184 



Fishery Bulletin 105(2) 



Discussion 



Forward stepwise DFA of parasite tags has been suc- 

 cessfully employed in describing the stock structure of 

 various flatfish species. Arthur and Albert (1993) used 



Table 2 



Results of forward stepwise discriminant function analy- 

 sis of samples of American plaice iHippoglossoides plates- 

 soides) from the southern Gulf of St. Lawrence (North 

 Atlantic Fisheries Organization Division 4T) showing 

 the relative importance of abundances of two species of 

 acanthocephalan in the classification of individual fish to 

 western and eastern 4T stocks. 



Standardized 

 coefficient of 

 first canonical 

 Variable variable 



df F score 



Eigen- 

 value 



Echinorhynchus 0.862 



a 



Corynosoma 0.499 



strumosum 



230 71.48 0.46 

 229 20.84 



Table 3 



Cross-validation of American plaice iHippoglossoides pla- 

 tessoides) from the western and eastern North Atlantic 

 Fisheries Organization division 4T (southern Gulf of St. 

 Lawrence) determined by discriminant function analysis 

 of abundances of two species of acanthocephalans. 



Location 



Western 4T Eastern 4T 9c Correct 



Western 4T 

 Eastern 4T 

 Totals 



117 

 28 

 145 



20 



67 

 87 



85 

 71 

 79 



Table 4 



Cross-validation of American plaice iHippoglossoides 

 platessoides) from the Smokey Channel and Louisbourg 

 Hole, North Atlantic Fisheries Organization subdivision 

 4Vn, as revealed by discriminant function analyses of the 

 prevalence or abundance of four helminth species (preva- 

 lence of Stephanostomum baccatum, and abundances of 

 Fellodistomum sp., Corynosoma strumosum. and Pseudo- 

 terranova decipiens). 



Location 



Smokey 

 Channel 



Louisbourg 

 Hole 



Correct 



Smokey Channel 

 Louisbourg Hole 

 Totals 



86 



20 



105 



42 

 57 

 99 



67 



74 

 70 



five species of larval helminth to classify widely sepa- 

 rated stocks of Greenland halibut from the Saguenay 

 fiord, St. Lawrence estuary, Labrador coast, and Baffin 

 Island, amd Boje et al. (1997) used six helminth species 

 to gain insight into the stock structure of Greenland 

 halibut from northeastern Newfoundland and Green- 

 land. The overall correct classification rate was >99% 

 in the former, and 77% in the latter study. Blaylock et 

 al. (2003) used eight helminth species to describe the 

 stock structure of Pacific halibut along the Pacific coast 

 of North America from the Bering Sea to northern Cali- 

 fornia and achieved an overall correct classification rate 

 of 83%. In all three studies, accuracy of classification 

 increased when the numbers of geographical categories 

 were reduced by pooling sampling locations. 



Although our survey was confined to a much smaller 

 geographic area than those surveyed in the studies 

 above, the accuracy of classification of plaice stocks 

 from the southern Gulf of St. Lawrence (NAFO division 

 4T) is remarkably high (79%) (Table 3). DFA with only 

 two markers, namely the abundances of the acantho- 

 cephalans E. gadi and C. strumosum, indicated that 

 southern Gulf plaice consist of discrete western and 

 eastern stocks on their summer feeding grounds; this 

 result supports the results of earlier mark-recapture 

 studies (Powles, 1965). Because the acanthocephalans 

 are acquired passively in prey, and 4T plaice feed pri- 

 marily in the Magdalen Shoals, spring through fall, 

 while fasting in winter (Swain et al., 1998), these 

 markers may also be employed in studies of seasonal 

 migrations and mixing of the two stocks. Analysis of 

 infection parameters of four helminth taxa (abundances 

 of C stru!7iosum. the digenean Fellodistomum sp.. the 

 larval anisakine nematode Pseudoterranova decipiens, 

 and prevalence of the larval digenean Stephanostomum 

 baccatum) reveals a somewhat lower degree of discrete- 

 ness (70% correct classification) between plaice from 

 Smokey Channel and Louisbourg Hole, within NAFO 

 subdivision 4Vn. 



Abundances of Fellodistomum sp., P. decipiens, E. 

 gadi, and C. strumosum, and prevalence of S. baccatum 

 are significant factors in a DFA of plaice from all four 

 locations, although the overall accuracy of classification 

 falls to only 48% (Table 5). Although an improvement 

 over random classification (25%), this latter result of- 

 fers little encouragement for the use of parasites tags 

 in future investigations of movements and possible mix- 

 ing of 4T and 4Vn plaice overwintering in Laurentian 

 Channel waters of the 4Vn subdivisioin. DFA of eastern 

 4T and Smokey Channel (4Vn) samples similarly seems 

 to indicate that helminth tags would have little value 

 in studies of the winter composition of plaice stocks in 

 the Cabot Strait area. In this analysis, with the use of 

 E. gadi and the larval nematodes Anisakis simplex and 

 Contracaecum osculatujn as markers, 98% of the 4Vn 

 plaice were classified correctly, but 52% of the plaice 

 from southeastern 4T were misclassified as 4Vn fish. 

 However, given the differences in infection parameters 

 of E. gadi in eastern 4T (prevalence [P] = 50%, abun- 

 dance [A]=1.30) and Smokey Channel samples (P=4%, 



