20 THE ABS01£1>TI0N AND EMISSION OK All:. 



tbe spectrograph could not seriously be feai'ed, considering the many years during 

 wliicli it lias been in use, I do not think that any of these sources of erroi- can be 

 suppDsetl to have affected the coiTectness and i)urity of the photographs obtained. 



The rays in theii- passage to the photographic plate traversed a hydiogen 

 atmosphere of moderate density, and not the exliausted interior of the spectrograjth 

 only. For the first five photographs, the gas, most sciu[)ulously purified and dried, 

 was led directly into the discharge-tube, and thence into the air-pumj). In this 

 way the tube was ooiitiiiuously traversed and eventually filled with gas which 

 must have been either altogether or very nearly fi-ee from mercurial vapor. For 

 the other seven [)hotographs the hydrogen, obtained by the action of the purest 

 zinc upon the purest suli>liuric acid, and aftervvai'd dried, was led through two 

 meters of the aii-pumj) tubing befoie entering the si)ectrograi)h and thence into 

 the Geissler tube. Consequently, it could not, in this case, be so fi'ee fi'om mer- 

 cury as in the other. I have not, however, been able to perceive any particular 

 disadvantage from the presence of the greater amount of mercury. 



The induction-current was given by a Ruhmkorff coil, fed by 5 Grove cells 

 and giving sparks 25 cm. long. The effective rays from the tube' traversed suc- 

 cessively the fluor-spar window, the collimator-slit, the first leus, a 60° prism, and 

 the second lens, over against which was the plate sensitive to ultra-violet light. 

 The lenses and prism are composed of white tiuorite. The lenses weie i)lano-convex 

 and had the same focal length, namely, 120 mm. for sodium yellow. The equality 

 of the focal lengths made proper positions of the lenses to be at equal distances, 

 the one from tbe slit and the other from the mid-plate. Since the refiangibility 

 of the short-wave rays increases very fast in fluor-spai- with decreasing wave-length, 

 the focus rapidly shortens Avith the wave-length. The consequence is that the 

 scale of the image of the line diminishes from the less refrangible to the more 

 refi'angible end of a photograph. In pliotogi'aphs patched together to show a long 

 stretch of the spectrum this difference of scale makes itself unpleasantly noticeable, 

 owing to the varying breadth. But on such short photographs as these of the 

 hydrogen spectnim it may ])e neglected. 



When thei'e was no particular i-eason foi- placing the prism symmetrically to 

 the course of certain rays, the rule was to give their minimum deviation to the 

 most refrangible lines of the part of the spectrum to be photographed at one 

 e.\[)osure; for this is the oidy way to get a uniformly shaip definition throughout 

 the stretch photographed. If I'ays are to traverse the prism symmetrically, it will 

 only be necessary, while preserving the equality of the foci of the two lenses and the 

 minimum of deviation, to turn the camera-tube until those rays fall on the middle 

 of the plate. For, as Fig. 9 shows, the camei-a is so consti-ucted that, whatever 

 may be the angle between the jilane of the photographic plate and the axis of the 

 lens, this a.\is must always cut the mid-vertical of the sensitive side of the plate, 

 which coincides with the geometrical axis of the camera-cone. The considerable 

 obliipiity of the plate to the axis of the lens has, along with the advantage of about 



' 'I'o jirotect the window from possible heating, the electrode farther away from the slit was 

 made the cathode. 



