13 



Again, in some experiments described by C. D. Child (5) 

 upon the arc through mercury vapour, clusters very similar 

 in nature to those described by Burke are met with, ajid their 

 behaviour was only explained on the hypothesis that recom- 

 binations between positive and negative ions were not always 

 complete or stable, and that the clusters so formed were, 

 under suitable conditions, readily broken up, giving rise to 

 ionization. 



It may not be out of place at this point to remark that 

 the nucleus of such clusters might possibly be formed either 

 by the partial recombination of a positive and a negative 

 ion which have each had time to become attached to neutral 

 molecules of the gas or by the immediate partial recombina- 

 tion of a parent molecule and the electron ejected from it by 

 the process of ionization. 



Just as this paper was near completion the results of an 

 investigation by Von Erich Barkow <6' came to hand, in which 

 the existence of neutral pairs of ions, evidently produced in 

 the former manner, offers a ready explanation of the effects 

 observed. 



It was, however, more with the idea of finding out whethei 

 a cluster is formed in the latter way that the work of this 

 present paper was carried out. When two oppositely charged 

 ions, approaching ea^h other, are at a distance apart "r" (in 

 the present case "/" is of the order of the molecular free path), 

 the particles will revolve in closed orbits around each other, 

 if the kinetic energy due to their relative motion is less than 



- (e being the charge on an ion). An electrical field will 



tend to disturb this motion, ajid it is one of the purposes of 

 the present paper to decide whether some such state of semi- 

 combination can persist for an appreciable time, and a separ- 

 ation of the ion be eventually effected by an electrical field, 

 or by a field conjointly with other c-auses. 



Returning to the experiment described by Rutherford, 

 little detail information is given in regard to the apparatus 

 used. It is stated that instead of measuring the current with 

 the uranium oxide covering one electrode, the air which had 

 passed over the uranium was forced between two concentric 

 cylinders between which the electromotive force was acting. 

 It seemed probable that the apparatus described by him 

 earlier in the same paper (p. 144) had been used, in which 

 air, after passing through cotton wool and over the uranium 

 surface, was forced through a wire gauze, and then between 



(5)Phvs. Rev., vol. xxii., pp. 221-231, April, 1906. 

 (6) Annalen der Physik, p. 317. No. 7. 1907. 



