248 



Fishery Bulletin 88(2). 1990 



Figure 6 



Scattergram of vertebral count plotted against number of lateral 

 plicae for Atnmodytes americanus and A. dubius. 



Plicae and vertebral counts, when plotted against one 

 another, were the most useful combination of char- 

 acters for the entire geographic range producing vir- 

 tually 100% separation (Fig. 6). 



The results of a PC A run on meristic data confirmed 

 the usefulness of these characters. Since factor 1 ac- 

 counted for 91% of the variation and all the meristic 

 characters loaded heavily on this factor, factor 1 was 

 the only factor used. Plicae had the highest component 

 loading score (8.09) followed by dorsal rays (2.44), 

 vertebrae (2.31), gill rakers (1.23), anal rays (1.03), and 

 pectoral rays (0.45). 



The histogram of component scores (Fig. 4) produced 

 a bimodal distribution with little overlap above or below 

 a score of 1378 for the first set of individuals evalu- 



ated. When the scores were compared with the earlier 

 tentative identifications (based only on meristics), only 

 10 (5 of each species) were either borderline or placed 

 into the wrong species grouping. These 10 individuals 

 were then reexamined and either the proposed iden- 

 tification was verified (3 cases) or errors in the counts 

 of certain meristic characters were found (7 cases). 



At most, 2% of the specimens were misclassified using 

 the component scores. Of 54 additional specimens tested 

 against the ec|uation, none were misclassified. This 

 methodology, therefore, was consistent and reliable. 



Based on these results, diagnoses for the two species 

 follow: 



Ammodytes americanus 



Diagnosis Ammodytes ajnericnmis tends to have 

 lower counts than A. dubius: lateral plicae 106-126 

 (X 117.4); vertebrae 62-70, usually 64-69 {x 66.4); 

 dorsal fin rays 52-61 {x 57.4), anal fin rays 26-33 (x 

 29.4), and pectoral fin rays 11-15 (x 13.2), with the 

 most common dorsal, anal, and pectoral ray counts 

 being 55-59, 27-31, and 13 rays, respectively; gill 

 rakers on the first arch 21-28 (x 24.3). 



Six meristic characters in combination provide a good 

 separation of this species from A. duhius. Number of 

 plicae, however, is the single best character for distin- 

 guishing >!. (nrieriranus from A. dubius in the western 

 North Atlantic. Vertebrae plotted against plicae separ- 

 ate individuals of the two species with little overlap 

 (Fig. 6). 



Distribution Based on our samples, A. ampriranus 

 ranges coastally from southern Delaware north through 



