334 
ASTRONOMY: ABBOT AND ALDRICH 
trie with the strip is a hollow hemispherical screen of ultra-violet crown 
glass, 26 mm. in outer diameter and 2 mm. thick. Its purpose is to 
admit rays of shorter wave-lengths, such as form essentially the whole 
strength of the direct and scattered solar rays, but to cut off rays of 
great wave-length proper to the emission of a body at ordinary tem- 
peratures. During measurements of nocturnal radiation this glass 
screen is removed. A nickel-plated hemispherical shell of polished 
nickel-plated copper encloses this glass, and is removable at pleasure.. 
If now the shutter is opened diffuse sky-radiation falls upon the strip 
and warais it, producing a deflection of a moving-coil galvanometer in 
the circuit of the thermo-couple. The shutter being then closed, an 
equal deflection may be produced by the electric heating current. As 
corrected to allow for losses by reflection of the glass and the imper- 
fect absorption by lampblack, the energy dissipated in the strip by the 
heating current measures the energy of radiation. As constructed the 
sensitiveness of this instrument is so great that it proves convenient to 
balance the deflection to zero by means of a potentiometer current in the 
galvanometer circuit, and so to reduce the operations to the zero method. 
A defect of this simple form of pyranometer is found to be caused by 
the slow warming of the glass-covered portion of the copper disk when 
the shutter is opened, which at other times shades that area of the sur- 
face. This warming induces a secondary deflection, because it affects 
the two differently situated ends of the thermo-couple differently. Ex- 
periments have shown, however, that practically the full deflection due 
to direct heating of the strip occurs in 20 seconds, and that the second- 
ary deflection begins to be sensible after 20 seconds. Accordingly the 
error is eliminated by balancing the primary deflection by the potentio- 
meter current after exactly 20 seconds, then closing the shutter and 
waiting two minutes for the secondary heating to subside, before adjust- 
ing the heating current. 
Fearing that this defect might prove more serious in nocturnal radia- 
tion work, we devised a second form of pyranometer. In this form 
there are two blackened manganin strips side by side, each 2 mm. wide, 
6 mm. long, separated by a nickel-plated copper bar 2 mm. wide, and both 
insulated as in the simple form by vertical mica strips coming exactly to 
the surface of the plate. Thermo-couples connect the two strips at the 
back, the hot junction behind one strip, the cold junction behind the 
other. As the two strips absorb radiation equally, there would be an 
equal rise of temperature, if it were not that one strip is 10 times as thick 
as the other. Owing to this the conduction to the ends is so much greater 
for the thick strip that a difference of temperature arises, and a deflec- 
