THE ARSORPTIOX AND KMISST01>T OF All!. 9 



that ttts to the slit-carriage of tlie spectrograph is connected air-tight with the l.ody 

 of tlie absorption apparatn-^ ; that is, with a tliick-walled hollow cylinder of brass. 

 In the interior of the hollow cylinder there is a tube of cast-steel which can be 

 displaced along the axis by the aid of a railled-edged disk provided with a female 

 screw-thread within. The measure of the displacement can be read upon the two 

 scales visible in the picture; namely, the whole millimeters upon the longitudinal 

 one, and the tenths of niillimeters upon the circumferential one. Uiion the face of 

 the steel tube tuinied toward the covei', e, is fixed tlie tluor-sivir window that is 

 moved with the tube, while u[u)n the inner plane surface of the cover, e, the other, 

 which is fixed, abuts uiton the centi'al light-opening, which leads to the slit. The 

 two windows remain in pai'allel planes, whatever their distance a[)art maybe. The 

 air-tight connection between the steel tube and the brass cylinder, a (Fig. 10), 

 is effected by means of a ring of sole-leather turned u[)on the lathe and having a 

 cross-section 3 millimeters square, this ring being pressed against tube and cylinder 

 by means of an annular female screw. For the same purpose, the greatest care 

 was taken in preparing the steel tube and its brass guide, to render the cylindi'ical 

 sui'faces precisely circular. To the milled disk, h, the Geissler tube, c, is fastened by 

 its rim. As the medium of attachment the above-mentioned mixture of wa.x and 

 vaseline is employed, a small amount being laid upon the rim of the tube. In this 

 case the tube cannot be adjusted by a steel pin as it is adjusted to the cover <f ; and 

 the following process has to be resoiied to: The apparatus by the annular shoulder 

 of e is stuck into a well-fitting ring, not too light, in wliicli it can be steadily turned 

 about its axis. A suitable pin-hole is uow ai'ranged, oi' bettei-, a shoi-t-focussed read- 

 ing-telescope is directed so as to place the middle of the mouth of the capillary 

 upon the cross-wires. Then, by turning the appai-atus round through 180', and 

 shifting the tube by a suitable amount, it will easily be accurately centered. Lest 

 the centering so attained should be lost, Avhich at first, before the exhaustion, might 

 easily happen, while the tube is only attached by the grease on its rim, four guides, 

 (/, sliding in the radial grooves of the disk, b, are delicately pushed up against the 

 wall of the tube, and are secured there by their screw^s, / Fui-ther, to render the 

 centering quite secure, the support (/ is required. The lateral part of the Geissler 

 tube that carries electrodes should rest against this support throughout the opera- 

 tion of centering, and should continue to do so thereafter. These precautions 

 would, indeed, be needless if the rim of the Geissler tube were precisely on 

 the circumference of the base of a right cone having its vertex at the centei- of the 

 mouth of the capillary. But this condition is not fulfilled, an<l could be only 

 at great cost ; so that "the cheaper arraugement with the support, ;/, is preferable. 

 What is indispensable is that the bearing surface of the rim of the tube shoulu 

 be optically plane and should lie in a plane to \vhich the axis of the capdlary 

 is perpendicular; the former condition being required for air-tightness, and the 

 latter for the uniform illumination of the slit. 



The fillincr of the Geissler tube is performed through the specti'ograph, which 

 is for this purpose connecte.l with an apparatus for the evolution of the gas 

 wanted The gas then flows in through the cock, i, and through a narrow channel 



