366 Knight, — Further Observations on the Transpiration i 
ment it was found possible to maintain the temperature constant within 
reasonable limits, but, as the humidity was variable, determinations were 
made of the evaporating power of the air by means of a form of paper 
atmometer which has been found satisfactory for the evaporation rates 
normally encountered in south-east England. It consists (see Fig. 1) of 
a circular piece of filter-paper (A), to the centre of which is attached, with 
sealing-wax or shellac, a small glass cup in the form of a thistle funnel (c), 
such as is used in porometer work. The filter-paper is supported at its 
edge by a wire frame (b) which is kept in position by wire stays fixed to the 
glass cup. t The cup and its stem are filled with water and the stem is 
passed through the rubber stopper of a bottle which acts as a water reser- 
voir and at the same time supports the cup and the paper. As water 
evaporates from the paper, which must be completely wetted at first, it is 
replaced from the reservoir, and it is found that no drying out occurs at the 
edges of the paper if the diameter of 
the cup is sufficiently large — about 
1 1 cm. for a 9 cm. filter-paper. The 
paper must be of a rather loose 
texture or the conduction of water 
through it is too slow, although the use 
of the soft paper shortens the life of 
the instrument. Evaporation rates are 
determined by weighings. The ap- 
paratus is distinctly fragile, but easy 
to construct, and with careful handling 
one will serve for several experiments. It has the advantage that it is 
very light, and naturally responds to temperature changes more rapidly 
than porous cup atmometers. It has proved very convenient for green- 
house work, but out of doors it would probably suffer in heavy rain 
or high wind. In the present experiments, owing to the short time 
available for observations, this paper atmometer was adapted for rapid 
volumetric readings in order to save the time which would be occupied by 
weighing. The leg of the glass cup was sealed into the side of a horizontal 
graduated capillary tube (e), with a reservoir (d) and three-way cock (f) 
attached, so that the apparatus resembled a Ganong potometer, the cup and 
paper replacing the plant. With this apparatus, readings of the evaporation 
rate could be made over successive periods, each of two or three minutes’ 
duration. Determinations of transpiration rate of the plant and of absorp- 
tion from the potometer were made eight or ten times per hour. As 
a weighing ordinarily takes between one and two minutes, the following 
procedure was adopted in order to save time. The balance was placed on 
the roof of the air flue, and the potometer was attached to the balance arm 
by a thread passing through holes in the top of the flue and the base of the 
A 
Fig. 1. Type of paper atmometer used 
in transpiration experiments. A. Filter-paper. 
B. Supporting wire frame. c. Glass cup. 
D. Reservoir, e. Graduated capillary tube. 
e. Three-way tap. 
