FISHERY BULLETIN: VOL. 85, NO. 2 



Table 6.— The impact of clam harvesting on average density ( + SE) of benthic macroinvertebrates 

 per 0.008 m^. n = 6 samples for each treatment matrix at each sampling date. Samples were 

 taken to 25 cm and passed through 1 mm mesh. Superscripts A-C indicate significant differences 

 among matrices in Duncan's test at a = 0.05, with those means sharing capital letter superscripts 

 not differing significantly. 



' *• - P < 0.01, •*• - P < 0.001, NS - P > 0.05 in one-way ANOVA's (for each separate date) and two-way 

 ANOVA's (for sums) on average macroinvertebrate counts per core (transformed by log (x + 1)). 



in species composition as a function of the intense- 

 kicking treatment. Table 7 presents the results of 

 these species identifications and shows that no 

 major shift in species composition of the most 

 abundant species occurred in either the sand-flat 

 or seagrass habitat following the application of 

 the intense-kicking treatment. Polychaetes domi- 

 nated the fauna of both habitats and the same 

 species of polychaetes tended to be represented at 

 similar densities both before and after intense clam 

 kicking. 



Bay Scallop Densities 



Bay scallops were never encountered in sampling 

 the sand-flat matrices, so we have no test of whether 

 clam harvest treatment affects bay scallops in areas 

 lacking seagrass. One-way ANOVA's on log {x + 1)- 

 transformed counts (which removed heteroscedacity 

 in Cochran's tests) demonstrated significant (a = 

 0.05) differences among seagrass matrices in aver- 

 age bay scallop density on only 2 sampling dates, 

 fall 1980 and fall 1983 (Table 8). Duncan's test on 

 the fall 1980 data showed that bay scallop density 

 in control I was significantly (P < 0.05) lower than 

 in every other matrix except intense-kicking II, and 

 that there were no other significant differences 

 between pairs of matrices. Because the fall 1980 

 sampling occurred before the major application of 

 clam harvest treatments (see Table 2), this sampling 



period may be considered a pretreatment sampling. 

 Extremely low seagrass biomass in control I in fall 

 1980 (Table 5) may explain the significantly lower 

 bay scallop densities in that matrix on that date. 



The fall 1983 sampling occurred after a period of 

 more successful bay scallop recruitment than oc- 

 curred before any other sampling date (Table 8) and, 

 thus, provided more "substrate" on which effects 

 of clam harvest treatments may have operated. Dun- 

 can's test on mean bay scallop densities for fall 1983 

 demonstrated that the matrices split into two 

 separate groups: a low-density group, made up of 

 control I and the 2 intense-kicking matrices, and a 

 high-density group, comprised of control II, the 

 raking, and light-kicking matrices (Table 8). Within 

 each group, no matrices differed significantly (a = 

 0.05) from any other, but all differences between 

 groups were statistically significant. Because fall 

 1983 bay scallop densities were so much greater 

 than at any other sampling date, the sums over all 

 five sampling periods also exhibited significant dif- 

 ferences among matrices in an analogous two-way 

 ANOVA, and Duncan's tests separated the matrices 

 into groupings virtually identical to those detected 

 for the fall 1983 data set alone (Table 8). 



A contrast of the bay scallop results of fall 1980 

 and fall 1983 demonstrates that after application of 

 the second intense-kicking treatment in the seagrass 

 habitat in winter of 1980-81, bay scallop densities 

 declined to join the already low value of control I, 



292 



