514 Mr Campbell, Discontinuities in Light Emission. 



accumulators was amply sufficient for the purpose : preliminary 

 observations had been made on the fluctuations obtained when 

 the two photoelectric cells were illuminated by two independent 

 lamps of this nature, one of which was placed in front of each 

 cell. But in order that the main experiment might be carried 

 out, it was necessary to divide the light from a single lamp into 

 two beams : and such a division was rendered very difficult by the 

 comparatively large area from which the light was emitted. It 

 was impossible, for instance, to obtain a parallel beam by means 

 of a lens and to divide that beam by reflection, as had been 

 attempted in the case of the Nernst lamp. Since, again, it was 

 necessary that the two beams, falling on the two cells, should 

 proceed originally from the same part of the incandescent surface, 

 it was impossible to obtain such beams as were required by the 

 simple expedient of placing the lamp between the two cells, so 

 that light emitted from the lamp in different directions should 

 fall on the ceils : for, with a wire lamp, the light emitted in 

 different directions does not come from the same part of the 

 incandescent filament. 



§ 2. An attempt was then made to attain the desired object 

 by means of the following device. A 110-watt Osrara lamp with 

 a bulb 12 cm. in diameter was silvered thickly on the outside, so 

 that the walls became very nearly perfect reflectors. Two holes 

 of 1 cm. diameter were made in the silver, and the light emergent 

 from these holes was allowed to fall on the two cells. It is easy 

 to see that the light emerging from either of the holes is made up 

 of rays emitted from every part of the wire, reflected in most cases 

 many times from the silvered walls. Accordingly the two beams 

 emerging from the two holes are " dependent " in the sense of the 

 word employed in the previous paper. The device was so far 

 successful that, on looking into the bulb through one of the holes, 

 no sign of the wire could be seen: the centre of the bulb ap- 

 peared to be occupied by a uniformly incandescent body, the 

 shape of which, in accordance with theory, was approximately 

 that of the surface enveloping the wire filament. 



The device was, however, unsuccessful, for the bulb became so 

 hot that it was necessary to keep it cool by immersion in a water 

 bath. After a few hours in such a bath the silver began to detach 

 itself from the walls, whatever method of silvering was used : 

 covering the silver with several kinds of varnish did not improve 

 matters materially. 



§ 3. An attempt was then made to substitute mercury for 

 silver. The lamp with the walls carefully cleaned was immersed 

 in an iron vessel containing clean mercury. Through two holes 

 in the iron vessel passed two glass tubes with ground ends held 

 against the surface of the bulb by springs. These tubes provided 



