S86 DISTRIBUTION AND MOVEMENTS OF DESERT PLANTS. 
As has been pointed out in Publication No. 50 of the Carnegie Insti- 
tution of Washington, the retaining power of the hill soil is high, being 
48.1 per cent of its dry weight.! The high moisture retaining power 
greatly retards percolation and also evaporation, as well as hindering to 
some extent the absorption of soil-water by plant-roots. The high evap- 
orating power of the air in the dry seasons at Tucson also tends to prevent 
water-loss from the deeper soil-layers. This is explained by the fact that, 
at the beginning of a dry period, evaporation removes moisture from the 
surface layers of the soil more rapidly than this can be supplied from 
below, the evaporation rate being much higher at the soil surface than 
the maximum rate of diffusion through the soil. Therefore, if a dry period 
follows after a rain, the surface of the soil quickly becomes desiccated and 
operates in the manner of a dry mulch to lessen the rate of water-loss 
from the deeper layers. The dry surface layer at length becomes thick 
enough practically to prevent further water-loss by evaporation. At 
the end of the spring dry season, the driest period of the year, the soil 
of the hill has been found to contain, at a depth of from 30 to 4o cm., a 
water-supply amply sufficient for the normal growthof the more xerophilous 
forms of plant life. The actual amount of soil-moisture contained at 
such depth, in July, 1904, was as much as 17.83 per cent of the dry weight 
of the soil; just preceding the advent of the summer rain in July, 1907, 
a sample of this soil from a depth of 40 cm. was found to contain 15.8 
per cent of its dry weight. At the same time, a sample from a depth 
of 15 cm. had a moisture content of only 9.1 per cent of its dry weight. 
At this season of the year the moisture-content of the deeper soils of the 
hill usually equals or surpasses the data just given. 
(2) The soil of the Larrea slope, which surrounds the hill, is even more 
shallow than that of the hill itself. It is underlaid almost continuously 
by caliche, which is usually found at a depth of 30 cm. or less. Above 
the caliche often occurs a layer, some 10 cm. in thickness, of coarse gravel 
and rock fragments mingled with comparatively little soil. The soil 
above this is more loamy than that of the hill, with an admixture, usually 
approaching half its volume, of finely fragmented lava and caliche. The 
soil here is thus seen to be exceedingly well drained; with its under- 
lying caliche, which is nearly impervious, it is unable to hold nearly so 
much water as does the adobe clay of the hill. Beneath the caliche 
there are occasional deposits of soil, but evidence has not been found that 
any considerable amount of precipitation water often penetrates to these. 
This soil has little or no tendency to crack when dried out from the 
wet condition, partly because of the lower moisture-holding power, and 
perhaps also in part because of the larger proportion of small stones. 
A sample of this soil showed its moisture-holding power to be 20.1 per 

' All samples used in deriving the soil-moisture data here to be presented, as well 
as for data of Publication No. 50, were unsifted, but no pebbles larger than 8 mm. in 
diameter were included. : 
