M. E. Goldstein on Spectra of Gases. 341 



white and surrounded by an aureola, does not appear suddenly if 

 the pressure of the gas under low pressure is increased slowly. We 

 observe first a yellow spark, which is of the same thickness and 

 which can be seen isolated in narrow tubes. The rotating mir- 

 ror shows that the discharge is discontinuous. The spectrum 

 consists of bands. Its thickness is much less than that of strong 

 discharges of the Leyden jar which give line-spectra. 



In wide tubes we see discharges going out from the positive 

 electrode which are of a red colour. They look like branches 

 of a tree, and appear to be about as thick as a thread. The spec- 

 trum consists of bands. 



All this shows that the supposition cannot be true which as- 

 sumes that the band-spectrum is only given out by thick layers, 

 and that it is produced by a superposition of light of feeble in- 

 tensity and absorption in the gas itself. 



Wullner's hypothesis on the production of band- and line- 

 spectra cannot, therefore, be maintained. 



A few objections, and the proof that for some gases Wullner's 

 propositions can a. priori be shown to be untenable, must be de- 

 ferred to another occasion. 



Wullner assumes in nearly all his papers that certain changes 

 in the spectra are produced by certain changes in the pressure. 

 Although he believes the pressure at which these changes take 

 place to be different according to whether it is the simple cur- 

 rent of the induction-coil that passes through the gas or the 

 discharge of the Leyden jar, yet these pressures are given as 

 Constant quantities for the same mode of discharge and the same 

 chemical composition of the gas. I cannot here refer to the lite- 

 rature upon this subject, but merely give my experiments. 



I constructed a tube, one end of which was closed hermeti- 

 cally by a caoutchouc stopper and mercury joint. A knitting- 

 needle was inserted in the stopper and used as electrode, which 

 could be approached to and removed from the second electrode, 

 which was fixed. When the two electrodes were near together 

 the band-spectrum was seen. As the distance increased, the 

 band- spectrum became more feeble, and at last the line-spec- 

 trum appeared. As I brought the two needles together again, 

 the same phenomena were seen in the inverse order. 



As already mentioned, the spark discharge with line-spectrum 

 has sometimes a little reddish point, the spectrum of which 

 consists of bands. 



It is found sometimes that a tube shows a band-spectrum 

 if the spark passes in one direction, and a line-spectrum if it 

 passes in the other. 



With discharges of jars the resistances can be so arragned 

 that in rarefied air the sparks just give a band-spectrum while 



