14 TRANSPIRATION IN A DESERT PERENNIAL. 
Unless the plants have some special physiological response to wind action 
the relative transpiration rates of potted plants ought to be comparable to 
the relative rates of the plants taken in situ in still air. A general observa- 
tional record of wind velocity was kept and it was found that whileamarked 
increase in wind velocity was always accompanied by an increase in the 
actual transpiration rate, this increase did not appear in the relative trans- 
piration, thus showing that the plant was affected in the same manner as 
the atmometer. The few conclusions drawn in this paper from a compari- 
son of the relative rates of plants in situ with potted plants are of course 
open to the objection that there may exist an unknown physiological re- 
sponse on the part of the plant to wind which did not appear. 
EXPERIMENTATION. 
All of the plants or branches used in the following experiments were in 
normal condition and were situated in exposed positions, so that the sun 
could fall upon them the entire day. Whenever over-night readings were 
taken the jar was completely covered with a black cloth. It will be seen 
that, as the experiments progress in chronological order, the agreement of 
meteorological conditions of the jars with each other and with the outside 
become better, due to the growing experience of the operator. A small 
amount of calcium chloride was exposed for the night readings, the amount 
being about tripled for the first reading in sunshine and then increased grad- 
ually as the day advanced and decreased in proportion as the afternoon 
progressed. In all cases it was necessary to spend one or two days in trials 
before the amount could be properly adjusted for a particular plant. Also, 
it frequently became necessary to take into account the changes of dew- 
point in the outside air. With the temperature kept within a degree or two 
of that of the outside air by means of the method of shading mentioned | 
above, and with the dew-point held near that of the outside air, the humidity 
varied only a small percentage from the atmospheric humidity. In no case 
were the differences between bell-jar and air conditions greater than the 
differences existing between the atmospheric conditions of any two days. 
Thebestresults were obtained when periods of one to two hours were used; 
in the long over-night periods results were never satisfactory. Of course 
two-hour periods might have been used at night as easily as in the day, if 
the purposes of the experiment had required it. It was thought advisable 
to eliminate so variable a factor as the wind and so no attempt was made to 
get readings in the wind except in the case of potted plants. 
In the tables which follow, unless otherwise defined, 7 represents the 
transpiration rate per hour per square centimeter of area; EF the evapora- 
tion per hour per square centimeter; and 7'/E the relative rate. All losses 
are given in fractions of agram. Unless stated otherwise all readings were 
taken in bright sunshine. In table 1, the original readings and some of the 
details of calculation are given, while for economy of space these details are 
omitted in the succeeding tables. 
EXPERIMENT I. 
The subject of this experiment was a small 1910 leafless seedling, growing 
on the north slope of Tumamoc Hill. The area was found to be 7.99 sq. em. 
