260 BOTANICAL GAZETTE [APRIL 
to a critical comparison this and other methods of measurement 
of solar energy. These discussions need not be repeated here. In 
spite of the shortcomings of this method, it has the advantage of 
being capable of easy manipulation and gives sufficiently correct 
results for a comparison of relative light values under the cheese- 
cloth shade and in the open. 
The observations were taken on December 17, December 23, 
and January 25. December 12 and January 25 were bright days 
with only a few hazy clouds, which are usual in Cuba toward the 
middle of the day. December 23 was cloudy, so that all the light © 
on that day was diffuse. On the first day, December 12, ten obser- 
vations were taken usually at each hour period for each kind of light 
under each condition, but on account of the length of time required 
to make that number of observations and the change of light mean- 
while, only five observations were taken at each reading on te 
other days. As a rule, the observations were taken alternately 
within and without the tent. Since the personal equation in the 
judgment of color is likely to play an important part in determining 
the time required for the sensitive paper to reach a standard tint, 
the probable error of the average was calculated for each set of 
observations, except some taken early or late in the day, when, on 
account of the weakness of the light, it was not possible in some 
cases to take more than one or two observations. An examination 
of the probable errors given with the column of averages shows that 
it is not of undue magnitude. It may also be stated that the ten 
or five observations from which the average is made up showed a 
very close agreement, usually within a fraction of a second of each 
other except when the exposures were very long, that is, 3° seconds 
or more, although, as has been stated, the observations were taken 
alternately in the two stations in such a way that the observer wa> 
not influenced by the previous observation and record. From the 
observations thus made each hour in each station, the average time 
of exposure was calculated. These averages with their probable 
errors are given in table I. 
The light intensity is of course proportional to the reciprocal of 
the time of exposure. These reciprocals were obtained, therefore, 
but in order to reduce the figures of each day to relative values, the 
