1(1 TJIK AUStdiPTIOJf AND EMISSION OK ATR. 



continuous with this, aud having its orifice near the Geissler tube. When the cock 

 is closeil, the tube continues to hold its gas while the spectrograiih is exhausted. 

 Although the gases of the tube and of the absor[»tion-chainl)ei' are only separated 

 Ijy the steel tube, I have never found that even hydrogen eiiteied the latter, dift'usi- 

 l)le as that gas is. The gas whose absorption is to be investigated flows into the 

 absorption-chamber through the tube, //, aud leaves it through a second tube placed 

 diametrically aud having its further end a little below the surface of some mercury. 

 The mercury forms an escape-valve effectually preventing the entrance of atmos- 

 pheric air, while a very slight over-pressure will cause an outflow. 



The circumstance that the steel tube and its bearing are of different materials 

 has the disadvantage that, owing to the different expansibilities of the two luetals, 

 the steel tube in the cooler part of the year does not move as readilj' as it should ; 

 and besides, the bearing may be injured. The beai'ing was originally fitted with 

 great accuracy at the mean temperatuie of the loom. E.xperiments that were 

 instituted upon the appearance of this inconvenience showed that the reduction of 

 the temperature by only a few degrees Celsius would suflSce to damage the beaiing 

 and sometimes even the tube itself, if this were moved. Since the whole utility of 

 an apparatus upon the design and construction of which no little pains had been 

 bestowed was thus placed in jeopard}', I provided it with a little heater consisting 

 of a water-jacket, Tc, to enclose the brass cylinder, a, at the point of the beai'ing of 

 the steel tube, which water-jacket, by the aid of a funnel-shaped projection that 

 stands vertical when in use, could be filled with water, aud then, by means of the 

 warming-tube, m, could be warmed to 27° C. by a little alcohol flame. At this 

 temperature, or even a couple of degrees higher, the ap[)aratus works faultlessly. 



It may be thought that it would have been better to construct the whole 

 api)aratus of one metal. But there were certain technical objections to this into 

 which it would lead me too far to enter here. It may be added that the a[)paiatus 

 has no pretension to universal applicability, being restricted to the study of such 

 gases or vapoi-s as are chemicall}- indift'erent to steel and to fluoi-spar. The pre- 

 paration of the atmospheric gases proceeded throughout in vessels connected by 

 ground-glass fittings. The same thing is true of experiments with the gases which, 

 as will be shown, enter as impurities into the Geissler tubes, such as CO aud C^, li„, 

 whose influence, as appealed in the course of my investigation, required searching 

 attention. 



For photographic experiments ultra-violet plates prepared according to my 

 process could alone be thought of, since the gelatine plates have not sufficient 

 sensibility for rays beyond 185 /'/'. Foi- although the gelatine plates are not 

 completely insensitive for those rays, yet they are so inefficient that nothing more 

 than a l)a?ely visible and worthless continuation of the spectrum appears in any 

 case. The cause of this want of sensitiveness is, as I formerly proved, the want of 

 transparency of the gelatine. By keeping this fact in view, I have succeeded, by 

 persistent experimentation, in reducing the proportion of gelatine on my plates 

 until it ceases to do any harm. The thickness of the sensitive layer is, at the same 

 time, tlius reduced to but one fifth of what it is upon the ordinary gelatine plates. To 



