In January 2001, mercury (Hg) was detected (500 nanograms per liter, ng/L, or greater) in the distribution system of the Long Neck Water Company (LNWC), Pot Nets, Delaware. By April 2001, two LNWC production wells had been taken off-line because discharge concentrations of total mercury (HgT) either had exceeded or approached the Federal limit of 2,000 ng/L. From October 2003 through January 2005, the U.S. Geological Survey, Delaware Geological Survey, and Delaware Department of Natural Resources and Environmental Control conducted a cooperative study to (a) determine if the Hg contamination was widespread, (b) identify possible forms of Hg in ground water, and (c) examine Hg occurrence in relation to (geo)chemical conditions and characteristics of ground water and sediment in the surficial aquifer on the Long Neck Peninsula, Sussex County, Delaware. An initial water-quality survey conducted with samples from 22 production wells revealed that concentrations of HgT in ground water in the surficial aquifer ranged from 0.11 to 1,820 ng/L. Shallow ground water (less than 120 feet below land surface) throughout most of the peninsula, including that which contained elevated concentrations of HgT (exceeding 100 ng/L), appeared to be affected by human activities. All samples contained volatile organic compounds (VOCs) and elevated nitrate-nitrogen (NO3-N, exceeding 0.4 milligrams per liter, mg/L). Most (16 of 22) samples had elevated specific conductance (SC, in excess of 100 microsiemens per centimeter at 25 degrees Celsius). Elevated concentrations of HgT, however, only occurred in five production wells in the Pot Nets Bayside and Lakeside communities. The vertical distribution of HgT in shallow ground water (less than 80 feet below land surface) was determined with samples collected at 5 to 6 vertical-nest short-screened (2 - 5-foot length) monitoring wells installed near Bayside and Lakeside production wells with the highest HgT concentrations (exceeding 1,000 ng/L). Elevated concentrations ofHgT (100-6,380 ng/L) occurred in the shallow aquifer near each well at different depths. Chemical analyses of selected soil, fill, and aquifer sediment samples, obtained during the installation of nested wells, indicated that little HgT occurred in soil or fill at either site (40 micrograms per kilogram, ?g/kg, or less). No HgT was detected (less than 20 ?g/kg) in aquifer sediment samples. These low HgT concentrations imply that neither the soil, fill, nor aquifer sediment was a likely source of the elevated Hg in ground water. Given Hg occurrence appeared to be a ground-water transport phenomenon, the forms of Hg in transport were investigated. Differences in HgT concentrations between raw and filtered (0.1- and (or) 0.4-absolute micrometer pore size) samples from nested wells were random in sign and similar in magnitude to the variability in measuring HgT attributed to field and laboratory methods (? 5-10 percent, for HgT concentrations exceeding 100 ng/L). Thus, Hg transport likely occurred in a dissolved or fine-colloidal nonparticulate phase. Methyl mercury (HgMe) only was detected at low concentrations (0.06 ng/L or less) in nested-well samples with low to moderate concentrations of HgT (less than 366 ng/L). Whether HgMe occurred at similar concentrations in samples with high HgT concentrations was unresolved due to a sample-matrix interference problem. Potential complex forms of Hg were investigated in relation to the occurrence of selected ligands (organic carbon, sulfide, and chloride concentrations) and geochemical conditions (for example, pH and dissolved oxygen concentrations). Only dissolved organic carbon (DOC) appeared directly related to Hg occurrence. Elevated concentrations of HgT and DOC co-occurred in ground water at both Pot Nets sites. The average concentration of DOC was about four times greater in samples from the Pot Nets wells with the highest HgT concentrations (exceeding 1,000 ng/L) than in most Pot Nets o