268 



Fishery Bulletin 104(2) 



CM 



< 



O 

 O 



B 



» 



" o^SAL STRAT 



^'"1^ 



c 

 3 

 3 



CD 



c\j 



< 

 O 

 O 



0) 



CM 

 < 



o 

 o 



3 



CD 



CCA1 



Figure 5 



Canonical correspondence analysis ordinations (portraying the first and second dimension scores) of the \Q% cross-shelf 

 data set showing the correlations between environmental variables, species, and station groups each season; spring (A), 

 summer (B), fall (C), and winter (Dl. The solid triangles mark the locations of taxa, and the polygons enclose the bound- 

 ary of each station group with three or more stations (as in Fig. 3). If stations were not grouped, circles mark the loca- 

 tions. The arrows depict the gradient of each environmental variable (temperature = btemp, salinity=bsal, density=bden, 

 stratification = strat, and depth = dep). The dashed lines intersect at the origin of the plot. Analyses were conducted by 

 using both juvenile abundance and environmental data. 



Seasonal patterns In juvenile fish assemblages 



Many of the species (>68'7r) were present on the shelf in 

 more than one season, which resulted in a large overlap 

 of the seasonal assemblages (ordination not shown), 

 including 12 of the 23 reef-associated species (Table 2). 

 The assemblage data could be explained in three dimen- 

 sions (as determined by CA of the 10% cross-shelf data 

 set [Table 5]) and portrayed seasonal and cross-shelf 

 patterns. Along the first dimension, there was separa- 

 tion between the winter inner-shelf assemblage and all 

 other assemblages. The second and third dimensions 

 portrayed a cross-shelf gradient in the spring, fall, and 

 winter juvenile assemblage data. The summer assem- 

 blage overlapped all the cross-shelf groups from the 

 other seasons except the winter inner-shelf assemblage. 

 High abundances of Larimus fasciatus in the fall and 

 Leiostomus xanthuriis and Mugil cephalus in the winter 

 (Table 2) caused these two seasons to be more dissimilar 

 than were spring and summer, which were very similar. 

 Many of the juveniles were found in the same cross-shelf 

 assemblages during each season, whereas others shifted 

 assemblages (Fig. 4). 



Relation between seasonal juvenile assemblages and 

 environmental variables 



There appeared to be no relationship between the 

 seasonal juvenile assemblages and environmental 



variables. The products of the CCA eigenvalues and 

 species-environment correlations were extremely low 

 (Table 5), and the CCA ordination of seasonal juvenile 

 assemblages and environmental variables was dif- 

 ferent from the CA ordination. Thus, over the entire 

 year, environmental variables did not help explain 

 among-season variation in the juvenile assemblage 

 data. 



Along-shelf and cross-shelf patterns in the inshore 

 juvenile fish data set 



The inshore data set was made up of estuarine, coastal, 

 and open-shelf species (Table 3, Appendix). There was 

 a cross-shelf gradient in the spring and winter (Fig. 6). 

 The innermost station group (8-m) separated from the 

 other station groups along the first dimension (Fig. 7A) 

 in spring. Most taxa from the 8-m assemblage (5 of 

 7) were not collected in high enough abundance to be 

 included in the cross-shelf data set (Tables 2 and 3). 

 In winter there was also a gradient from shallow to 

 deeper stations (Fig. 6C). The assemblage nearest the 

 origin in the winter (8-m assemblage) included estuarine 

 species that were abundant at all the inshore stations 

 (Table 3, Appendix). The coastal species, however, were 

 more abundant at the deeper (12-m, 15-m, and 18-m) 

 station groups (Table 3). There was no pattern in the 

 along-shelf transects, from north to south, during either 

 season (Fig. 6, B and D). 



