42 Prof. S. P. Langley on the 



own * enable us to convert this arc into wave-length. This 

 fixes the position of bands in the spectrum. The amount of 

 heat in any portion of the spectrum is ; within the narrow 

 limits of errors of observation, strictly proportional to the 

 deflexion on the galvanometer-scale (the conditions of the 

 bolometer, battery-current, galvanometer, &c, remaining 

 constant). As these degrees are arbitrary, they are converted 

 into thermometric degrees by a process fully detailed in the 

 original memoir. 



The preliminary record of the humidity, state of the sky, 

 temperature, &c, is nearly self-explanatory. We need only 

 explain that " Eock-salt lenses at 37 centim. " refers to the 

 fact that the focal length of these lenses increases from 35 

 centim. in the visible spectrum to one indefinitely greater 

 with heat of great wave-length, and that this focus accord- 

 ingly needs to be adjusted for the particular part of the heat- 

 spectrum under study. 



" Deflexion per degree Centigrade " refers to the use of a 

 constant determined for each evening, giving the actual 

 deflexion the galvanometer produces, for each degree of 

 excess of temperature in a certain standard Leslie cube at a 

 certain standard distance from the bolometer. 



The order of observation consists first in noting the time. 

 (We will suppose, in the example which follows, that the time 

 is 9 h '08 M. T., on the evening of February 9, 1887) ; next, 

 in noting the prismatic deviation corresponding to the actual 

 position of the bolometer in the spectrum, which in this 

 particular case is 41° 08' 30", or that of the D line in the 

 visible spectrum. Previously to observation the bolometer 

 has been radiating through the spectroscopic train and mirror 

 to the special screen described, which, on this particular 

 evening, is at the constant temperature of +18° Cent. 

 Under these circumstances, the needle will take up some 

 position representing radiation to the screen, which at this 

 first exposure we will call A, its position here on our 

 arbitrary scale being at the 215th millimetre. During this 

 time the siderostat-mirror has been so placed as to be sending 

 towards the bolometer radiations from the sky on the east of 

 the moon t. Call the effect of these particular sky radia- 

 tions B. They have been intercepted by the screen, but 

 now the screen is withdrawn, and the bolometer radiating to 

 this eastern adjacent sky receives less heat than the screen 



* See Phil. Mag. xxii. August 1886. 



t The area of sky observed is virtually the same as that of the moon. 



