.»(; TIIK ABsoltlTlUN AXIJ li.MlSSluX UK AIK. 



r,.r the iiicasuiemeut ot \v;ivc'lfiii,'tlis, the wave-lengths can be determined by giat- 

 iiitr and lens, aceoiding to the method devised l)y Coinvi and applied by Liveiug 

 and Dewar. It is a method of coincidences resting upon the fact that the wave- 

 lengths of two lines whicli coincide in the axis of the lens are invei-sely propor- 

 tional to the numerical orders of the spectra. This method has the advantage of 

 not requiting any ray-filter, such as is required with a concave grating for the separa- 

 tion of mixed sjiectra. Foi- in my case, the lens not being achromatic, the spectra 

 do not at all coincide, being separated by the considerable difference of their foci. 

 This is an important advantage; for we artr acquainted with no suitable filter for 

 these short-wave rays, and a search for such a filter could hardly fail to be a very 

 laborious task indeed. Unfortunately, there is a serious drawback to the advan- 

 tages of this grating-and-lens method in the circumstance that the one of the two 

 spectra which is sharply focussed is so frequently overlaid with the out-of-focus spec- 

 tra as to be doubtful. This, of course, will pievent the lines which so suffer from 

 beins employed for measurement. The operation, moreover, takes up a gieat deal 

 of time. Kvery determination of a wavelength i-equires a series of photographs 

 which because of numerous particular observations occupies from 5 to 8 hours. 



The following i-emarks i-elate to the suitable airangement of the two tubes foi' 

 measuring wave-lengths in vacuo. The slit-jaws are microuietrically movable inde- 

 pendently of one anothei', so that the slit may be symmetrically widened and nar- 

 rowed. It thus becomes possible, in case the lines to be compared demand very 

 different exposures, to shoi-ten the exposure of the weaker line by symmeti-ically 

 widening the slit. A rotating shutter is provided for half closing the slit at times. 

 The camera-tube carries the caniei'a with a plate-holder for plates of 12 mm. square. 

 The plate-holder during the photographing is carried by a slide which is vertically 

 moval)le both micrometrically and by the free hand, and which is also capable of 

 beint' turned over in the plane of the plate very accui'ately through 180°, about 

 the optical axis of the camera lens as its axis of rotation. This is I'equisite in order 

 to ascertain the point at which that optical axis cuts the plane of the plate. Special 

 care has been taken to secure accurate centering of the lenses and their draw-tubes, 

 on which the accuiacy of the measures essentially depends. 



Whenever the intensity of light is sufficient, the concave grating should be 

 preferred in the determination of wave-lengths. It is important that with it there 

 is but a single surface at which the intensity of the rays is reduced. For the fewer 

 times the light is weakened, the stronger it will be when it reaches the photo- 

 graphic plate, other things being equal. Now the more energetically the extreme 

 ultra-violet light acts ui)on the photographic plate, the easier it will be to obtain 

 photographs of 'this light. Nor ought it to be forgotten that these short-wave rays 

 suffer nuich more by reflection than do long-wave rays. My recommendation of the 

 concave grating stipulates two provisos, viz.: 1st, that the matei-ial of the grating 

 should siifficieiitly reflect these short-wave rays; and 2d, that the grating-spectrum 

 should V)e sufficiently bright in this part. My photogiai>hs with plane grating and 

 lens prove that speculum-metal does in fact sufficiently reflect the short waves. But 

 the second condition relates to an individual character of the grating selected for 



