PRODUCED BY HOT PLATINUM IN DIFFERENT GASES. 61 



The only general conclusion which these numbers point to is that the work required 

 to liberate a positive ion tends to be smaller than that required to liberate a negative 

 ion. This is probably due to the fact that the negative ions are corpuscles, whereas 

 the positive ions are associated with a certain amount of matter. 



Throughout this paper, the view has been maintained with regard to the negative 

 ioiiisation that it is due to the escape of corpuscles which are present in a more 

 or less free condition insido the metal, the variations in the amount of negative 

 ioiiisation caused by different gases being due to the effect the gases have on the 

 amount of work necessary for a corpuscle to get through the surface. The positive 

 ionisation, on the other hand, has been supposed to be caused entirely by traces 

 of gas absorbed by the metal. There is, however, another view of the origin of 

 the positive ionisation which cannot be lightly dismissed. This is that it is due 

 to the escape of positive ions which are moving freely inside the metal in much 

 the same way as the negative corpuscles have been supposed to be, and that the 

 effects of different gases are simply due to the changes they produce in the 

 surface. There are a number of considerations which make this hypothesis plausible 

 at first sight. It would give an obvious explanation of why the positive obeyed 

 the same temperature law as the negative ionisation, and would also account for 

 the work necessary for a positive ion to esca{>e being in general smaller than that 

 for a negative ion. For since the particles at the surface of the metal are all at 

 the same temperature, the energy of a moving ion will tend to become equalised 

 with the average value for the metal at each collision, so that, to take an 

 exaggerated case, an ion which had found its velocity reduced to nothing, owing 

 to the work it had done in getting through a certain fraction of the surface, would 

 get a fresh start if it made a collision. Thus the work function in question will 

 l>e equal to the average amount of work to be done between the last collision 

 inside the metal and a point outside ; it will thus be greater for an ion with a long 

 free path, such as a corpuscle, than for one with a short free path. On this view, the 

 effects of different gases are to be explained by the changes they produce in the work 

 required for an ion to escape from the surface layer. This might occur by the 

 formation of an electrical double layer, or simply by a change in the physical 

 properties of the bounding region, or by both. The former would act differentially 

 on the positive and negative ionisations, whilst the latter might be expected to act in 

 the same way, though not necessarily to the same extent, on both. A combination of 

 the two could obviously be made to account for any observed simultaneous change in 

 tlir two leaks. If we calculate the number of ions per cubic centimetre inside the 

 metal from tin- ionisation in different gases according to the formula previously* given 

 by the author, we get the numbers given in the columns N + and N_ in the last table. 

 These mnntars do not mean much, owing to the ignorance of other data which really 

 enter into the calculation, but on the view at present discussed they should range 



* 'Camb. Phil. Proc.,' voL 2, p. 286 it Kq. 



