REVERSIBLE AND IRREVERSIBLE SYSTEMS UNDER INFLUENCE OF LIGHT. 349 
Crown carbons” used were 8 millims. thick on the top, 14 millirns. thick below. The 
carbons should not be placed one at the side of the other to throw more light in the 
horizontal direction, as this leads to a rapid deformation of the incandescent surface 
of the lower carbon, but should be placed one over the other, so that the crater should 
come into the centre of the carbon, with only a small inclination to the side which 
we wish to illuminate ; less light for the purpose, but more constant and for longer 
periods, is thus obtained, and trimming is required only at longer intervals. New 
carbons should never be used at once for the experiments, but only after some time, 
when the Rubens thermopile indicates that their intensity has come already to rest, 
i.e., that the carbons have reached their permanent shape. 
Direct measurements with the Rubens thermopile show that the intensity of the 
light sent by the arc horizontally in different directions differs as far as 30 to 50 per cent, 
according to the position of the crater with respect to the thermopile (and therefore 
also the quartz vessel). Similar measurements show that a variation of the voltage 
may lead to an essential increase of the space between the carbons, considerably 
change the brilliancy and intensity of the crater, and not change, or only very 
immaterially change, the amperes in the circuit. It is for this reason evident that at 
all times the intensity of the arc light cannot be indicated by the amperemeter, but 
must be directly measured with the Rubens thermopile (or bolometer), care being- 
taken that exactly the same most advantageous side of the arc which was turned to 
the quartz vessel should be also turned to the thermopile. For the reasons given 
above it is possible to get constant light for shorter periods, but it is very difficult to 
get every time anew the same constant intensity, as the variation in the shape and 
quality of carbons and the movement of the crater alone are more than enough to 
frustrate all attempts to get a constant light with the arc for longer periods. 
The Quartz Vessel .—In this the plates and solutions of the different combinations 
were exposed to light. The front and back plates (45 millims. by 45 millims. and 
removed 13 to 15 millims. from one another), which were exposed to light, \vere 
made of quartz, all other parts were of glass, cut to the correct shape and ground at 
their edges. The quartz plates were cut perpendicularly to the optical axis and were 
optically pure. The different parts of the vessel were stuck together by means of 
Crookes cement and the seams covered with asphaltum. An ebonite piece fitted well 
into the upper part of the quartz vessel, and had two holes drilled through it to 
allow two copper or glass rods, to which the plates were fixed, to pass through them. 
The copper or glass tubes were held on the top by means of screws or shellac. In the 
wider part of the ebonite piece below two thin parallel slots were cut to receive the 
plates and keep them always in position ; the front plate was about 1 millim. apart 
from the front quartz window. 
hhe Plates .—They were all taken 30 millims. wide, 42 millims. long; at the top 
the plates had strips about 3 to 4 centims. long, both, plates and their strips, having- 
been cut out of one and the same metal sheet. To the lower end of the thin strips of 
