474 COBLENTZ: NEW DESIGNS OF EADIOMETERS 



The efficiency of such a device was tested in the following 

 manner. Two strips of very thin platinum, such as is used in 

 bolometers (thickness about 0.001 mm.), about 6 by 20 mm. in 

 area were mounted over slits cut in cardboard which was 0.45 

 mm. thick. Both sides of these strips were painted with a thin 

 coat of lamp black and covered with soot from a sperm candle. 

 The thermopile receiver was 1.8 by 16 mm. Slits of bright 

 aluminum, 0.85 mm. thick, were placed in front of the thermo- 

 pile or in front of the blackened platinum strips when they were 

 in front of the thermopile. The distance between the thermo- 

 pile receiver and the platinum strip (and between the two plati- 

 num strips) was 0.45 mm. When the thermopile was exposed 

 directly to a standard of radiation the deflection was 12.15 cm.; 

 when one platinum strip intervened the deflection was 5.88 cm. ; 

 and when the two platinum strips were in front of the thermopile 

 the deflection was 3.57 cm. 



The multiple bolometer receiver. Since there is but little differ- 

 ence between the radiation sensitivity of a bismuth-silver ther- 

 mopile and a bolometer, the above tests show that the radiation 

 sensitivity of a bolometer can be increased by 50 per cent by 

 having the receiver (the arm of the bridge) consist of two strips, 

 one back of the other, the front strip being exposed to radiation. 

 By using three strips placed one back of another the galvanom- 

 eter deflection would be increased by about 80 per cent, and 

 by using four strips (joined in series, or a single strip folded three 

 times) the deflection will be double that produced by the front 

 strip. The sensitivity of the whole combination would be further 

 increased by placing this multiple receiver at the focus of a hemi- 

 spherical mirror. 



The bolo-pile. This is a combination of a single bolometer 

 strip, close back of which is placed the receiver of a thermopile. 

 The latter is so constructed that the pairs of receivers are in two 

 rows corresponding to the two bolometer strips. In this manner 

 the heating produced by the current passing through the bolome- 

 ter strips will produce no deflection in the thermopile circuit. 

 The manner of connecting the bolometer and the thermopile 



