1910] LIVINGSTON—SOIL MOISTURE 255 
evaporation on the hill, such as the presence of numerous rock 
fragments on or in the soil. 
An interesting question is suggested by the fact that the creosote 
bush dominates the slope and occurs on the hill, but is almost 
entirely absent from the plain, excepting along the margin where 
it abuts against the slope. No evidence as to why this shrub should 
not thrive on the plain is as yet at hand. The most vigorous 
specimens of the area occur along the lower margins of the slope, 
which agrees with a similar behavior always observed in irrigated 
specimens, in seeming to indicate that a more adequate water supply 
produces an abnormally great growth. SPALDING mentions the 
physiographical relation (which of course cannot directly affect the 
vegetation) that Larrea usually dominates gentle slopes which are 
being eroded at the present time. It would seem that there must 
be some condition in the soil of the plain which is antagonistic 
to the growth of this shrub. A study of the oxygen conditions of 
these soils may possibly throw some light on the subject. The 
moisture condition alone is inadequate to explain the facts. 
The results obtained from a study of the graphs would place the 
hill soil first, as the best suited for géneral plant activities; the soil 
of the plain would occupy second place, and the two other soils 
would lie rather close together and far below either of these. This 
general arrangement agrees well with the vegetational characters 
of the four soil types if we consider that the deep-rooting forms of 
wash and plain obtain moisture from levels deeper than those here 
considered. It is also in fair agreement with the simple series of 
the water-retaining powers of the four soils. From this retaining 
power alone the soils would have been interpreted in the same 
general way. This emphasizes the water-holding power as an 
easily obtained and very important factor in studies of distribu- 
tion. 
The last-named factor has been used by the writer (4) with 
Some success in a search for the determining conditions of forest 
distribution in Michigan. It appears that for different upland 
habitats of the same region, where the level of underground water is 
| far from the surface, and the precipitation and evaporating power 
