Nizinski et al : Separation of Ammodytes amencanus and A dubius in western North Atlantic 



249 



Labrador (Fig. 2). Previous studies reported its occur- 

 rence in Chesapeake Bay (Richards 1982, Norcross 

 et al. 1961); however, no specimens were examined 

 from Chesapeake Bay during this study. 



This species occurs in shallow coastal waters and in 

 protected bays and estuaries. Frequently, specimens 

 were collected with seines or dipnets on sandy beaches 

 in less than 2 m of water. Identification of individuals 

 based only on locality (i.e., inshore vs. offshore) can be 

 made with discretion, but some collections (approx- 

 imately 20%), mainly from inshore stations, contain 

 both species. 



Ammodytes dubius 



Diagnosis Ammodytes dubius shows geographic vari- 

 ation in meristic features (Tables 1-6), and counts for 

 this species are higher than those of A. americanus: 

 lateral plicae 124-147 (i 132.1); vertebrae 68-76 (x 

 70.8) with 69-74 being the most common; dorsal fin 

 rays 56-67 (x 61.8), anal fin rays 28-35 {x 31.1), and 

 pectoral fin rays 12-16 {x 14.0). The majority of indi- 

 viduals examined have >60 dorsal, >29 anal, and 14 

 pectoral rays. More gill rakers (23-31, i 26.6) are 

 present on the first arch than in A. americanus. As 

 described previously, plicae count is the best single 

 distinguishing character; however, vertebrae plotted 

 against plicae separate the species with little overlap 

 (Fig. 6). 



Distribution Ammodytes dubius ranges from North 

 Carolina to Greenland (Fig. 2) and is found in deeper, 

 more offshore waters than A. americanus. Ammodytes 

 dubius is occasionally found inshore but is generally 

 taken in deeper, open waters. This species has a broad 

 bathymetric distribution in coastal waters and ranges 

 in depth from 7 to 80 m. Both species were taken 

 together in approximately 20% of the samples exam- 

 ined, with the majority of these samples being collected 

 at inshore stations or around islands just offshore. 

 Winters and Dalley (1988) reported co-occurrence in- 

 shore oiA. americanus andyl. dubius, particularly in 

 Nevirfoundland waters. Although none of the collections 

 that we examined from Quebec-Nova Scotia contained 

 both species, mixed collections were found in all other 

 geographic areas. Several mixed collections (6 of 11 

 collections) contained only one or two individuals of 

 A. dubius, with the remainder of the lot being 

 A. americanus. Two of the 11 collections were all 

 A. dubius except for one specimen of ^. americanus. 



Greenland Ammodytes Since doubts exist with 

 regard to the number of species and the appropriate 

 names of species occurring in Greenland, these speci- 

 mens were analyzed separately using the PCA equa- 



tion and diagnoses used in analyzing and describing the 

 species in other regions. Only those specimens (A^ = 51) 

 with complete meristic data were included in the 

 analysis. The majority of specimens fit our definition 

 of A. dubius. Vertebral counts ranged from 66 to 75 

 (x 70.8); plicae counts from 124 to 156 (x 133.4); PCA 

 scores from 1388 to 1636. Two specimens, however, 

 fit our definition oi A. americanus; one specimen had 

 70 vertebrae, 123 plicae, and a PCA score of 1374, and 

 the other had 68 vertebrae, 124 plicae, and a score of 

 1372. 



Discussion 



Results of this study demonstrate, in accordance with 

 the majority of previous research (Richards et al. 1963, 

 Leim and Scott 1966, Winters 1970, Scott 1972, Rich- 

 ards 1982), that two species of sand lances occur in the 

 western North Atlantic: an inshore species, A. ameri- 

 canus, with low meristic features, and an offshore 

 species, A. dubius, characterized by high meristics. 



No unequivocal method has been demonstrated 

 previously to consistently identify individuals of 

 A. americamis and A. dubius. Counts purportedly 

 delineating the two species varied between studies, and 

 considerable variation in meristic features for either 

 one or both of these species has been reported. Addi- 

 tionally, earlier studies relied principally on vertebral 

 counts (with supporting data from dorsal and anal ray 

 counts) but this approach was inadequate to accurate- 

 ly identify all individuals. Plicae count, the most useful 

 character in our study, was not used in the majority 

 of previous studies. Furthermore, the lack of published 

 detailed locality data in the majority of previous 

 studies, consistent inshore/offshore designations, and 

 a full understanding of migration patterns (Reay 1970) 

 have added to the confusion in taxonomic status and 

 ability to accurately identify species. 



With no clear-cut definition of the western North 

 Atlantic species, identification problems have hindered 

 previous investigators. A case in point is the reported 

 discrepancy in geographic distribution of A. dubius and 

 resultant interpretations of species. Since the earliest 

 studies by Reinhardt (1838) and DeKay (1842), species 

 designations have not been consistent. Reinhardt 

 (1838) described A. dubius from Greenland. Jordan and 

 Evermann (1896) later reported that this species oc- 

 casionally reached as far south as Cape Cod. In the 

 revision of the Ammodytidae by Duncker and Mohr 

 (1939), A. dubius was reported from throughout the 

 North Atlantic. However, the majority of recent re- 

 searchers disagree and have limited the range of 

 A. dubius to the western North Atlantic. In 1963, 

 Richards et al. reported A. dubius as extending 



