REPORT OF THE SECRETARY. 91 
PREPARATION OF APPARATUS, 
Sky radiation instruments—The director and Mr. Aldrich have 
devoted much time to the design and testing of apparatus for meas- 
uring the scattered radiation of the sky by day. What is desired is 
an instrument exposing horizontally an absorber of radiation in such 
a manner that the rays of the entire visible hemisphere of the sky 
would be received upon it, all rays not of solar origin would be 
excluded by a suitable screen, and the total energy of the scattered 
sky radiation originally emitted by the sun would be measured 
accurately. This is a more difficult problem than the measurement 
of the direct solar radiation, and it is unlikely that quite as high 
precision can be attained with the sky radiation instrument as with 
the pyrheliometers used for measuring direct solar radiation. From 
experiments with several instruments of the kind which have been 
constructed in the Observatory shop by Mr. Kramer and tested by 
Messrs. Abbot and Aldrich it now seems probable that the sources 
of error can be so far eliminated that sky radiation measurements 
accurate to about 2 per cent will be made. An instrument embody- 
ing what are thought to be the final improvements of design is now 
under construction, and it is hoped it will be used a great deal in 
the coming year. 
Balloon pyrheliometers.—Still more time has been devoted by 
Messrs. Abbot, Aldrich, and Kramer to the reconstruction and test- 
ing of balloon pyrheliometers. Mention was made in last year’s 
report of the proposed measurements of solar radiation by apparatus 
attached to sounding balloons and raised to great elevations. As 
stated below, the first trials in August, 1913, while unexpectedly 
successful in many ways, did not enable us to obtain measurements 
above the elevation of about 14,000 meters, or 45,000 feet. At this 
elevation the mercury froze in the thermometers. Also, the clock- 
work proved not sufficiently accurate for best results. Still the 
results obtained were so promising that it was thought well to repeat 
the experiments. 
Accordingly the five balloon pyrheliometers were reconstructed. 
Excellent French clocks were substituted for those used in 1913, and 
many improvements of the instruments were introduced. Two de- 
vices were employed to prevent the freezing of the mercury in the 
thermometer. In some instruments water jackets, having numerous 
interior copper bars to act as heat conductors, were arranged. In 
these it was hoped to make available the latent heat of freezing of 
the water and thus to prevent the surroundings of the pyrheliometric 
apparatus from descending far below the freezing point of water. 
In other instruments electrical temperature regulators were provided. 
Many experiments were tried to obtain a constant, powerful, and 
