LOCAL DISTRIBUTION OF SPECIES. 37 
of small ridges and troughs, a topography usually typical of Larrea. 
But such instances serve to direct attention to the fact that underlying 
the differences of topographical features there must be one or more con- 
trolling physical factors to which the widely different distribution of 
these plants is primarily due, and the first of these is apparently the 
amount and location of soil-water. 
A large number of observations, too numerous to be given in detail, 
go to indicate that the presence or absence of caliche has, locally at least, 
an influence on the distribution of these two species. All about the 
Laboratory domain the most exclusively Larrea areas are those where 
caliche comes most nearly to the surface in its characteristic layer, while 
the mesquite is, as a rule, absent from such areas. Nevertheless a com- 
parison of the map of Prosopis with that of Larrea (plates 14 and 18) 
shows that over considerable areas where neither is abundant the two 
may grow together or in close proximity. All this is consistent, however, 
with the principle stated above, since differences in amount of available 
soil-water are certain to result from the presence of the caliche, as it occurs 
at the base of Tumamoc Hill, and from the disposition of rocks and soil 
elsewhere. 
COMPARATIVE STUDIES. 
It is obvious from the foregoing that soil conditions, even within the 
narrow limits of the Laboratory domain, exercise a marked influence on 
the local distribution of plants, and this becomes still more evident when 
the plants of a wider area, with more pronounced differences of soils, are 
considered. 
For the sake of wider comparison in this direction, a study of the vege- 
tation of a number of areas in Arizona and New Mexico has been under- 
taken. That part of the Gila Valley lying between Solomonsville and 
Fort Thomas was visited by the writer in November, 1906. The work 
there was carried on with particular reference to ascertaining how far 
the plant habitats of the valley and their most characteristic plant species 
show a definite correspondence in their distribution with that of the 
various soils as they have been mapped by Lapham and Neill (1904), of 
the Bureau of Soils, U. S. Department of Agriculture. As shown by 
the map referred to (plate 19), a section of the Gila Valley at some place, 
Pima for example, where the different soils are well represented, crosses 
successively between the river bed and the lower mountain slopes on 
either side: (1) The Pecos sand, uniformly present on either side of the 
river in the valley trough; (2) the Gila fine sandy loam; (3) the Maricopa 
silt loam, a soil of very fine texture, with some of the peculiarities of 
adobe; these three constituting the alluvial soils of the district in dis- 
tinction from the three following, sometimes distinguished as colluvial 
soils, which are derived from the products of erosion of the mountain sides, 
