Abstract-Physicochemical characteristics of sediment unrelated to contaminant levels and bioavailability may influence the outcome of toxicity tests. In particular, sediment organic matter content has the potential to be a confounding factor in toxicity tests using the midge larva Chironomus tentans because the larvae are infaunal and feed on organic matter in the sediments. To examine this possibility, we conducted a series of tests using formulated sediments with varying organic matter contents (0-20%) following the standard U.S. Environmental Protection Agency (U.S. EPA) 10-day C. tentans growth and survival protocol. Formulated sediments made with peat moss, ␣-cellulose, and maple leaves were tested. An organic-rich natural sediment diluted with formulated sediment to achieve a range of organic matter contents was also examined. In a final experiment, sediments containing each of the four organic matter sources at the same concentration were tested against one another. Survival was not greatly affected by concentration of organic matter, except at the lowest concentrations in natural sediment (6-7%), where survival dipped below 70%. In experiments using peat moss, ␣-cellulose, and maple leaves, significant differences in C. tentans growth (measured as ash-free dry weight) were found at different organic matter concentrations. In contrast, concentration of organic matter in the natural sediment dilution series had little effect on growth, perhaps because much of this material was highly refractory. In the comparison experiment, growth differed significantly among the four sediments, with best growth achieved with ␣-cellulose and leaves. These results suggest that both organic matter quantity and quality can be confounding factors in toxicity tests using C. tentans.

Abstract-Relatively few case studies have been conducted demonstrating the utility of multiple endpoint approaches to sediment quality assessment in areas with moderate contaminant levels and heterogeneous conditions. Because variability is commonplace in natural systems, assessment approaches must be designed explicitly for these conditions. In an assessment of highly heterogeneous freshwater sediments in Lake Champlain, Vermont, USA, we measured multiple endpoints, including physical and chemical characteristics of the sediment, acute and chronic toxicity in three tests (Ceriodaphnia dubia, Chironomus tentans, and Pimephales promelas), and benthic community composition, at 19 sites throughout Inner Burlington Harbor. Multiple regression techniques were used to investigate whether significant relationships existed between biological endpoints and the physical and chemical characteristics of the sediments. Although all three laboratory exposures indicated toxicity at some sites, little correspondence was found among the tests. No changes in the benthic community could be attributed to trace contaminants. Multiple regression was useful in showing associations between contaminants, grain size, toxicity, and benthic community composition. Although not demonstrating causal linkages, these statistical analyses suggested which factors might be important in driving measured responses. In heterogeneous sites, an individualized assessment approach based on a preliminary analysis of variability might provide the most useful information.

Relatively few case studies have been conducted demonstrating the utility of multiple endpoint approaches to sediment quality assessment in areas with moderate contaminant levels and heterogeneous conditions. Because variability is commonplace in natural systems, assessment approaches must be designed explicitly for these conditions. In an assessment of highly heterogeneous freshwater sediments in Lake Champlain, Vermont, USA, we measured multiple endpoints, including physical and chemical characteristics of the sediment, acute and chronic toxicity in three tests (Ceriodaphnia dubia, Chironomus tentans, and Pimephales promelas), and benthic community composition, at 19 sites throughout Inner Burlington Harbor. Multiple regression techniques were used to investigate whether significant relationships existed between biological endpoints and the physical and chemical characteristics of the sediments. Although all three laboratory exposures indicated toxicity at some sites, little correspondence was found among the tests. No changes in the benthic community could be attributed to trace contaminants. Multiple regression was useful in showing associations between contaminants, grain size, toxicity, and benthic community composition. Although not demonstrating causal linkages, these statistical analyses suggested which factors might be important in driving measured responses. In heterogeneous sites, an individualized assessment approach based on a preliminary analysis of variability might provide the most useful information. Keywords-Lake ChamplainToxicity tests Sediment quality assessment Benthic invertebrates Multiple regressionEnviron. Toxicol. Chem. 16, 1997 M.C. Watzin et al.

Physicochemical characteristics of sediment unrelated to contaminant levels and bioavailability may influence the outcome of toxicity tests. In particular, sediment organic matter content has the potential to be a confounding factor in toxicity tests using the midge larva Chironomus tentans because the larvae are infaunal and feed on organic matter in the sediments. To examine this possibility, we conducted a series of tests using formulated sediments with varying organic matter contents (0–20%) following the standard U.S. Environmental Protection Agency (U.S. EPA) 10‐day C. tentans growth and survival protocol. Formulated sediments made with peat moss, α‐cellulose, and maple leaves were tested. An organic‐rich natural sediment diluted with formulated sediment to achieve a range of organic matter contents was also examined. In a final experiment, sediments containing each of the four organic matter sources at the same concentration were tested against one another. Survival was not greatly affected by concentration of organic matter, except at the lowest concentrations in natural sediment (6–7%), where survival dipped below 70%. In experiments using peat moss, α‐cellulose, and maple leaves, significant differences in C. tentans growth (measured as ash‐free dry weight) were found at different organic matter concentrations. In contrast, concentration of organic matter in the natural sediment dilution series had little effect on growth, perhaps because much of this material was highly refractory. In the comparison experiment, growth differed significantly among the four sediments, with best growth achieved with α‐cellulose and leaves. These results suggest that both organic matter quantity and quality can be confounding factors in toxicity tests using C. tentans.