ACTION OF LIGHT ON SELENIUM. 
347 
Fig. 3 represents another experiment, with the marked end heated ; and fig. 4 an 
experiment with the unmarked end heated. 
It appears from most of these experiments that on heating the marked end there is a 
current from P to S at that junction, and on heating the unmarked end there is a current 
from S to P at the heated junction. Hence, on heating the whole piece, it would appear 
that a continuous current is produced. The whole piece may be used as a thermo- 
electric pile, and would form a very delicate thermometer, since the exposure of either 
end of this piece to heat causes a current in it from the marked to the unmarked end. 
This agrees with the results of exposure to the lime-light, as recorded in the paper. 
On referring to the experiments with lime-light (p. 337), we see that, when the light 
was brought to a focus on any part of this piece, there was a current in it from the 
marked to the unmarked end. 
In one or two other pieces of selenium we have found that heat applied at one end 
produced a current from platinum to selenium, and heat applied at the other end pro- 
duced a current from selenium to platinum ; so that, on the whole being heated, there 
was a current from the first to the second end through the selenium. 
The other pieces of selenium whose thermoelectric properties have been tested are 
Nos. 18, 20, 21, 22, 23, 24, and 25. At each junction of each of these pieces it was found 
that the application of heat gave rise to a current from selenium to platinum, so that 
the process of annealing had in these cases raised selenium from the bottom and placed 
it above platinum in the thermoelectric scale. 
In most instances the currents due to heating the two ends did not differ much from 
one another. No. 21 is exceptional. 
On warming with the hand the unmarked end there was a current of 1 0 only, whereas 
on applying the same source of heat to the marked end there was a current of 150. 
With the hot blast the values were 15 at the unmarked end and 65 at the marked 
end. 
It seems to be pretty well established by these experiments that the more complete 
the process of annealing, the higher selenium is raised in the thermoelectric scale ; and 
on referring to the Table of Resistances (p. 323), we see that the more the resistance of 
a piece of selenium is diminished the higher is it raised in the thermoelectric scale. 
Thus all the earlier pieces, except Nos. 11 and 15, were of high resistance. 
Nos. 11 and 15 are of smaller resistance, but appear to have been unequally annealed 
at the two ends, so that one end is above, or near, platinum, whilst the other is below 
platinum in the scale. 
All the later specimens appear to have been well annealed at both ends, and are of 
very small resistance, except No. 21, which was of high resistance, and which appears 
to have been only partially annealed at one end at the time of the experiment. 
It would seem from these experiments that the process of annealing raises selenium 
in the thermoelectric scale at the same time that it diminishes its electrical resistance. 
