IONIZATION POTENTIALS IN DISSOCIATED GASES. 217 



In atomic hydrogen the secondary spectrum was entirely absent, 

 but the series spectrum appeared strongly when the arc struck at 

 13.5 volts and could be detected down to lo.i volts. 



These results confirm the prevailing opinion that the series spec- 

 trum is due to atoms and the secondary spectrum is due to molecules. 

 The voltages at which the spectrum appears in atomic hydrogen are 

 exactly those to be expected from Bohr's interpretation of the series 

 formula. The exciting voltage in molecular hydrogen is that to be 

 expected if the critical 16.3 voltage is interpreted as dissociation of 

 the molecule plus ionization of one of the atoms as the result of 

 single electron impacts, giving 16.3 — 13.5 = 2.8 volts as the heat of 

 dissociation in equivalent volts. This is almost exactly the value 

 given by Bohr's theory, but is lower than the value 4.06 volts calcu- 

 lated from Langmuir's measurement of the heat of dissociation of 

 hydrogen.* 



The interpretation of the secondary spectrum is puzzling. It is 

 certainly due to molecular hydrogen, but the nature of the emitting 

 molecule is uncertain. It is probably not due to neutral Ho mole- 

 cules, since these have no absorption in the visible spectrum. Hg 

 molecules are known to be present in fairly large concentration in an 

 arc, but more in a high voltage discharge in a large vessel at low 

 pressures. H3 molecules are also known to be present. Possibly 

 one of these may give rise to Fulcher's Group I. and the other to his 

 Group II., but evidence on this point is rather conflicting. An obser- 

 vation of possible interest in this connection is that we found the 

 secondary spectrum to disappear, when the outer tube was heated, at 

 temperatures certainly too low to have dissociated H^ molecules to 

 any great extent. This suggests that the molecule responsible for the 

 observed secondary spectrum lines is less stable than H^. As far~as 

 this is concerned, it may be either H^ or H3. The absence of a 

 Doppler shift for these lines may possibly point to the neutral H3 

 molecule as the agent. In this connection it may be mentioned that 

 R. W. Wood has very recently succeeded in drawing off pure atomic 

 hydrogen from the center of a long Geissler tube, where only the 

 series spectrum is visible, and he finds that the presence of a tungsten 



* Am. Chcm. Soc. Jour., 37, p. 417, 1915. 



