Austin et aL: Morphometric separation of annual cohorts within mid-Atlantic Pomatomus saltatnx 



415 



ra 

 o 

 'c 

 o 

 c 

 <B 



o 



6.0- 



4.0- 



a 2.0 



.0 



-2.0- 



-4.0 



-6.0-" 



Bluefish 200-400mm 



/$\ 1987 Centroid 

 [*] 1988-89 Centroids 

 * 1990 Centroid 



/ 



Summer 

 Hatteras, Chesapeake Bay, 

 Long Island 1988-1989 / A 



Summer Chesapeake Bay and 

 ~ ^- Hatteras 1987 

 \ 



\ 



\ 



"V / Summer Chesapeake Bay 



/ sV ^^ — ^ '^^^ 



I Bll \ / 



\ zf-^ ] ( ^ \ 



V 



/ 



Summer Chesapeake Bay 

 1990 



-6.0 -4.0 



— 1 — 

 -2.0 



— I 1 — 



2.0 4.0 



6.0 



Canonical discriminant function 1 



Figure 1 



Scatterplot of canonical di.scriminant functions 1 and 2 for bluefish 200-400 

 mm showing locations of group centroids for 1987-90 by geographic area of 

 collection. 



groups as Wilk (1977) found, we conducted the DFA 

 without a priori assignments. 



Finally, an unweighted paired group method 

 (UPGM) of clustering with arithmetic averages, a 

 form of cluster analysis, was run to examine the rela- 

 tive distance between group centroids. Centroids that 

 appear in a cluster are morphologically similar. 



Results 



The first five canonical functions are listed because 

 they carry the analysis through the 95^f cumulative 

 percentage for fish 200-400 mm (Tables 4 and 5); 

 seven functions are included for the fish >400 mm; 

 however, the eigenvalues (Tables 5 and 6) show that 

 only the first two functions are important in each 

 case. The discriminant scores and centroids from 

 functions 1 and 2 were plotted against each other to 

 develop a graphic representation of the relationship 

 among gi-oups (Figs. 1 and 2). 



Wilk's 1960's ( 1977) analysis found that the inter- 

 dorsal space was a discriminating character. To pro- 

 vide a comparison, we computed an "inter-dorsal 



