Trends in PCBs, DDT, and other contaminants have been monitored in Great Lakes lake trout and walleye since the 1970s using composite samples of whole fish. Dramatic declines have been observed in concentrations of PCB, ΣDDT, dieldrin, and oxychlordane, with declines initially following first order loss kinetics. Mean PCB concentrations in Lake Michigan lake trout increased from 13 μg/g in 1972 to 23 μg/g in 1974, then declined to 2.6 μg/g by 1986. Between 1986 and 1992 there was little change in concentration, with 3.5 μg/g observed in 1992. ΣDDT in Lake Michigan trout followed a similar trend, decreasing from 19.2 μg/g in 1970 to 1.1 μg/g in 1986, and 1.2 μg/g in 1992. Similar trends were observed for PCBs and ΣDDT in lake trout from Lakes Superior, Huron and Ontario. Concentrations of both PCB and ΣDDT in Lake Erie walleye declined between 1977 and 1982, after which concentrations were relatively constant through 1990. When originally implemented it was assumed that trends in the mean contaminant concentrations in open-lake fish would serve as cost effective surrogates to trends in the water column. While water column data are still extremely limited it appears that for PCBs in lakes Michigan and Superior, trends in lake trout do reasonably mimic those in the water column over the long term. Hypotheses to explain the trends in contaminant concentrations are briefly reviewed. The original first order loss kinetics used to describe the initial decline do not explain the more recent leveling off of contaminant concentrations. Recent theories have examined the possibilities of multiple contaminant pools. We suggest another hypothesis, that changes in the food web may have resulted in increased bioaccumulation. However, a preliminary exploration of this hypothesis using a change point analysis was inconclusive.