We have used electrical resistivity tomography (ERT) as a tool to identify locations for four rural boreholes in southern Malawi. Integrating the geophysical data with local geology, we determined a preferred siting location and a larger area likely to produce an acceptable borehole. We then engaged in a dialog with local stakeholders to help them determine a socially and politically acceptable site that we were confident would produce an adequate water supply. This presentation concentrates on the local geology, the ERT methods and results, and the completed boreholes. The geology of southern Malawi is dominated by the southern arm of the East African Rift. Our study site in rural Zomba district is east of the rift valley. Most of the study area is underlain by crystalline metamorphic rocks that predate the rifting, but several large plutons (including Zomba Mountain) and smaller dikes are related to Jurassic rifting. The shallow-water Lake Chilwa occupies a large plain east of Zomba Mountain. At one time the lake occupied much of the plain, but is slowly receding to the northeast. Hills consisting of Cenozoic igneous rocks and resistant quartz-rich meta-sediments protrude from the otherwise flat-lying veneer of sediments within the Lake Chilwa basin. Three of the four villages in our study are situated near the base of two of these hills. An important part of our investigation in these villages was to determine whether sufficient saturated thickness could be expected from new boreholes to sustain continuous and year-round pumping. The other village is a fishing village along the shore of the lake. With its internal drainage, the lake water is seasonally brackish to saline. Most of the fishing villages along the shores of the lake have difficulty finding reliable, fresh, groundwater sources. We acquired multiple ERT profiles in each village using an ABEM Terrameter 1000 with 64 electrodes at a nominal spacing of 5 meters and a Wenner configuration. Typical resistivity inversions were 45 to 55 m in depth before elevation corrections. In total, we acquired 15 ERT profiles covering 7,260 meters. ERT profiles at two villages near Chanda Hill clearly distinguished the shallow weathered and water-bearing material from the underlying unweathered metamorphic rocks. At one of these villages we were also able to identify a buried, weathered dolorite dike. ERT profiles at the third hillside village were acquired parallel to the hill and showed a generally constant depth to the unweathered rock. Near the lake, very low resistivity values (4 to 8 ohm-m) likely indicate saline water intrusion into the groundwater. Further inland, resistivity values abruptly increase, suggesting the presence of fresh-water sands. Based on the ERT profiles and conversations with local stakeholders, we drilled successful boreholes in each of the four villages. Water levels in the boreholes ranged from 19 to 35 m and total depth varied from 39 to 52 m. Each borehole was completed with a brick and concrete apron and equipped with a standard AfriDev hand pump. To ensure sustainability of the hand pumps, water point committees have been established, trained on borehole maintenance, and linked to suppliers of spare parts and existing local governance structures.