10 BOTANICAL GAZETTE [yuLy 
equilibrium. The tests were confined finally to the region of soil 
moisture from air-dry to the wilting coefficient, for with a higher 
moisture content the seeds always became saturated with water. 
In the case of this sand it was not until the water content was 
reduced to about 1 per cent that a noticeable “back pull’’ was 
developed by the soil. 
This method is obviously open to the criticism that friction 
retards the movement of water in dry soils, and that the seeds 
therefore do not reach actual equilibrium with the total soil mass, 
but only with the soil lying very near them. In order to meet this 
difficulty a rotation method was adopted which brings the seeds 
constantly into contact with fresh soil particles. 
A definite amount of dry soil, usually 60 gm., was taken, and 
the desired amount of water thoroughly mixed with it. In this 
condition the soil was divided finely enough to pass through a | 
2 mm. sieve. The moist soil was then placed with dry seeds in 
a wide-mouthed 200 cc. bottle, without completely filling it, so 
that rotation would constantly mix the soils and seeds. The bot- 
tles were carefully sealed with heavily shellacked corks to prevent 
loss of water during the period of rotation. 
These bottles were arranged on rotating wheels, driven by a 
motor and controlled by a speed reducer (fig. 1). The range from 
air-dry to wilting coefficient was divided into 10 fairly equal 
divisions, giving 11 tests in each series. The rotation was con- 
tinued for 15 days, this time having been chosen after making, 
tests as to the effect of differences in duration of the experiment 
on the amount of intake by the seeds. For instance, no. 2/o sand 
with about 0.2 per cent moisture added permitted an intake of 
22 per cent of their weight by the seeds in 5 days, and a parallel 
test showed 21.7 per cent intake in 10 days. Fifteen days, there- 
fore, seems ample time for the establishment of equilibrium. At 
the end of the time the bottles were opened, and the seeds very 
rapidly separated from the soil and brushed free of all dust with a 
camel’s-hair brush. The soil and seeds were both placed in weigh- 
ing bottles as quickly as possible, to prevent serious loss of water 
by evaporation. The soil was weighed carefully and dried at 
104° C. until loss ceased. The seeds also were weighed. The 
