476 BELL SYSTEM TECHNICAL JOURNAL 



was left almost alone upon the scene. These results signify that an 

 800-volt electron operates ionization in hydrogen by detaching an 

 electron from a molecule; other kinds of ions appearing in the gas are 

 due to subsequent adventures of these ions. 



The method of H. D. Smyth is suitable for investigations into this 

 question. In apparatus such as his, hydrogen bombarded by (say) 

 40-volt electrons is found to contain all three ions H+, H2+ and H3+; 

 but as the density of h>'drogen is reduced, the first and the last of these 

 ions become less abundant and finally insignificant by comparison with 

 the ion H2+. As the bombarding- voltage is reduced towards the value 

 (about 16) at which ionization commences, all three kinds of ions be- 

 come less plentiful; but with high densities and sufificiently sensitive 

 apparatus it is found that Hs"^ makes its appearance as early as H2'*', 

 and there is no reason not to suppose the same about H"^. In hydrogen, 

 therefore, and also in nitrogen, it is agreed that an electron-impact 

 against a molecule results, if in any sort of ionization at all, in the de- 

 tachment of an electron from the molecule, not (for instance) in a dis- 

 sociation into one ionized atom and another atom ionized or neutral. 

 Dissociation and new sorts of association may result from the further 

 adventures of this molecule-ion in the gas. In certain compound gases^* 

 of which the molecules consist of two or more atoms of different kinds, 

 there is reason to expect the contrary: that is, that an electron-impact 

 against a molecule would result directly in splitting it into a positively- 

 charged atom (or group of atoms) and a negatively-charged atom (or 

 group of atoms). Certain experiments indicate this: in ZnCU vapor, 

 for instance, CI atoms bearing an extra electron and ZnCl molecules 

 minus an electron are found as soon as ionization commences; but the 

 question can hardly be deemed settled until comparative measure- 

 ments are made at various gas-densities. 



From these experiments it follows that a measurement of the energy 

 just sufificient to produce ions in a molecular gas, while interesting in 

 itself, can hardly be interpreted without additional data regarding the 

 nature of the ions produced. There are other difficulties in determining 

 ionizing-potentials in such gases; for instance the likelihood that the 

 hot filament will itself dissociate the gas. The published determina- 

 tions are frequently contradictory; the various published values for 

 the ionizing-potentials of hydrogen, for instance, form one of the most 

 discouraging sets of irreconcilable data to be found in physics. 



According to thermochemical measurements the "heat of dissocia- 

 tion" of hydrogen, in other words the energy-difference between a sys- 

 tem of two free H atoms and an H2 molecule, amounts to 3.5 equivalent 

 " Those designated by chemists as hcteropolar. 



