Large-scale degradation has resulted in a widespread conversion of Chihuahuan Desert grasslands to shrublands dominated by mesquite (Prosopis glandulosa).  Efforts to control mesquite and re-establish native grassland vegetation have had limited success, possibly because essential water and soil resources are not retained on site in these mesquite dune-dominated systems.  Microcatchments may facilitate restoration of natural processes by reducing soil and water losses from the site.  In 2003, we initiated an experiment to determine the integrated effects of vegetation manipulation treatments on mesquite control and grassland restoration in southern New Mexico. Treatments tested were herbicide application (Triclopyr and Clopyralid), and creation of microcatchments, followed by reseeding in June or August; the experiment was repeated in 2004 and used a randomized complete block design. Growing season (July-September) precipitation was highly variable: plots received 22%, 110%, 37%, and 310% of average growing season precipitation in 2003 – 2006.  In 2006, perennial grass cover, mesquite cover and bare ground were affected by herbicide treatment (p< 0.001), but no effects of microcatchments or planting date were detected (p>0.05).  Herbicide treated plots had 35% perennial grass, 2% mesquite, and 38% bare ground cover, compared to 2% perennial grass, 20% mesquite and 70% bare ground cover on unsprayed plots.  New plant establishment in 2006 indicates the seedbank persisted through at least 2 growing seasons.  Observations of perennial grass establishment under mesquite canopies and in interspaces, and lack of microcatchment effects suggest that controlling mesquite was more important than altering microsites for revegetation success.