AND STATICS UNDER THE INFLUENCE OF LIGHT. 
359 
The thermopile has 40 thermo-junctions # of iron and constantan, on an area 
of 2 cent inis., 20 of them in one line being exposed to the light, and 20 (10 in one 
line on each side of the line of the junction exposed to the light) remaining in 
the dark. The wires used are very thin (OT to 0 15 millim.), and the places of the 
junctions (those exposed to the light as well as those remaining in the dark, are 
hammered to little thin round plates (0 - 5 to 0'8 millim. diameter) so as to increase 
their sensitiveness to variation of temperature, and are made dead black. Before 
the line of junctions exposed to the light there is a small and then a larger cone of 
polished nickel, placed opposite and near the line of junctions, which causes a greater 
quantity of light to fall upon the junctions. The electromotive force of the thermo¬ 
pile is O’OOIOG volt per 1° C. 
In connection with the Rubens’ thermopile a Crompton’s dead-beat galvanometer 
was used, in order to avoid the numerous disturbances experienced by galvanometers 
other than that of the D’Arsonval type, and its sensitiveness was adjusted to get a 
sufficiently great deflection of the spot of light—not to get the greatest possible 
deflection, but to arrange the measurements of intensity so that after they were 
brought to a great accuracy, say of 0T per cent., the measurement should be made in 
an easy, steady, and reliable manner, and independent of numerous obstructive 
influences lying beyond the thermopile itself. This meant that special precautions 
had to be taken with both the thermopile and the galvanometer. 
Whatever form is given to the thermopile, the next precaution, especially when a 
sensitive thermopile (or bolometer) is used, is to protect it from the influence of the 
surrounding medium, so as to secure concordant results. The junctions which are 
exposed to light are necessarily much more exposed to all sorts of air currents than 
those protected and covered from the light, and are also more subjected to the 
influence of the variation of the temperature of the room. The unavoidable 
continuous use of the acetylene flame to be measured by the thermopile from time to 
time, produces air currents, set up by local differences in the temperature of the room. 
The sensitiveness of the whole arrangement was necessarily so great that if the hand 
was put before the thermopile at the distance of 1 or 2 decimetres, the deflection of 
the galvanometer was considerable. This difficulty was evaded in the following 
manner : — 
The thermopile was enclosed in a cylinder of thin copper (about 1 millim. thick ; 
the two circular sides of the front and back being made of the same thin copper. 
In the centre of the front side a circular opening of about 4\ centims. was cut and in 
this a copper tube of the same diameter (for the quartz plate) was fixed. In the 
centre of the circular copper plate behind, a thick but very narrow tube (of about 
3 or 4 millims. internal diameter) was fixed for the purpose of seeing and directing 
the thermopile upon the light of the burner. The thermopile was fixed on a piece of 
* Through an accident our thermopile had only 38 junctions. 
