no 
Journal of Agricultural Research 
Vol. XXIV, No. 2 
PROCEDURE 
The trees were taken from the nursery and potted about April 15, 1917. 
From that time until April 26 they were kept in a warm room, without 
sunlight, to encourage root growth and establishment in the soil. On 
April 26 the cans were first brought to standard moisture content, and 
measurements of transpiration losses were begun. From that date to 
June 3 they were kept in the window of a warm room, where they received 
light for only a few hours each day. The pots were frequently but not 
regularly shifted in position. 
On June 3 they were placed on a revolving table in the greenhouse, 
where they remained until the close of the test on November 14, with 
the exception of one day out of doors. This table was handled in several 
ways, the power available for rotating it at the outset being inadequate. 
At first a small motor was used, the motor being cut in each minute for a 
period of a second or more, so as to give the table a fraction of a revolu¬ 
tion. For several short periods the table was turned by hand, a quarter 
revolution about once each hour. A water motor was finally used, which 
for a time kept the table revolving continuously. This, however, seemed 
to have a theoretically objectionable feature in that the trees were 
constantly passing from light to shadow, in a very unnatural manner. 
The driving belt was therefore arranged so as to move the table a perim- 
etral distance of about 6 inches each minute, or, say, a complete 
revolution in about 25 minutes. 
The pots occupied Ae periphery of the 4-foot table, various types of 
evaporimeters being placed between them. Within this circle was 
placed an air-and-soil thermograph, the arm of the air register being 
shaded, while the soil bulb was blackened and so placed, with its long 
axis horizontal, as to receive as much sunlight as the trees. With the 
assistance of maximum and minimum registering air thermometers and 
a thermometer attached to the blackened bulb of the soil thermograph, 
there were thus recorded both air temperatures and '*sun’' temperatures. 
In addition, a psychrometer was used during the morning observation 
each day, giving a rough indication of prevailing vapor pressures. 
The most important question of procedure, of course, concerns the 
method of determining water losses. As shown by Table I, each pot had, 
at the outset, a known gross weight when its soil contained 20 per cent of 
moistinre. The aim was to keep the moisture to this standard by replac¬ 
ing losses each day. It was o^y necessary to determine the amount of 
water required to bring the pot up to standard weight in order to record 
the loss for the preceding period. This was accomplished by placing 
the pot on one pan of the scale, the standard weights on the other pan, 
and filling from a titrating burette until a balance was reached. The 
amount drawn from the biuette was, therefore, the measure of the loss. 
The measurements of transpiration, it is thus seen, were actually volu¬ 
metric, even though scales were used. This introduced no error worth 
considering, as the temperatures at the observation hours varied scarcely 
at all from about 50° F. 
The burette was graduated to o.i cc. The scales were barely sensitive 
to O.I gm. under the usual load of 4,000 gm. However, errors from this 
source should be compensating, "yi^enever the filling was carried too 
far, as not infrequently happened, the overload was determined and 
allowed for, and also carried to the record for the next period. 
