852 PROCEEDINGS OF THE AMERICAN ACADEMY 



Description of Balance Case, Figijre 2. 



To protect the Bolometer effectually from air-currents and sudden 

 changes of temperature, and to enable it to be handled more easily, 

 it is enclosed in a chamber situated in the axis of a long cylin- 

 drical holder of non-conducting material (wood or ebonite), which can 

 itself be held by a suitable clamp pointed accurately in any direction 

 or laid in a horizontal position on Ts. This chamber is lined with 

 copper, to secure an equable distribution in the heat of its walls. 

 Through a circular opening, 15 mm. wide in front, the grating of strips 

 can be seen, and over all is slipped a hollow cylindrical cover of wood 

 or ebonite 15 cm. long, within which is a second tube containing four 

 or more concentric cardboard diaphragms pierced by apertures 6 mm. 

 in diameter. These disks of cardboard are separated by ebonite rings, 

 and form a succession of drum-like chambers, through whose apertures 

 the radiation passes unobstructed, but by which the entrance of air- 

 currents from without is effectually stopped. In front of all, a re\^olv- 

 ing cardboard diaphragm with suitable stops admits or shuts out the 

 radiation at pleasure. At the back of the copper-lined chamber is a 

 layer of solid non-conducting material, through which pass the connect- 

 ing wires, terminating in metal plugs insulated from the copper lining. 

 With these plugs the four terminals of the double grating are connected 

 by clamping-screws. The Balance Case is prolonged yet 15 cm. far- 

 ther back, forming a tube in which may be included and protected 

 from air-currents an adjustable resistance or rheostat by which the two 

 arms may be brought to perfect equality. It is advisable to have the 

 two halves as nearly as possible equal at the first ; since, if unequal, 

 the increment of resistance in the larger, caused by a general rise of 

 temperature, exceeds the corresponding increment in the smaller, ne- 

 cessitating a frequent readjustment of the variable resistance and pro- 

 ducing a " drift " in the galvanometer needle, which slowly changes its 

 direction according as the temperature of the room rises or falls. This 

 " drift " is not a peculiarity due to the use of the Bolometer, however, 

 as a similar drift, due to different causes, affects the galvanometer (if 

 equally delicate) when used with the thermopile. 



DEscRirxioN OF Connections, Figure 3. 



The Bolometer, A, is connected witli the distant coils, forming the 

 rest of the Wheatstone's Bridge, by four insulated copper wires, u, i\y. z 

 (twisted together and covered with flannel to reduce the effect of vary- 



