THE ABSORPTION AND EMISSION OK AITi. 27 



use. It must be tested by special photographs made for this purpose. For such 

 test a vacuura-spectrograph of suitable dimensions is I'equired. Theie is no shorter 

 way to the ascertainment of the serviceableness of a diflt'raction e-ratiiiir for use in 

 the ultra-violet. 



The body of the spectrograph should be so constructed as to afford the bi'oad- 

 est guarantee for its air-tightness. Moreovei-, thei'e should be no flexure upon 

 e.xhaustiou. Should there be flexure, however, it is still possible to obtain useful 

 photographs under unaltered pressui'e, since to this would correspond a constant 

 bending. But shouhl leakage or any other cause alter the pi'essure fallacious re- 

 sults would be obtained, in consequence of the shifting of lines due to the bending 

 of the apparatus. The exterior best adapted to such a spectrograph would be that 

 of a tube having its walls not too thin and composed of di'awn or rolled material. 

 For a tube presents equal resistance to outward pressure, and therefore, supposing 

 its walls to be equally strong, and to be subjected to equal pi-essui'e in all direc- 

 tions, will show the smallest bending. A tube drawn of brass or a Mannesmann 

 tube of steel is preferable to a cast tube on account of its greater homogeneity. 

 The open ends must be closed with permanent covers. At one end the concave 

 gi-ating should be set up so as to be capable of rotation about its mid-ruling as an 

 axis, while at the other end should be the slit-cai'riage, movable in the direction of 

 the rays. To one side of the slit there should be a horizontal slot in the cover of 

 the tube, giving exit to the rays coming back fi'om the grating. Before this slot 

 there should move vertically an air-tight slide containing the plate-holder, so that a 

 series of photogi'aphs could be taken one under the other without altering the 

 pressure within. The focussing of the different photographic fields would have to 

 be effected by displacement of the slit-carriage, which to this end must be provided 

 with a sufficiently long air-tight di'aw-tube. 



The most important part of the whole apparatus is the slide carrying the 

 plate-holder. Upon the sufficiency and permanence of its air-tightness depends the 

 minimum pressure obtainable by exhaustion of the apparatus and thus the success 

 and the cost in time of the photographs. But the displacement of the plate and 

 the prevention of the entrance of air when it is taken out are not the only purposes 

 subserved by the plate-holder slide. It has also to make it possible to exhaust the 

 air that comes in with the plate before the latter is exposed to the light. For it is 

 clear that this small quantum of air can be removed much more quickly from the 

 narrow space of the interioi' of the slide than from an apparatus of many litei's' 

 volume. The plate should not be brought before the air-slot until in this way the 

 slide and the plate-holder have been exhausted. Thus when the apparatus has 

 once been exhausted, supposing its other parts to hold, it will only have to be 

 thoroughly pumped out now and then. I must not, however, be understood to 

 assert that in this way the vacuum will hold in the spectrograph for weeks oi' even 

 for days, as a good mercurial pump, for example, will. Such a degree of tightness 

 is not often met with among vacuum-spectrographs. At best, it will be necessary 

 to pump a little after every change of plate in order to remove gas that has been 

 set free ; as, for example, absorbed air. Even when the closing surfaces have been 



