12 



Summary sediment grain size data (e.g., % silts, % gravel) are presented as stacked 

 bar graphs. Infaunal assemblage parameters; taxa richness (taxa/sample), total numerical 

 abundance/m^ and total wet- weight biomass/m^ were tested using ANOVA. Logarithmic 

 transformations were required for abundance and biomass. Where significant differences 

 were detected between main effects (Site or Year) or by linear contrasts, mean values jf 

 one standard error have been plotted. 



Infaunal taxonomic structure was examined using the nonparametric ordination 

 technique, Nonmetric Dimensional Scaling (NMDS) and Similarity Percentage (SIMPER), 

 a procedure that estimates the relative contribution of each taxon to overall similarity. The 

 total species list was reduced by considering only those taxa that comprised 1 % or more of 

 total abundance or were present in 50% or more of the cores. In order to conform to 

 computational limits of the statistical software, the number of samples was reduced by 

 randomly selecting only 10 cores from each sample date and site for inclusion in the 

 analyses. For NMDS, abundance values were logarithmically transformed (log x+l) and 

 Bray-Curtis (BC) similarity values calculated for all possible combinations of samples. 

 Stress, a goodness-of-fit measure, was calculated for all NMDS comparisons. Stress values 

 less than 0.2 are considered to be adequate for interpretation of results (Clarke and 

 Warwick, 1994). For SIMPER analyses, abundances were fourth-root transformed as 

 recommended by Clarke and Warwick (1994). SIMPER calculates average sample 

 dissimilarity using the Bray-Curtis (BC) dissimilarity index (dissimilarity = 1- BC value). 



Because of the oligochaete identification problem with Sheep Island 1990 data, it 

 was impossible to directly compare all years simultaneously. Instead, two separate analyses 

 were performed. First, 1990 and 1991 data were compared using the 1990 taxonomic 

 classifications (all oligochaete taxa pooled) and second, 1991 and all later samples were 

 compared using the full range of oligochaete identifications. Differences between 1990 and 

 post- 1991 data are inferred from their relationship to the 1991 results. 



Ecological Monitoring of a Constructed Intertidal Flat at Jonesport, Maine 



