126 BOTANICAL GAZETTE [AUGUST 
difference could be shown to exist in their transpiration rates per 
sq. cm. of leaf surface. The following table gives the details of 
this work, The column marked “check”? contains the data derived 
from the apparatus set up to measure the quantity of water in the 
normal air. 
An examination of Table V shows a constantly higher rate for no. 64 
over that of no. 67; the rate of difference, however, was very variable. 
Unfortunately a larger leaf area was taken each time for no. 67, so 
much so that the relative humidity in its cylinder was higher than that 
of no. 64 in the first two experiments. This fact would place no. 67 
at a disadvantage in these two experiments, since the higher relative 
humidity would retard to a greater degree its transpiration rate. Had 
it not been for the third experiment, therefore, in which, notwith- 
standing the greater leaf area of no. 67, the total transpiration of 
no. 64 was greater, thereby causing a higher relative humidity within 
its cylinder, the series would have been inconclusive. This third 
experiment, however, confirms the results of the other two, by showing 
that plant no. 64 may even overcome the disadvantage of a higher 
relative humidity and still transpire nearly twice as much per sq. C™. 
of leaf surface as the other. 
It may be well to add here several suggestive details for those who 
wish to use this method. On account of the strong affinity of P05 
for water, it must be kept very tightly sealed. Moreover, it 1s 4 
difficult substance to handle, especially to transfer to the mouth of a 
small U-tube. I have found it very convenient to keep it in such . 
flask as is made especially for holding anhydrous copper oxid. = 
has a small neck out of which the P,O, can readily be poured into 
the test tube, is fitted with a good ground-glass stopper within, a” 
a ground-glass cap without. I have found that phosphorus pentoxid 
keeps perfectly dry in this form of bottle; the stopper does not stick 
fast, and it is ready for use at any time. 
In filling the U-tubes it is well to push a bit of glass wool into the 
tube with a glass rod, pour on this about o. 5" of P,O,, then _ 
wool and more P,O,, until the tube is filled. A bunch of glass ¥0 
is put on the top, so that in handling the tubes the P,O, will not get 
into and clog the intake and outflow tubes while they are being 
handled, between the times of weighing and setting up the apparatus: 
