Lenihan and Michell; Biological effects of shiellfish harvesting on oyster reefs 



91 



portion of deaci, naturally occurring clams (P=0.003; 

 Table 4). The proportion of dead clams in both creeks 

 was much higher on harvested than on unharvested (i.e. 

 control) reefs (SNK P<0.05 for both contrasts; Fig. 4) 

 but was similar among the three harvest treatments 

 (SNK, P>0.05 for both contrasts; Fig. 4). 



After harvesting, the density of live and dead hatch- 

 ery-raised clams transplanted to reefs at the beginning 

 of the experiment tended to vary with the interaction 

 of creeks and harvest treatment, although not signif- 

 icantly (ANOVA, creek x harvest treatment interac- 

 tion, P=0.07-0.08; Table 5 ). However, the density of live 

 transplanted clams varied between creeks (P=0.03; 

 Table 5) and among harvest treatments (P=0.04; Table 

 5). More transplanted clams were recovered alive in 

 Pages Creek (mean ± ISD: 3.21 ±1.62/m2) than in Whis- 

 key Creek (2.22 ±1.45/m''). Fewer live transplanted 

 clams were recovered from clam-harvested plots than 

 from control plots in both creeks ( SNK, P the interac- 

 tion of <0.05; Fig. 5). The number of dead transplanted 

 clams found after harvesting also varied between 



creeks (Pages Creek>Whiskey Creek; P=0.0001; Table 

 5) but did not vary significantly with harvest treat- 

 ment (P=0.10; Table 5). At Pages Creek, there was 

 a slight trend for greater mortality of transplanted 

 clams on clam-harvested and clam- and oyster-har- 

 vested plots than in oyster-harvested and control plots 

 only (Fig. 5). Most transplanted clams placed on reefs 

 at the beginning of the experiment were not found at 

 the end of the experiment ("missing" clams; Fig. 5). The 

 number of missing transplanted clams differed with the 

 interaction of creeks and harvest treatment (ANOVA, 

 creek x harvest treatment interaction, P=0.03; Table 5) 

 because fewer clams were recovered in our census in 

 the oyster-harvested plots than in clam-harvested plots 

 at Whiskey Creek only (SNK; P<0.05; Fig. 5). 



Discussion 



Our results clearly demonstrate that both clam and 

 oyster harvesting significantly reduce oyster popula- 



