PHYSICS: MILLIKAN AND SOUDER 
23 
alkali metals, lithium, potassium and sodium, were arranged on the 
periphery of a wheel which could be rotated electro-magnetically by an 
electromagnet outside the chamber. Fresh surfaces of these metals 
were produced by means of a knife operated from outside by another 
electro-magnet. A Hilger thermopile used in connection with a Cob- 
lentz galvanometer sensitive to 1 X 10~^i amperes intercepted a portion 
of the beam and permitted energy measurements to be made simul- 
taneously with the measurements of the photo currents. 
The apparatus was the same as that used in the photo-electric deter- 
minations of Planck's 'h^ with the aid of the alkali metals.^ As indicated 
above, the incidence was always normal. 
The photo-currents are all saturation currents with — 12 volts applied 
to the lithium. 
The pressures at which the observation in the experiments corre- 
sponding to curves in figure 2 were made ranged from 0.04 to 0.0001 mm. 
While the rate at which the maximum grows depends somewhat up- 
on the pressure, any individual curve is found at a given time to be not 
essentially modified, so far as the phenomenon here under consideration is 
concerned, by changes in pressure within the specified Hmits. 
Changes in photo-sensitiveness bearing some similarity to these were 
observed by Pohl and Pringsheim^ in the cases of barium, aluminium, 
magnesium and calcium but the maxima observed by them were not 
shown to be in the neighborhood of a wave length at which a selective 
effect had been found. Indeed Pohl and Pringsheim did not interpret 
their results as meaning that these four metals possess any selective 
effect at all. It will be observed, however, that the maximum in curve 
2 coincides, within the limits of accuracy of Pohl and Pringsheim's de- 
termination, with the frequency 280 ju/x at which they located the selec- 
tive effect of lithium. It is not at all likely, therefore, that our normal 
effect can differ in any essential particular from their selective effect in 
lithium. It is true that we have not shown that our Hthium surface shows 
the characteristic of yielding an energy-frequency curve having a maxi- 
mum when light is polarized perpendicular to the plane of incidence 
combined with no maximum when it is polarized parallel to the plane of 
incidence. But this was not the criterion which Pohl and Pringsheim^ 
used when they discovered the existence of the selective effect in Hthium. 
Their criterion was simply the existence of a pronounced maximum at 
280 fxfx when they used obHque incidence and polarized Hght. It is 
scarcely conceivable that our maximum at 280 /x/x would disappear if 
we made our tests under precisely the same conditions which they used. 
Our experiments have, as yet, thrown no light on the cause of the ex- 
