292 Dr. Norman Campbell on the 



apparatus are those imposed by the batteries ; but there are 

 almost certainly other differences of potential. It appeared 

 that the value o£ Y x — Y 2 for which I becomes zero is not 

 always exactly 0, and yet it is presumably when: the rays 

 reach the P with zero speed that I is zero. Accordingly if 

 v is the value of Y 1 — Y 2 for which I is (the value, that is, 

 in the neighbourhood of Yi— Y 2 — for which this occurs ; 

 there is another value, as we shall see), the speed of the rays 

 was taken to be Y x — Y 2 — v. v was usually negative and about 



I volt. It is possibly due to a neglect of v that other ob- 

 servers have found values for the ionization potential less than 



II volts. 



7. The result of the measurements made immediately after 

 the apparatus was set up and evacuated is shown in fig. 2, 

 curve A. Its form agrees generally with that of Gehrts. 

 It differs from that curve in showing much less distinctly the 

 maximum of R at about 5 volts, in showing the minimum at 

 11 volts rather than at 10, and in rising rather more quickly 

 after that minimum. The platinum in this condition is said 

 to be in state A. 



Heating P to about 1000° for three days, when it gave off 

 large quantities of gas, which were removed, did not change 

 the form of the curve. Pure oxygen was then admitted to a 

 pressure of about 1 mm. and heating resumed ; a day of this 

 treatment had no effect, but later within the space of 1^ hours 

 the curve changed from the form A to the form B (fig. 2). 

 It will be observed that all the values of R have decreased, 

 even those which depend only on true reflexion *; but the 

 values for the higher speeds of the incident rays, which are 

 due to ionization, have decreased markedly. The curves do 

 not perhaps show well how marked the change is ; it is more 

 noticeable if expressed in another way. When the speed of 

 the incident rays is increased sufficiently above that necessary 

 to produce ionization, a point is reached at which the electrons 

 liberated are equal in number to those in the incident rays, 

 so that the current flowing from P is zero. The value of" 

 Yi — Y 2 for which this change in sign of the current from P 

 occurs is, of course, independent of the magnitude of the 

 current, and, like all null-points, can be detected and repro- 

 duced with considerable accuracy. In state A this value of 

 Yi — Y 2 (which will be called Y ) was 18 volts; in the 



* I am not certain that this decrease in R for the small speeds took 

 place at exactly the same stage as that for the high speeds ; unfortunately 

 observations at the low speeds were not made on every occasion during 

 the preliminary heating. This change is probably connected with a 

 roughening of the surface of the metal, such as is known to occur with 

 prolonged heating. 



