546 Wisconsin Academy of Sciences, Arts, and Letters. 
and the closing cord in the other. The receiver is so suspended 
that the long side of the frame is turned toward the sun and the 
side with the opening and closing cords is next to the boat. 
At the signal of the recorder, the operator pulls the cord which 
turns back the shutter so that the thermal couples are exposed to 
the sun. The full deflection is reached in a few seconds. The 
recorder notes it and then signals to close the cover. When the 
galvanometer has returned to zero, the reading is repeated. Two 
or more swings of the galvanometer are ordinarily recorded from 
each depth. Readings are usually made at half-meter or meter 
intervals. There is no difficulty, but little profit, in using smaller 
intervals. However, the hemispherical form of the glass cover 
of the thermal couples is not suitable for use in determining ab¬ 
sorption by thin strata of water that lie close to the surface. 
With this instrument readings can be made very rapidly and 
very accurately. It is not hard to estimate to tenths of a divi¬ 
sion of the scale. With the coils in use one division of the scale 
equals about 0.011 calories per square centimeter per minute. 
This is when the shunt is included in the circuit. When the shunt 
is cut out, the value of one division is 0.0059 cal. cm, 2 min. 
Ordinary errors in estimating tenths of a division make little 
difference in the results. 
A portion of a set of readings is added (table 3) in order to 
show the general results obtained from the pyrlimnometer. 
The pyrlimnometer was first used in 1912. In its original form, 
the glass cover of the thermal couples was flat. Two receivers 
were made, one to read in the water and the other to read in the 
air. The plan was to read the instrument in the air alternately 
with that in the water. This method, however, proved unneces¬ 
sarily complex. If the sun’s radiation is changing rapidly, as 
by haze or drifting clouds, the changes are too rapid to be fol¬ 
lowed by the instrument. If the sun is fairly steady, as on an 
ordinary sunny day, the second instrument adds little to accuracy 
and much to trouble. 
The regular receiver was built in 1913 and was then 
furnished with the hemispherical glass cover shown in the figure. 
This change made it unnecessary to compute the varying amounts 
of radiation reflected from the glass at different altitudes of the 
sun. 
The smaller receiver (text fig. 5) was kept in its original form 
and used in determining the absorption of the sun’s radiation by 
