*Undisturbed soil cores from the rooting zone of three trees in each ecosystem were removed using a split steel, continuous sampling tube, 1.56 m long with 6.5 cm i.d. This coring device, mounted on a truck, was modified from Kelley et al. (1947). The split-tube bit fits into an outer, rotating auger bit that served as a continuous casing to prevent cave-in. As the outer bit cut through the soil, the inner, nonrotating bit was pressed into the soil. Cores were collected at the edge of the mesquite canopy. The core retainer and the two halves of the split sampling tube were cleaned of all residual soil. Their interior surface was flame sterilized with 95% ethanol before being put together for sampling. Soil samples were removed from the surface 1 m of soil in 0.5 m increments, and thereafter in 1 m increments. Flame sterilized trowels and spatulas were used to replace each depth increment into a clean plastic bag. These were put into icecooled chests, and transported to the Univ. of California, Riverside, where they were subdivided for analysis. Drilling depth for each core was determined by either the absence of roots in two consecutive 1.56 m sampling tube lengths, or the presence of coarse, dry loose soil that could not be retained in the tube. The number of cores (three per ecosystem per sampling) collected was limited by the expense of obtaining the specialized drilling equipment used in this study. Sampling dates at New Mexico were in January 1986, the midpoint of the dormant season; late May 1986, during peak growth; and early October 1986 following the summer rains. The grassland site was not sampled on the Jan-Feb drillings. Each bag of soil representing a depth increment was mixed thoroughly before subsampling. Using trowels and spatulas flame sterilized with 95% ethanol, each soil sample was sub-divided for various analyses. Gravimetric water content of the soil samples was determined at the time of subsampling(weight of water/weight of dry soil). Soils for chemical analysis were then air-dried in a glass-house, ground to break-up clay and caliche aggregates, and passed through a 2 mm sieve. Potassium chloride-extractable NH4-N and NO3-N (Keeney and Nelson, 1982) and NaHCO3-extractable P (Olsen and Dean 1965) were measured colorimetrically using a technicon Auto analyzer (technicon Instruments Corp). Roots were separated from field-moist soil without sieving by flotation using an elutriator fitted with a 20 mesh seive (Byrd et al. 1976). Fall soil Total Kjeldahl N, Total P, NH4-N and NO3-N analyses were done on field moist soils rather than dried soil. Total N (and Total P) was digested by Kjeldahl digestion block techniques (Bremner and Mulvaney, 1982). Digest NH4-N was analyzed using an automated salicylate procedure (Technicon Industrial Method No. 329-74W/B). Concentrations of mesquite-nodulating rhizobia in field-moist soil samples were estimated using the plant-infection, most probable-number (MPN) technique (Vincent 1970). Tenfold dilutions (10 -1 to 10 -6), starting with 10 g of subsmple in 90mL of sterile, buffered saline (0.15 M NaCl, 0.002 M KH2PO4, 0.004 M Na2HPO4, pH 7), and four replicate mesquite plants per dilution, were used for the MPN tests. Mesquite seedlings were germinated from seed collected from the Jornada. Seeds were surface sterilized and scarified by exposure to concentrated sulfuric acid for 3 min followed by six washes in sterile, distilled water. One sterilized seed was planted in sterile vericulite contained in plant tubes (Garvin and Lindemann 1983). The plant tubes had been sterilized with 0.5% sodium hypochlorite and rinsed thoroughly in sterile water. Each mesquite seedling was inoculated with 1 mL of an appropraite soil dilution. After inoculation, the plants were randomized, and grown in a glasshouse where temperatures ranged from 25 to 35 C under natural lighting conditions. During winter, supplemental light (cool-white flourescent lamps) situated 1 m above the plants was provided to maintain a 14-h daylength. On alternate days the plants were irrigated with sterile, 0.25 strength Hoagland's nutrient solution lacking N. Noninoculated plants served as controls. Plants were grown for 6 to 8 weeks before they were examined for nodulation. The MPN for each sample was calculated using Table 3.5A in Vincent (1970).