Because of a change in contractors processing the samples, differences arose in the 

 level of taxonomic identifications between the 1990 and post- 1990 sample sets particularly 

 in the identification of oligochaete worms (Annelida). The initial contractor was apparently 

 unfamiliar with the group so all specimens were recorded simply as Oligochaeta. 

 Examination of later samples indicated the presence of a number of species, including 

 Tubificoides benedini and Tectadrilus gabriellae , two of the most numerically abundant 

 taxa. Attempts to locate the 1990 specimen collection were unsuccessful making it 

 impossible to reexamine the specimens or measure biomass. 



2.5 Statistical Analyses 



Soft-clam and bait-worm abundances are reported on a per square-meter basis by 

 sampling method: core or pit and rake. Pit and rake data could not be analyzed statistically 

 due to differences in sample area and sampling method, however, abundances from core 

 samples could be evaluated using Analysis of Variance (ANOVA). Abundance data for 

 bait-worms and soft-clams were tested for normality and heterogeneity of variance prior to 

 ANOVA and transformed where necessary to conform to assumptions of the test. Sheep 

 Island Nereis virens and Mya arenaria data required fourth-root (x^''*) transformations as 

 did clam- worm densities from Reals Island. Too few specimens were collected to permit 

 analysis of M. arenaria at Beals Island and Glycera dibranchiata at either site. Data were 

 tested using a two-way ANOVA with sampling date and site (constructed flat or reference) 

 as the main effects. When either site or year effects were significant (p<0.05) Tukey's test 

 was empoyed to determine differences between means. Where the Site by Year interactions 

 were significant, the main effects could not be interpreted (Zar, 1996). Linear contrasts 

 were performed in order to determine where significant differences occurred between sites 

 among the sampling dates using the Bonferroni adjustment (p = 0.05/no. comparisons) to 

 correct for multiple comparisons (Underwood, 1997). Since there are five relevant site by 

 date combinations (e.g.. Sheep Island Constructed Flat 1991 vs. Reference 1991), a p 

 value of 0.01 was required for a comparison to be considered statistically significant. 

 Where only four comparisons were possible (e.g., biomass data) a p value of 0.0125 was 

 required. 



Soft-clam and bait-worm population structures were examined by construction of 

 size frequency histograms. Measurements for individual species were pooled by site and 

 date and the relative abundance of animals in each of at least 10 size classes were plotted. 

 A minimum sample size of 30 animals was required in order to reduce the influence of a 

 few very large or very small animals. In general, this restricted the size frequency analyses 

 to the 1991-1992 sample collections and excluded consideration of Glycera populations. 



Ecological Monitoring of a Constructed Intertidal Flat at Jonesport, Maine 



