60 DISTRIBUTION AND MOVEMENTS OF DESERT PLANTS. 
As the tap-root, or the roots clustered about it and of the same char- 
acter, was no longer adequate for support, the lateral roots increased 
very greatly in diameter and assumed a part of the burden. Should 
the laterals die and decay before the destruction of the subaerial portion 
of the plant, as very frequently happens, this support is removed and 
the plant is easily thrown to the ground. From the circumstances given 
above it is clearly incorrect to make a sharp distinction at any period 
in the life of the cactus between water-absorbing and anchoring roots, 
since at no time does such distinction exist. So far as safe anchorage 
for the plant is to be counted a factor in limiting the distribution of Cereus 
guganteus it is apparent, therefore, that such can be attained wherever 
the permeable substratum is always firm and as deep as I meter or even 
less. This may be the case in certain places on the mesa-like slopes, or 
preferably on rocky places where crevices in the rock or spaces between 
bowlders afford depth of soil and additional support. 
The root-system of Cereus giganteus, in its relation to the substratum 
and to the roots of neighboring plants, may be best presented by giving 
a concrete example. The habitat of the specimen to be described is the 
eastern foot of the Tucson slope, running eastward from the base of the 
Tucson mountains; it is 0.5 mile west of the Laboratory. The ground is 
of a dark color and rather thickly strewn with dark volcanic bowlders 
of various sizes. A section of the ground where the cactus was growing 
shows the following characters: The uppermost soil, 30 cm. in thickness, 
is of malpais, which is derived from volcanic rock (basalt) ; in this stratum 
are bowlders of various sizes. Beneath the malpais stratum is a thinner 
one, 20 cm. in thickness, of caliche, in which also bowlders are embedded. 
Beneath the caliche is the solid rock. 
At this place the lay of the surface and its relation to the surrounding 
physiographic formations are such that no water comes to it from seep- 
age or drainage; the plants have to depend for their water-supply entirely 
on what rain falls on the particular spot. In the immediate vicinity 
of the cactus are to be found Parkinsonia microphylla, Larrea tridentata, 
Kramerva canescens, and near by Encelia farinosa and several species of 
cacti, notably Cereus fendleri and Mamillaria sp. The specimen of Cereus 
giganteus studied was a perfectly healthy and very vigorous plant 1.2 m. 
high and 35 cm. in diameter at the largest part. 
As already stated, the root-system of this species is of a double nature; 
it is in part deeply placed, which insures safe anchorage for the plant, and 
it is in part superficial, which provides a large absorbing surface and at 
the same time includes a relatively large area from which to extract the 
needful water-supply. The extent of the latter kind of roots is far in 
excess of that of the former. In the individual studied the deeply placed 
system consists of a single tap-root which goes straight down to a depth 
of 30 cm. when it gives off one branch that runs at right angles for a 
