A comprehensive review of published information on the distribution and behavior of pesticides and their transformation products in ground water indicates that pesticides from every chemical class have been detected in ground waters of the United States. Many of these compounds are commonly present at low concentrations in ground water beneath agricultural land. Little information is available on their occurrence beneath non-agricultural land, although the intensity of their use in such areas (on lawns, golf courses, rights of way, timberlands, etc.) is often comparable to, or greater than agricultural use. Information on pesticides in ground water is not sufficient to provide either a statistically representative view of pesticide occurrence in ground water across the United States, or an indication of long-term trends or changes in the severity or extent of this contamination over the past three decades. This is largely due to wide variations in analytical detection limits, well selection procedures, and other design features among studies conducted in different areas or at different times. Past approaches have not been well suited for distinguishing "point source" from "nonpoint source" pesticide contamination. Among the variety of natural and anthropogenic factors examined, those that appear to be most strongly associated with the intensity of pesticide contamination of ground water are the depth, construction and age of the sampled wells, the amount of recharge (by precipitation or irrigation), and the depth of tillage. Approaches commonly employed for predicting pesticide distributions in the subsurface--including computer simulations, indicator solutes (e.g., nitrate or tritium), and ground-water vulnerability assessments--generally provide unreliable predictions of pesticide occurrence in ground water. Such difficulties may arise largely from a general failure to account for the preferential transport of pesticides in the subsurface. Significant improvements in understanding and predicting the occurrence and fate of pesticides in ground water are likely to depend on: (1) greater coordination of ground-water sampling across the nation to ensure consistency of study design, and thus comparability of results; (2) more extensive analyses for pesticide transformation products during ground-water monitoring studies; (3) substantially enhanced communication among investigators conducting laboratory experiments, small-scale field studies and large-scale monitoring studies; and (4) more routine testing of predictions of pesticide behavior and ground-water vulnerability against actual field observations of pesticide occurrence in ground water