AX1> STATICS UNDER THE INFLUENCE OF LIGHT. 
361 
exceedingly small, while all the junctions should be at the same temperature in 
the dark. 
There still remained other sources of error. In the first instance, there was the 
quartz plate fixed in front of the larger tube closing up and protecting the inner air 
space from the air currents of the room. The quartz plate is exposed on the inside 
to the temperature of the air in the inner cylinder, on the outside to that of the 
room. There was additional reason why the temperature of the water between the 
cylinders and of the air space in the inner cylinder had to be brought to the 
temperature of the room; because it is evidently necessary that the thermopile 
should be exposed to the same temperature on all sides. In the second instance, the 
quartz plate and the inner air space between the same, the cone and the copper ring 
upon which the quartz plate is fixed, are being heated by the rays of light while 
measurements of the intensity of light are carried out, the rest of the inner cylinder 
remaining unexposed. 
To counteract these sources of error the air space exposed to the light was made 
very small in comparison with the total air space of the inner cylinder. The narrow 
circular copper plate (cp, fig. 5), on which the quartz plate (</) was fixed, was also 
water-jacketed by means of the india-rubber tube (y), so that the four ends of the 
quartz plate, equal to about one-third of the total surface of the plate, were directly 
cooled by the narrow circular plate. In front of the quartz plate a large wooden 
screen was placed, and to its back (not to be seen in this drawing) a copper 
cylinder was fixed of about 1 inch in thickness and of the same diameter as the 
outside copper cylinder of the thermopile, filled with water of the same temperature 
as that contained between the two copper cylinders of the thermopile. 
In filling all the apparatus with water from the main, the space between the two 
copper cylinders was first filled with water through E. From the top of the outer 
cylinder the water passes through the india-rubber tube (y) behind the quartz plates 
and through the india-rubber tube (f) to the lower part of the cylinder which forms 
the water screen. From the top of the water screen the water passes through an 
india-rubber tube and pewter pipe ( p) back to the tank, and then the run of water 
is stopped. The india-rubber tubes f and y allow the screen to move up and down, 
and when it is down the water screen quite covers the whole front surface of the 
copper cylinder of the thermopile, being removed from it by only about half a centi¬ 
metre. Direct tests of these arrangements showed them to be successful; after the 
light has been used and the quartz plates again screened, the deflection returns to the 
same zero quickly (the thermo-electromotive force in the dark never exceeding about 
2 millims.), and remains so for any length of time, and the same deflection is obtained 
every time the screen is opened again. It takes only a fraction of a minute for the 
spot of light to attain its maximum deflection or return to zero. 
3 A 
VOL. CXCIX-.—A. 
