measured at 30-sec. intervals in the same sequence as the tests were initiated. Between each extract 

 dilution level, the blank of the corresponding concentration was used to adjust the photometer for the 

 contribution of the extraction solvent. To conclude the tests, light emission was measured again at 15 

 min., and these data were used to calculate the 50% inhibition concentrations (i.e., the EC50s). 



In addition to the tests described above that were performed with all of the samples, several others were 

 performed on selected samples as a part of methods development. Tests of the growth of a polychaete 

 (Armandia brevis) and an adult sand dollar (Dendraster excentricus) were performed with 17 of the 

 samples (Rice et al., in press). Also, nine of the samples were tested with the freshwater amphipod 

 Diporeia spp. by the Great Lakes Environmental Research Laboratory (Dr. Peter Landrum). The test 

 animals were acclimated to 20 ppt salinity seawater by the addition of 5 ppt seawater/day and held for 

 48 hours. Two cm of sieved sediments, in replicates of three per sample, were placed into one-liter 

 beakers, containing 600 mL of seawater at 20 ppt salinity. Twenty animals were used in each replicate. 

 Tests were performed at 4°C, maintained by a constant temperature water bath. The amphipods were 

 monitored daily for sediment avoidance, signs of stress, and mortality. Avoidance of the sediment was 

 observed as the absence of burrowing and migration to the water surface, which resulted in adherance 

 to the surface film. Stress was observed as animals lay on the sediment surface. Dead animals were 

 recorded and removed from the exposure chambers. After 28 days, the beakers were removed from the 

 water bath and the sediment was wet sieved through a 1 mm screen. The numbers of live and dead 

 animals were recorded and the percent mortality and percent survival were calculated. 



Estimates of the Spatial Extent of Toxicity. The spatial extent of toxicity within the survey area was 

 estimated using methods similar to those of the Environmental Monitoring and Assessment Program 

 (EMAP) of the U.S. EPA (Schimmel et al., 1994). However, the design of the sampling plans differed 

 between Phases 1 and 2. During Phase 1, the dimensions of each sampling zone (Figure 4) were 

 outlined on navigation charts during the design phase. The locations of each sampling site were deter- 

 mined a priori to represent conditions within each zone. These site locations were chosen following a 

 review of existing information of sediment types, bathymetry, and proximity to previously sampled 

 sites. The size of each zone was determined with a planimeter. The toxicity data were weighted to the 

 size of each zone (divided by three, the number of sites in each zone), and the cumulative distribution 

 functions of these data were prepared. Using critical values of toxicity results less than 80% of the 

 control responses (as in the EMAP) and less than 20% of controls (reciprocal of 80%), the size(s) of the 

 area(s) that were significantly toxic and highly toxic, respectively, were estimated. 



The principles of a probabalistic sampling design require that the sampling locations be chosen ran- 

 domly and without knowledge of site-specific conditions (Schimmel et al., 1994). However, that type 

 of sampling design was not strictly adhered to in Phase 1 of this survey. The boundaries and dimen- 

 sions of each zone were established a priori, but the locations of the sampling sites were not selected 

 with a strictly random process. Some sites were chosen to coincide with the locations of sites previ- 

 ously sampled by other investigators. However, none were chosen to represent conditions near any 

 point sources or waste disposal sites. All sites were chosen to represent conditions in the nearby 

 vicinity of the sampling location and within the respective zone. The locations of the individual sam- 

 pling stations at each site were chosen by the vessel operator in the field as three points on a compass 

 radiating from the site center. 



Highly disturbed areas that obviously had been recently dredged were avoided. Also, samples with 

 excessive amounts of coarse sandy materials were avoided, where possible. Within each site, attempts 

 were made in the field to avoid a mixture of stations from deep, dredged channels and shallow, undredged 



22 



