266 



Fishery Bulletin 104(2) 



fluenced by the single station 7 (Fig. 3E), which had 

 a high abundance of Monacanihiis hispidus. Bothus 

 ocellatiis and B. lunatus were also collected only at 

 this station (Table 2). The inner-shelf consisted of 

 one dominant species (Fig. 4C), and the mid-shelf had 

 the greatest number of species (Fig. 4C, Table 2). The 

 outer-shelf assemblage consisted of estuarine, coastal, 

 and open-shelf species (Fig. 4C). Six reef-associated 

 species were found in the mid- and outer-shelf assem- 

 blages. The distributions of species in all three winter 

 assemblages indicated cross-shelf mixing. An open- 

 shelf species, Ogcocephalus nasutus, was present in 

 the inner-shelf assemblage and Mugil cephalus, which 

 uses estuaries as adults, was present in the outer-shelf 

 assemblage (Table 2, Appendix). Finally, two slope 

 species (Argentina striata and Antennarius radiosus) 

 were present in the mid- and outer-shelf assemblages 

 (Table 2, Appendix). 



Cross-shelf patterns in juvenile fish assemblages and 

 environmental variables 



The juvenile fish assemblages and environmental vari- 

 ables all showed a cross-shelf gradient, which was cor- 

 related with depth during three of the four seasons 

 (Table 6). The relationship between the other environ- 

 mental variables and assemblages varied with season. 

 In spring, the gradient in the inner- and mid-shelf 

 and outer-shelf was mainly along the first dimension 

 (Table 5) and was influenced by depth, bottom salin- 

 ity, and bottom density (Fig. 5A). Inner- and mid-shelf 

 stations were shallower and had less saline and dense 

 bottom water than outer-shelf stations (Fig. 2). The 

 spread along the second dimension in the outer-shelf 

 group was less important (Table 5) and was influenced 

 by the higher stratification of the outer stations (Fig. 2). 

 A Monte Carlo permutation test of the environmental 

 variables for spring indicated that depth, stratification, 

 and bottom density were significantly related to the 

 juvenile assemblages and the grouping of stations in 

 the CCA (Table 6). 



The summer juvenile assemblage was not related to 

 these environmental variables. The ordination of the 

 station and species data changed with the inclusion of 

 environmental data to the analyses, indicating a poor 

 relationship between the juvenile data and the environ- 

 ment (Fig. 5B) The low eigenvalues also indicated that 

 very little of the observed variation in the data was 

 explained by the environment (Table 5). These results 

 agree with the results from the CA, both of which indi- 

 cate that the entire shelf is a single assemblage. 



The fall cross-shelf patterns in the juvenile fish as- 

 semblages were related to depth and bottom salinity 

 (Fig. 5C). The first dimension had the highest eigen- 

 value and species-environment correlation (Table 5) 



