134 PROCEEDINGS OP THE AMERICAN ACADEMY 



thermograph. A telescope, with cross wires, is also attached to the 

 box, and may be used to direct it, but the former method is much the 

 more satisfactory. 



The entire apparatus is mounted upon a massive telescope tripod. 

 The mounting has tangent screw motions, and is provided with a ver- 

 tical circle for reading the altitudes. An hour or two before it is in- 

 tended to observe, the apparatus is mounted in the open air, in order 

 that it may have time to acquire the external temperature. Two 

 strands of heavy electric light wire pass from the thermogi-aph to the 

 laboratory, where they are attached to the galvanometer, which is 

 mounted upon a firm pier. 



For some reasons the effect of phase should precede all other con- 

 siderations as a subject of investigation, but it has been postponed 

 until later. The subjects that have engaged our attention, up to the 

 present time, may be for the most part classed under two heads : 

 1. The comparative intensity of radiation of the full moon and ©f the 

 sun. 2. The absorptive eifect of the earth's atmosphere upon the 

 radiation of the moon. These two topics will be considered in turn. 



It is obviously impossible to confine the observations to the time of 

 full moon without giving undue time to the investigation ; hence the 

 effect of phase demands to be considered. To avoid the introduction 

 of errors from this source the following precautions have been taken : 



1. To confine the observations to the three days nearest full moon. 



2. To measure always the radiation from a constant area of the moon's 

 surface. To fulfil the conditions of the second precaution, tliere is 

 placed, close to the strip of the thermograph, a diaphragm, the open- 

 ing in which is much smaller than the moon's image which falls upon 

 it. Thus there may be a large change of phase, and the opening still 

 remain full of the moon's image. This arrangement would completely 

 remove the difficulty, provided that the moon reflects like a flat disk, 

 as Zolner thinks, — which is not at all certain. It is believed, how- 

 ever, that, when taken in connection with the first precaution, the out- 

 standing error cannot be larsre. The method has the disadvantage of 

 reducing the available heat by about one half; but, as it is, the heat 

 is sufficient to drive the needle off the scale with the galvanometer in 

 its most sensitive condition. The controlling magnet is set so that a 

 defiection of between one and two hundred scale divisions is obtained, 

 corresponding to a period of between five and ten seconds' vibration 

 of the needle. 



