Perennial grass production in the semidesert grass-shrub type (with and without a velvet mesquite (Prosopis juliflora var. velutina [Woot.] Sarg.) overstory) was dependent primarily on current summer rainfall and previous summer rainfall. The influence of previous summer rainfall was an interaction effect-not a direct effect. The best overall relationship involved current August rainfall, previous June through September rainfall, and the interaction product of these two. However, in interaction product alone yielded estimates essentially as good as the multiple regression, and explained 64%-91% of the year-to-year variability in grass production. Winter precipitation had no consistent effect on perennial grass production the following summer. The depressing effect of mesquite on perennial grass production was most noticeable at low rainfall levels, and became minor at high rainfall levels.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.. Allen Press and Society for Range Management are collaborating with JSTOR to digitize, preserve and extend access to Journal of Range Management.

Measurements of phenological development, herbage production, basal area, and density of annual and perennial grasses and of the half—shrub burroweed (Aplopappus tenuisectus (Greene) Blake), over a 4—year period show that production of each class of plant was affected to some extent by each of the others, except that annual grasses had no effect on burroweed crown area. Production of Arizona cottontop (Trichachne californica (Benth.) Chase), the dominant perennial grass, was restricted about 25% on plots with annual grass or burroweed competition and 46% by both together. Annual grass production averaged 18% lower with burroweed competition and 44—54% lower with perennial grass competition. Burroweed crown area increased 220% on plots with no perennial grass, but only 111% on plots with perennial grass competition. Presence of burrowed reduced perennial grass yield only moderately, because the root systems of burroweed and grass do not overlap greatly, and their main growth periods are at different seasons. High evaporation rates during the summer growing season masked most differences in moisture extraction between species. During the winter—spring growing period, on the other hand, burroweed depleted the available soil water rapidly, while water loss on perennial grass plots was little more than from bare soil.

DWIGHT R. CABLE Highlight: Mesquites used water consistently to a depth of 3 m and outward to 10 m beyond the crowns, but use at 15 m was limited mainly to drier periods when water supplies closer to the trees were depleted. With the start of spring growth, water was extracted most rapidly from the surface layers. As the season advanced, the watersupply zone became increasingly thicker. Rates of extraction were highest immediately after recharge in early spring and early summer, and lowest in late fall. Differences in available water in the soil accounted for 72 to 88% of the variation in rates of extraction. The competitive effect of velvet mesquite on perennial grasses is most severe in the upper 37.5 cm of soil under and near the mesquite crowns, and gradually decreases with distance into adjacent openings. The competitive effect in the openings is much more severe in dry years than in wet years. Velvet mesquite (Prosopis juliflora var. veZutin42) competes severely for soil water with perennial grasses. When mesquite is killed or reduced in amount, perennial grass production increases, sometimes dramatically (Parker and Martin 1952; Cable and Tschirley 1961; Martin and Cable 1962, 1974; Cable and Martin 1964). These results have been attributed to the extensive lateral root system of velvet mesquite, which occupies the same soil mass from which grasses must obtain their water. The assumption is that mesquite roots exert a stronger "pull" on the soil water than do the grasses. The present study was undertaken to determine the water use pattern of mature velvet mesquite at all stages in the annual growth cycle within a soil mass 6 m deep extending 15 m horizontally beyond the edge of the tree crowns. Water moves from the soil into the plant roots, and through the plant as liquid flow in response to gradients in water potential between the soil and the-____