REPORT OF THE SECRETARY. 87 
single point of time, so that changes of the atmospheric transparency 
during the period of observation are avoided. 
This new method will hereafter be employed by the Smithsonian 
observers at Calama in combination with the old, not only for air 
mass 2, but for air mass 3, and they will check one against the other 
frequently for a considerable period of time until we are abundantly 
satisfied of the accuracy of the new method of observation. Hith- 
erto the new method has enabled us to save at Calama a number of 
days of observation which, owing to the obvious changes in trans- 
parency of the atmosphere, due to formation or disappearance of 
clouds, would otherwise have been lost. 
So far as we have as yet been able to compare the results by the 
old and the new methods, they are on the average closely identical. 
For instance, on July 1 three values of the solar constant were com- 
puted: (1) By the old process; (2) from observation at air mass 2; 
(3) from observations at air mass 8. The results obtained were as 
follows: 1.948, 1.940, 1.955, all agreeing within less than 1 per cent, 
and the mean of the results by the new process agreeing identically 
with the result by the old. 
The new process requires but two or three hours of work, where 
the old required about 15, so that if it continues to appear as satis- 
factory as now a very great gain in labor will result from it. Not 
only is this so, but a still greater gain we think will come in accu- 
racy, for we have now eliminated the fruitful source of error, de- 
pending on the variability of the atmospheric transparency during 
the observations. 
The new method of determining the solar constant of radiation 1s 
not applicable to other stations than Calama without a new series 
of contemporaneous solar constant determinations by the old method 
and pyranometer observations at air mass 2 and air mass 3 to use 
with them. We have not at present available the necessary pyra- 
nometer observations at Mount Wilson, but we shall undertake to 
obtain them at the earliest practicable moment, and hereafter it is 
probable that the new method of determination will be employed 
there as well as in South America. 
On the whole, the expedition to South America was unexpectedly 
fruitful. First, satisfactory observations were made of the eclipse, 
including both photographic observations of the eclipse phenomenon 
and pyranometer observations of the brightness of the sky during 
its progress. Second, a very interesting conference was held with 
the chief and chief forecaster of the Argentine meteorological serv- 
ice, in which they explained their investigations of the application 
of solar radiation measurements to weather forecasts and indicated 
their high sense of the value of solar radiation work for weather 
