Complete hydrochemical data are rarely reported for coal-mine discharges (CMD). This report summarizes major and trace-element concentrations and loadings for CMD at 140 abandoned mines in the Anthracite and Bituminous Coalfields of Pennsylvania. Clean-sampling and low-level analytical methods were used in 1999 to collect data that could be useful to determine potential environmental effects, remediation strategies, and quantities of valuable constituents. A subset of 10 sites was resampled in 2003 to analyze both the CMD and associated ochreous precipitates; the hydrochemical data were similar in 2003 and 1999. In 1999, the flow at the 140 CMD sites ranged from 0.028 to 2210 L s-1, with a median of 18.4 L s-1. The pH ranged from 2.7 to 7.3; concentrations (range in mg/L) of dissolved (0.45-??m pore-size filter) SO4 (34-2000), Fe (0.046-512), Mn (0.019-74), and Al (0.007-108) varied widely. Predominant metalloid elements were Si (2.7-31.3 mg L-1), B (<1-260 ??g L-1), Ge (<0.01-0.57 ??g L-1), and As (<0.03-64 ??g L-1). The most abundant trace metals, in order of median concentrations (range in ??g/L), were Zn (0.6-10,000), Ni (2.6-3200), Co (0.27-3100), Ti (0.65-28), Cu (0.4-190), Cr (<0.5-72), Pb (<0.05-11) and Cd (<0.01-16). Gold was detected at concentrations greater than 0.0005 ??g L-1 in 97% of the samples, with a maximum of 0.0175 ??g L-1. No samples had detectable concentrations of Hg, Os or Pt, and less than half of the samples had detectable Pd, Ag, Ru, Ta, Nb, Re or Sn. Predominant rare-earth elements, in order of median concentrations (range in ??g/L), were Y (0.11-530), Ce (0.01-370), Sc (1.0-36), Nd (0.006-260), La (0.005-140), Gd (0.005-110), Dy (0.002-99) and Sm (<0.005-79). Although dissolved Fe was not correlated with pH, concentrations of Al, Mn, most trace metals, and rare earths were negatively correlated with pH, consistent with solubility or sorption controls. In contrast, As was positively correlated with pH. None of the 140 CMD samples met all US Environmental Protection Agency (USEPA) continuous-concentration criteria for protection of freshwater aquatic organisms; the samples exceeded criteria for Al, Fe, Co, Ni, and/or Zn. Ten percent of the samples exceeded USEPA primary drinking-water standards for As, and 33% exceeded standards for Be. Only one sample met drinking-water standards for inorganic constituents in a public water supply. Except for S, the nonmetal elements (S > C > P = N = Se) were not elevated in the CMD samples compared to average river water or seawater. Compared to seawater, the CMD samples also were poor in halogens (Cl > Br > I > F), alkalies (Na > K > Li > Rb > Cs), most alkaline earths (Ca > Mg > Sr), and most metalloids but were enriched by two to four orders of magnitude with Fe, Al, Mn, Co, Be, Sc, Y and the lanthanide rare-earth elements, and one order of magnitude with Ni and Zn. The ochre samples collected at a subset of 10 sites in 2003 were dominantly goethite with minor ferrihydrite or lepidocrocite. None of the samples for this subset contained schwertmannite or was Al rich, but most contained minor aluminosilicate detritus. Compared to concentrations in global average shale, the ochres were rich in Fe, Ag, As and Au, but were poor in most other metals and rare earths. The ochres were not enriched compared to commercial ore deposits mined for Au or other valuable metals. Although similar to commercial Fe ores in composition, the ochres are dispersed and present in relatively small quantities at most sites. Nevertheless, the ochres could be valuable for use as pigment.