192 HOWARD E. PULLING 



that it is protected against losses by convection, especially 

 convection losses due to extraneous air currents, that it is 

 sufficiently uniform over its entire extent or that the radiation 

 is incident always upon the same portion, and that it is placed 

 in the position most favorable for the reception of radiation 

 (i.e., in the plane at right angles to the direction of every ray 

 incident upon it), for any other position, by decreasing the 

 efficiency of the surface, may seriously affect the results. The 

 effects that all the devices, intended as protection to the in- 

 strument, have upon the incoming radiation must be sufficiently 

 well known to insure the requisite accuracy. The extent, 

 character and rate of change of the surface with time must 

 be definitely known not to exceed the limits of allowable varia- 

 tion or else the instrument must be re-standardized with suf- 

 ficient frequency to meet this requirement. If auxiliary mir- 

 rors are employed the difficulties increase because of refraction 

 in the mirror, intensified by the bending under intense illumi- 

 nation, defective reflecting surfaces, selective absorption and 

 selective reflection. Mirrors offer exceptional difficulties in 

 many cases, for example in measuring sky radiation, because 

 such radiation is polarized and is liable to partial extinction 

 upon reflection. 



The heat capacity of the entire instrument should be small 

 in order that the heat received may promptly produce as near, 

 the maximum temperature effect as is desired. Conduction 

 losses between the receiving surface and the indicating device 

 should be kept within desired limits. 



The simplest instruments of this type are modifications of the 

 Arago-Davy actinometer, 38 a pair of thermometers, one with 

 blackened and one with unblackened bulb, either exposed or 

 in a vacuum. There are many sources of serious error to be 

 investigated before they can be accepted as of any value al- 

 though the principle is correct. 



Instruments depending upon the utilization of the heat de- 

 veloped for the vaporization of liquid may prove to be of value 



38 Levy, M. A., L'Actinometer Arago-Davy. Deh6rain Ann. Agron. 4: 

 510-530. 1878. 



