation of the emission tubes (Geissler tubes), several difficulties 

 arose. The first of these lay in attaching a suitable window ; for glass is entirely 

 opaque to such shoi-t waves. Quartz is eminently proper for the pui'pose, and 

 finor-spar is still more so, since it allows the most refi'angihle rays to [)ass freely, 

 which quartz, as is well known, does not. Quartz has the advantage of greater 

 hardness, which is no small consideration owing to the frequency with which the 

 windows have to be taken off atul cemented on again. Fluor-spar is relatively 

 soft and in constant use, especially when its surfaces are often cleansed, is haid to 

 preserve from damage. Moreover, there is a circumstance which has nari'owly 

 limited its ap[)licability for the windows of Geissler tubes in my experiments. The 

 state of things is this. With Geissler tubes we are led to employ the end-on ftirm 

 for the most I'efrangible ultra-violet, because it s'ives a erreater emission of lii^ht 

 tiiau the cross-wise tubes, and, besides, it is better adapted to the in.sertion of a 

 plane window. I follow the usual practice of giving such a tube a plane rim at 

 one end upon which the window is readily fastened. But this closure, to be ab- 

 solutely staunch, is a more difficult matter with the rays here concerned than one 

 would think. The closure iiuist fulfil three conditions, to wit: 1st, it nuist cut off' 

 the photographic energy as little as possible ; '2d, it must not soil the purity of the 

 gas in the tube ; 3d, it must easily be removed without damage to tube or window. 

 But none of the j^lans for closing tubes hitherto known satisfy these conditions, not 

 even the grinding out of quartz in the manner published by me in 188G, since, be- 

 cause of its great thickness, it absorbs too much refiaugible I'ays. For my experi- 

 ments, therefore, a more suitable method of closing had to be invented. Of several 

 processes to which my thorough trials led me, the following has satisfied me best : 

 The rim of the tube and the closing plate, which latter should only be 1 ram. to 

 •J mm. thick, after being cai'efully cleaned, are put together and held with a clamp 

 which must not be too large, so that there can be no lateral displacement. Then 

 the joint between the tube and i»late around the latter on the outside is to be 

 painted with a thick coat of soluble soda-glass. After two or thi'ee hours, wheu it 

 is dry, a second coat is to be applied. A third would make the tube stauncher, but 

 it is not absolutely necessary. Such a tube will bear the highest exhaustion that a 

 Geissler tube can take, while by scraping away the crust of water-glass with a 

 knife, which is easy, the tube is opened without trouble and readily closed again. 

 The convenient opening and closing is an important point in observations of the 

 ultraviolet, because tlie inner surface of the <piartz window often after a brief use 

 gets a deposit upon it, which partly absorbs the most refrangible rays, and, indeed, 

 sometimes entirely cuts them off, an inconvenience which is the more troublesome 

 because the deposit is geneially invisible to the eye. AVater-glass proved most 



