Taylor et al ; Habilat use by summer-spawned Pomatomus saltolrix on the inner continental shelf off southern New Jersey 



595 



Figure 1 



Map of inshore and inner continental shelf habitats off southern and central New Jersey. Plus 

 symbols ' + i demarcate the 4x4 station grid (total area -825 km-) sampled with a surface- 

 deployed Methot trawl on the inner continental shelf between 11 August and 9 October 1998. 

 lAi Inner continental shelf sampling sites were grouped into four transects (I-IVi aligned 

 parallel to the coastline. Estuarine (•), inlet (O), and ocean beach iD) sites sampled with a 

 beach seine were located at the margins of Great Bay (GBi, Little Egg Harbor iLEHi, and 

 Little Egg Inlet (LEI). (B) Solid lines (^) demarcate six depth strata sampled with an otter 

 trawl on the inner continental shelf. Location of the Rutgers University Marine Field Station 

 (Tuckerton, NJ; Bi and depth contours of 10, 20, and 30 m are provided. 



Bluefish collected with Methot trawls were counted 

 to estimate fish density (no. of bluefish/1000 m-M, mea- 

 sured to the nearest mm SL, and classified into distinct 

 ontogenetic stages based on the individuals body size 

 and presumed scale development (Silverman, 1975; Able 

 and Lamonaca, in press). For the purposes of our inves- 

 tigation, several life history stages were demarcated: 

 larval, transitional, and juvenile. The onset of scale 

 development was recognized as the end of the larval 

 stage and the beginning of the transitional period (12 

 mm SL; Silverman, 1975; Hare and Cowen, 1994). The 

 transitional period in bluefish development, in turn, 

 concluded with complete scale formation and the juve- 

 nile stage began at 34 mm SL (Silverman, 1975). 



Spatial and temporal patterns of bluefish density and 

 mean size were analyzed independently with multivari- 

 ate repeated-measures analysis of variance (ANOVA) 

 models by using station transects (transects I-IV; Fig. 

 lA) and sampling date (day of year) as the between- 

 subject and within-subject factors, respectively. Profile 

 transformation of the within-subject factor was used 



in the repeated-measures ANOVA model, and statisti- 

 cal significance was estimated from the Greenhouse- 

 Geisser-adjusted probability to avoid violating the as- 

 sumption of circularity (sphericity) of the within-subject 

 variance-covariance matrix (von Ende, 1993), The mean 

 density and body size of bluefish across four levels of 

 transects and ten levels of sampling dates were con- 

 trasted with a Ryan-Einot-Gabriel-Welsch (Ryan's Q) 

 multiple comparison test (Day and Quinn, 1989). More- 

 over, natural log (x-fl) and natural log (x) transforma- 

 tions were performed on bluefish density and mean-size 

 data, respectively, to meet assumptions of normality 

 and homogeneity of variance. When data transforma- 

 tions did not achieve homoscedasticity, hypotheses were 

 rejected at alpha values lower than the P-values of 

 Levene's test for homogeneity of variance (Underwood, 

 1981). 



The effects of several environmental parameters on 

 the spatial and temporal distribution of bluefish density 

 and mean size were analyzed with a stepwise multiple 

 regression. The variables included in the regression 



