also tends to rule out a single treatment technology. Selection of a treatment regime that includes in situ bioremediation may be precluded by elevated levels of toxic metals that could constrain microbial growth or by the limited biological and chemical availability of contaminants.

Pilot studies for the in situ biological treatment of sediments are limited to a few examples in which the sediment volumes were small and the contaminant composition limited. The committee was told of four projects, two involving PCBs and two involving polyaromatic hydrocarbons (Thoma, 1994).

The in situ bioremediation of PCBs has been carried out in freshwater sediments in the Housatonic River in Massachusetts and the Hudson River in New York (Harkness et al., 1993). The volumes involved were less than a few cubic yards; only small portions of the contaminated sites were studied. The Massachusetts field studies were carried out based on laboratory data indicating that bromobiphenyls stimulate anaerobic microbial attack on PCBs and that highly chlorinated congeners are dechlorinated to produce molecules with fewer chlorine atoms. The pilot studies were successful in showing that after 373 days, the concentration of highly chlorinated congeners (containing 6 to 9 chlorine atoms per molecule) declined from 68 to 18 percent of all PCB molecules, with a corresponding increase in the species with fewer chlorine atoms. Although total PCB levels did not change, the data suggest that toxicity was reduced because the most toxic congeners with "dioxin-like" properties were preferentially dechlorinated. Whether this result constitutes remediation depends on regulatory requirements. Given the current regulations, which are based on total PCB content, the novel capability of stimulating anaerobic PCB transformation may have limited practical use, and further treatment would be needed.

Field studies in the Hudson River showed that in situ aerobic biodegradation is limited by physical and chemical factors unrelated to the microbial community. These factors include, for example, the sorption of contaminants into the sediment matrix and the consequent reduction in contaminant biogeochemical and biological availability, oxygen and nutrient availability, mixing, and the survival of externally amended active organisms.

In sum, the in situ bioremediation of PCB-contaminated sediment has only recently been recognized as a potential alternative. Although the technology looks promising, given the current level of application and the regulatory focus on total PCBs, it is unclear whether in situ bioremediation can achieve the cleanup levels required at a reasonable cost. If additional nutrients are needed, the sheer volume of and contaminant mixtures of most marine sediments will present difficulties for handling and monitoring.

Available evidence suggests that PCB dechlorination and biodegradation occur more slowly in marine sediments than in land-based systems, but the in situ degradation rates of sediments have not been measured with any reliability. Furthermore, bioremediation rates would be affected by site-specific characteristics, such as sediment composition, hydrodynamics, pore water composition, and