22 
indicate that a limiting root system may have an important bearing 
on the water-stress behavior of the largest plants. 
Most of the taproots of both large and small plants extended to a 
depth greater than 7 feet, and the only one followed to its greatest 
depth measured 10 feet 8 inches in length. A considerable number 
of secondary roots extended to great depths and apparently shared 
the function of the taproot. 
The greatest quantity of water was removed by the plants in plat 1, 
where the moisture content of the soil had been kept highest through- 
out the season, the amount removed during May and June being 
about twice as much as in plat 3. Plat 1 produced dry matter at the 
rate of 1,413 pounds per acre more than plat 3. 
- It appears that in Arizona the severe water stress frequently ex- 
hibited by plants possessing large areas of leaf surfaces and the shed- 
ding of squares and bolls from such plants are not always induced 
by a lack of " available " moisture throughout the soil mass, but may 
be due to a reduction of moisture in the portion of the soil im- 
mediately surrounding the roots more rapidly than it can be restored 
by capillarity. 
There appears to be an important relation between the increase of 
the shedding rate (based on the number of bolls available for shed- 
ding) during the first two months of the flowering season and the 
gradual reduction of the moisture content in the lower soil layers. 
The mean interval between the opening of the flower and the com- 
pletion of the shedding process was approximately 10 days. The 
general tendency was for the period to increase as the season ad- 
vanced, but fluctuation was great even near the end of the season. 
A heavy rain materially shortened the interval between flowering 
and shedding. 
In late summer a depression in the rate of growth and flowering 
was shown in the large plants much earlier than in the smaller plants. 
After this retardation began the large plants never again exhibited 
as prompt or vigorous a response to irrigation as the smaller plants. 
Plants to which irrigation water was supplied soon after the first 
flower buds began to form, and frequently thereafter, produced a 
greater stem growth and a greater number of flowers during the first 
45 'days of flowering than plants to which water was not supplied 
until 16 days later. 
Plants growing in soil which was supplied with water sparingly 
throughout the season produced a greater number of bolls late in the 
season than plants provided with larger supplies of soil moisture. 
The highest " potential seed-cotton production " per plant by Octo- 
ber 1 was on plat 3, which had been supplied with water sparingly. 
The highest actual yield of seed cotton was produced on plat 1, on 
which the plants had been stimulated by comparatively early irriga- 
