Two municipal supply wells in Lakewood, Washington, were found to be contaminated with trichloroethylene, transdichloroethylene, and tetrachloroethylene. Sequential samples were taken for chemical analyses, in conjunction with drawdown measurement during aquifer (pump) tests designed to help characterize the problem site. Plots of the chemical concentrations against duration (or volume) or pumping yielded patterns which lent insight into the orientation of the contaminant source. Analogies are drawn here between interpretation of these chemical time-series patterns and interpretations of breakthrough curves in tracer experiments. Examination of the data also infer the distinct possibility of serious interpretation errors arising from conventional sampling.