of the Selentum Cell. 29 
slabs of Atchison graphite, an especially pure form, were 
separated by a layer of mica, securely bound and shellacked 
together and then baked so as to forma compact mass. While 
the cell was still hot a very thin layer of molten selenium was 
spread over one side, as shown in the diagram. In the 
average the selenium layer was 15 mms. long, 1 mm. wide, 
and 0°08 mm. thick. Before being applied to the cell the 
selenium had been ground up to a fine powder in an agate 
mortar with 3 per cent. of a definite selenide *, for this pro- 
portion of selenide in selenium, as shown by Bidwell f, is 
approximately that necessary to develop the highest sensi- 
bility in a cell. The selenides of copper, lead, mercury, and 
silver were used for the reason that they form with great 
readiness and conduct electricity well. 
The details followed out in the process of forming and 
annealing were similar to those described by Bidwell f, and, 
in general, this procedure yielded cells of fair, though not 
excessive sensibility. In the average a cell had a resistance 
of about 20 megohms, and, when new, increased its con- 
ductivity from three to twelve times upon being illuminated 
by a 32 candle-power incandescent lamp at a distance of 
30 cms. 
Apparatus, 
The spectrometer was of the usual fixed-arm type, designed 
for work in the visible and infra-red spectrum. 
Light from a Nernst lamp L (fig. 2) was dispersed by a 
rock-salt prism P, the resulting spectrum being brought to a 
focus on a brass plate B which contained a vertical slit. By 
rotating the prism, light of any desired wave-length could be 
made to pass through the slit. Directly behind this a Rubens 
thermopile, T, and a selenium cell, 8, were rigidly mounted 
on a vertical bar, pivoted near its lower end and working 
between stops so that the selenium cell and thermopile could 
be mutually interchanged in position. By the use of the slit 
at B there is no question but that light from the same region 
of the spectrum which fell upon the thermopile would also 
illuminate the selenium cell, thus avoiding numerous com- 
plications which are certain to arise if the one instrument does 
not replace the other with absolute exactness. IK is a device 
placed in the path of the parallel beam to cat down the 
intensity of the energy. While a rotating sector would have 
* Obtained from Dr. Theodor Schuckhardt. 
+ Shelford Bidwell, Phil. Mag. vol. xl. p. 233 (1895), 
t Ibid. 
