From these two general observations (exceptions to the usual conditions), that on the flood-plain of the Oued M’Zab and the cemeteries at Ghardaia, it is concluded that the grazing of animals has had a very marked influence in modifying the flora of southern Algeria. So far as could be told from the limited opportunity to observe, the modification has gone along on at least two lines, which are, of course, closely related. The size of the plants eaten is much under normal for the particular locality, and at the same time the capacity of the plant for reproduction has been greatly lessened. That such species as are not touched by animals have not spread more rapidly, or even have not become the dominant forms, is an interesting problem, and one that would have to be worked out for each species. It can be suggested, however, that the restricted distribution probably lies in the fact that the struggle of desert plants is mainly with an adverse physical environment rather than with one another, and that such conditions would not be affected by grazing animals. For example, large shrubs do not occur on the hamada, hence _Tamarix_, _Zizyphus_, and _Rhus_, as well as _Peganum harmala_, are limited to situations where the soil is fairly deep and the water relatively favorable. _Euphorbia guyoniana_ also occurs only in sandy soil, and the number of plants limited in their distribution to soil characters is necessarily a large one. COMPARISON OF SOME GENERAL FEATURES OF THE VEGETATION OF SOUTHERN ALGERIA AND OF SOUTHERN ARIZONA. When we compare the most striking characteristics of the vegetation of the Algerian Sahara with that of the Tucson region where the Desert Laboratory is situated, we find some interesting differences, which may be summed up in the terms “desert” and “semi-desert,” as applied to the two widely separated regions. What is meant by these terms will be apparent from the following short characterization: