TRANSPIRATION STUDIES. 13 
0.5 mg. In the final results no important conclusions are drawn from 
numbers obtained from differences in weighings smaller than 10 mg. 
The curves are all drawn to a scale which shows only readings which are 
surely significant and well out of the range of experimental error. The solid 
horizontal lines show the actual time and duration of the intervals measured 
and the broken lines connect the mid-points of these. This scheme is used 
because the irregularity of the time required for taking the dew-points made 
_.it impossible to take readings at stated intervals. The time was always 
read to the nearest minute and the hourly rate calculated. It will be noted 
that the experiments made in 1912 show a greater agreement in the time 
of readings, an advantage gained by the use of the carbon dioxide for cooling 
the dew-point apparatus. 
The extremely small size of the leaves prevented the use of any of the usual 
methods for determining their area, and so a direct method was used as 
follows. From 10 to 20 leaflets of approximately the same size were taken 
at a time, placed carefully with their long diameters in a straight line, and 
their aggregate length read to0.1mm. These leaflets were then turned so 
that their short diameters could be measured in the same way. All of the 
leaflets used in an experiment were thus measured and the average long 
and short diameters were found. The leaflets were then treated as ellipses, 
which they are so far as the eye can tell. The area of an average leaflet 
was calculated and was then multiplied by the total number of leaflets. The 
rachises were placed end to end and their total length was read. Their 
average diameter was found from many readings with a vernier caliper. 
The rachises were treated as cylinders, the groove which exists along the 
top beingignored. The stems were treated in the same way as the rachises, 
readings of diameters in this case being taken at regular intervals of 1 cm., 
as closely as the presence of thorns and branches would permit. Live thorns 
were treated as branches and all dead ends of branches were left out of 
account. Small errors doubtless occur in the operation of this method, 
but the consistency of the results goes to show that the errors may be 
disregarded and that the areas determined are comparable. 
For the sake of comparison, plants raised in the green-house from seeds 
and then placed in the open for six weeks were sealed in their pots and their 
water loss was determined by weighings. The evaporation rates in these 
cases were determined by the Livingston type of atmometer.* Atmometers 
which had been standardized under Dr. Livingston’s direction were used 
and the proper corrections for reduction were made by using the coefficients 
furnished by him. The rates were reduced to unit area of water-surface 
in order that the relative rates might be compared with those obtained from 
trees by the bell-jar method. This standardization was done in a dark 
constant-temperature room, an open petri dish, with a reservoir to keep the 
water level constant, being employed. Two of the cups were run with an 
open dish in this manner and the other cups were reduced to the open- 
water equivalents by the use of the coefficients referred to above. 
*Livingston, B. E., Carn. Inst. Wash. Pub. 50. 
