92 DISTRIBUTION AND MOVEMENTS OF DESERT PLANTS. 
percentage (ordinates) and are taken directly from table 5. The pre- 
cipitation has also been plotted, in the form of a gradatory graph, the 
vertical lines denoting the amounts of precipitation occurring on the 
corresponding dates. The figures above the graph show the amount of 
precipitation recorded. Traces are denoted by a hachure above the 
graph, accompanied by the letter 7. The horizontal lines serve merely 
to guide the eye. Hachures and figures below the graph denote the 
dates on which soil-moisture determinations were made. This notation 
facilitates comparison of the precipitation graph with the curve of soil- 
moisture. To further facilitate this comparison, the precipitation graph is 
repeated with each curve of soil-moisture, it being assumed to be identical 
for the several stations. Vertical lines through the curves denote the 
position of every tenth day, beginning with October 1. 
In attempting to interpret these curves of soil-moisture it must be 
remembered that the samples for the different dates are not from exactly 
the same point in the soil, and that where local variations in the mois- 
ture-content occur these may lead to unexplainable irregularities in the 
curves. This consideration is especially important in the case of the 
hill soil, where the irregularities of the underlying rock-surface, as well 
as the presence of large rock-masses in the soil, cause marked variations 
in soil-moisture at any given time. Masses of rock which lie partly in 
the air and partly in the soil are effective in the control of the soil- 
moisture about them in two ways: they facilitate the downward pene- 
tration of rain-water along their surfaces, and they greatly hinder the 
loss of water which has once penetrated to soil which lies beneath them. 
This disturbing effect of local variations in soil-moisture is of less impor- 
tance when the question at issue is the average moisture-content of the 
soil for a period of several observations than when the fluctuations in 
soil-moisture are to be considered. But at best, the method here used 
is to be regarded as giving results which are only approximate; its prob- 
able error must be relatively very large, although we have at present no 
means of determining its magnitude. 
From the curves it is apparent that the moisture-content at the lesser 
depth varies from time to time much more markedly than does that at 
the greater depth. This of course is to be expected, for the surface soils 
are first to be affected by precipitation as well as by evaporation. The 
deeper samplings, however, are more valuable than the others in relat- 
ing soil-moisture to plant activity, for most plants of this region gain 
access to a depth of 30 cm. in a very short time after germination, and 
the soil-moisture at this depth is probably a fair measure of the water 
available for the general activity of most forms. The cacti are, in general, 
a marked exception to this statement, as are, indeed, many other forms 
provided with storage organs. 
All of the curves exhibit also the very important lagging of the effect 
of precipitation behind the march of the precipitation itself. Thus, it 
