grounded and in the upstream control plot were much lower, 11.4% 

 and 6.6% respectively. As previously noted, corrals in both of 

 these areas were destroyed. 



PROBLEMS WITH SMALL SAMPLE SIZES 



We found no statistically significant differences (P < 0.050] 

 in shell damage and mortality of mussels from the four plots 

 (Table 4) . When data from the two fleeted plots were pooled and 

 compared to the downstream control, the P value for mortality in 

 Amblema plicata was 0.096 and for shell damage in all species 

 pooled was 0.074 — none of the other values were close to 0.05. 

 Sample sizes were small, and in fact no damaged Leptodea fragilis 

 were recovered. Mussels with heavy shells, such as Amblema 

 plicata , are scraped and pushed down in the mud when barges are 

 grounded on them. We believe fragile-shelled species are crushed 

 to pieces. Since divers did not recover small shell fragments 

 (even if they had the particular fragment with the number on it 

 might not have been recovered) , our results probably 

 underestimate the number of fragile-shelled species damaged or 

 killed. With larger sample sizes, between-plot differences in 

 damage and mortality of heavier-shelled species are more likely 

 to be significant. 



Shell growth rates differed between plots for Amblema 

 plicata , the species with the largest sample size (Table 5) . The 

 P value for Leptodea fragilis , the species with the next largest 



25 



