[PATTERSON] CALLENDAR SUNSHINE RECORDER 55 
mounted in a wooden box of large cross section but of the same 
length as the tube and protected from radiation from the sides by a 
series of dead black diaphrams: the value of a scale division was then 
found to be -0417 gramme calories per min. or practically the same as 
before. The results show that the Callendar receiver when unpro- 
tected by glass is about 6% more sensitive than when screened by 
the glass globe. Kimball (loc. cit) found in one experiment that 
the unprotected receiver was about 10% more sensitive than when 
protected by the glass envelope. Even applying this correction the 
readings given by the Callendar sunshine recorder are still much below 
those given by the Angstrom Pyrheliometer except when the radiation 
is weak. Miller' has shown that when the sunshine receiver is exposed 
in its normal position (disk horizontal) that the caustic formed in the 
bulb by reflection affects the readings by making them too high when 
the reflected part falls on the black surface and too low when on the 
' bright. As the receiving disk was perpendicular to the sun’s rays, 
in the observations described in this paper, this source of error was 
eliminated. 
In the Callendar sunshine recorder, the radiation is measured 
by the difference in temperature’ between black and bright surfaces; 
the black surface consists of platinum wires covered with a black 
enamel and the bright surface the bright platinum wires, but as these 
wires are in contact with the mica disk on which they are wound the 
bright surface is in reality that of the mica. The various comparisons 
that have been made of the Callendar sunshine recorder show the 
desirability of making an attempt to get its readings more nearly 
absolute by laboratory investigation. 
1U. S. Monthly Weather Review, June, 1915, p. 264. 
