LOCAL DISTRIBUTION OF SPECIES. 41 
The evidence is such that we may unhesitatingly name soil-water content, 
percentage of alkali, and texture of the soil as three of the more efficient 
edaphic factors controlling the choice of habitat exhibited by the plants 
of the Gila Valley. The occurrence of the cottonwoods on the river- 
banks, of the mesquite on the flood-plain, and the creosote-bush on the 
lower mountain slopes is certainly conditioned by the first, and the case 
of the arrow-weed, which grows in any soil whatever where there is water 
enough, is still more striking. The constant association of Sueda moquini 
and Atriplex nuttalli1, almost alone over miles of salt-spots, both drained 
and undrained, is sufficient evidence of the efficient action of the second 
factor, and the behavior of certain other species points as clearly to a 
choice based on physical peculiarities of the soil. Thus Aériplex poly- 
carpa in its close adherence to the Maricopa sand suggests neither water- 
supply nor percentage of alkali as controlling factors, but more probably 
a demand on its part for more perfect aeration than is afforded by the 
heavier soils beyond. 
(3) The absence from the mountain slopes of the upper Gila Valley 
of plants so conspicuously present in similar situations on the Laboratory 
domain as the sahuaro, Encelia farinosa, and Parkinsonia microphylla 
points to the fact that general climatic conditions are also potent in 
determining what plants shall and what shall not hold their places here 
in associations of which they are elsewhere important constituents. From 
evidence obtained on Tumamoc Hill (p. 47), it appears probable that 
temperature is the determining factor in the cases referred to. The 
consideration of this phase of the subject is reserved for another place. 
(4) But whatever differences of characteristic constituents may appear 
in the same associations, as they are represented in the valleys of the 
Santa Cruz and the Gila, the capital fact remains that in the local dis- 
tribution of their plants these areas are essentially alike. We may add, as 
subsequent studies of the writer have shown, the Salt River Valley, so 
that what has been said applies generally to all the great valley systems 
of southern and southwestern Arizona. Beyond these valleys are the 
mountains with their mesophytic vegetation. It is in the valleys and 
on the adjacent slopes that the characteristic desert vegetation attains 
its best development and exhibits the peculiarities of local distribution 
that have been described, and since one great valley is fundamentally 
the counterpart of every other, a thorough study of one becomes the 
means of interpretation of all the rest. 
(5) In the Gila Valley, as in that of the Santa Cruz, and in general 
as far as these observations have extended, everything indicates that causes 
now in operation have determined the actual distribution of plants in 
the associations and habitats where they are now found. The general 
physiographic features remain constant, but boundaries are continually 
shifting, and coincidently with these changes the plant associations retreat 
