the Nature of Spectra. 91 



The tubes were then fixed insulated and horizontal in an 

 air-bath of iron, 16 centims. deep, 11 centims. wide, and 14 

 centims. long, by means of glass tubes which passed through 

 the two side walls of the bath. The front and back walls of 

 the bath were formed of plates of mica; and another mica 

 plate served as a cover. The heating was effected by a gas- 

 burner placed beneath. The progress of thQ phenomena was 

 examined both during the heating and especially during the 

 cooling. The temperatures were measured by an ordinary 

 thermometer placed immediately on the Geissler tube. 



For the examination of the spectra either a Bunsen spec- 

 trum-apparatus or a Browning direct-vision spectroscope was 

 used. To the former, in the place of the hair cross in the ob- 

 serving-telescope, a slit-diaphragm was applied, in order to 

 blind off certain portions of the spectrum. 



A middle-sized Ruhmkorff induction-apparatus served as 

 the source of electricity, its primary current being furnished 

 by 3 or 4 Bunsen elements; a small induction-apparatus, 

 however, or a Holtz electrical machine was also employed. 

 In all cases the same phenomenon was shown. At a not too 

 high temperature (between 100° and 200°) the lines corre- 

 sponding to the gas contained besides the mercury in the 

 Geissler tube disappeared. At the same time a clear differ- 

 ence between the two electrodes was observable. When, on 

 slowly heating, at the positive electrode and in the whole of 

 the luminous tuft issuing from it mercury-lines alone were 

 already to be seen, at the negative hydrogen- and nitrogen- 

 lines were still distinctly shown, which at higher temperatures 

 (at which larger quantities of mercury were present in the 

 gaseous space) likewise vanished. 



The best idea of the progress of the phenomenon will be 

 given by the following series of experiments. The tube was 

 of uniform width, the electrodes spherical ; the large induc- 

 tion-apparatus served as the source of electricity. The pres- 

 sure of the enclosed air amounted to about 10 millims. The 

 tube was first heated to 240°, and then permitted to cool 

 slowly. 



Above 240°, everywhere only the mercury-lines are to be 

 seen, both in the middle of the tube and at the electrodes. 

 The discharge issues as a zigzag line of light from the fore 

 end of the positive knob, and plays round the negative, without 

 any perceptible dark space between. At the wire to which 

 the negative knob was fastened, no light is to be seen. 



Between 230° and 210° the nitrogen-lines, together with 

 those of mercury, begin to appear at the negative electrode ; 

 and at the same time the dark space unfolds itself, the wire of 



