The marine environment is characterized by few physical barriers, and pelagic fishes commonly show high migratory potential and low, albeit in some cases statistically significant, levels of genetic divergence in neutral genetic marker analyses. However, it is not clear whether low levels of differentiation reflect spatially separated populations experiencing gene flow or shallow population histories coupled with limited random genetic drift in large, demographically isolated populations undergoing independent evolutionary processes. Using information for nine microsatellite loci in a total of 1951 fish, we analyzed genetic differentiation among Atlantic herring from eleven spawning locations distributed along a longitudinal gradient from the North Sea to the Western Baltic. Overall genetic differentiation was low (theta = 0.008) but statistically significant. The area is characterized by a dramatic shift in hydrography from the highly saline and temperature stable North Sea to the brackish Baltic Sea, where temperatures show high annual variation. We used two different methods, a novel computational geometric approach and partial Mantel correlation analysis coupled with detailed environmental information from spawning locations to show that patterns of reproductive isolation covaried with salinity differences among spawning locations, independent of their geographical distance. We show that reproductive isolation can be maintained in marine fish populations exhibiting substantial mixing during larval and adult life stages. Analyses incorporating genetic, spatial, and environmental parameters indicated that isolating mechanisms are associated with the specific salinity conditions on spawning locations.

No abstract

Polymorphisms at five microsatellite DNA loci provide evidence that Atlantic cod Gadus morhua inhabiting Gilbert Bay, Labrador are genetically distinguishable from offshore cod on the north-east Newfoundland shelf and from inshore cod in Trinity Bay, Newfoundland. Antifreeze activity in the blood suggests that Gilbert Bay cod overwinter within the Bay. Gilbert Bay cod are also smaller (weight and length) for their age and consequently less fecund for their age, than cod elsewhere within the northern cod complex. The productivity and recruitment potential of coastal cod off Labrador may thus be much lower than that of offshore northern cod or of inshore cod farther south, implying that a more conservative management strategy may be required for cod from coastal Labrador than traditionally practised for northern cod inhabiting less harsh environments. Relatively high F ST and R ST measures of population structure suggest that important barriers to gene flow exist among five components that include two inshore (Gilbert and Trinity Bay) and three offshore cod aggregations on the north-east Newfoundland Shelf and the Grand Bank. D A and D SW estimates of genetic distance that involve Gilbert Bay cod are approximately three-and 10-fold larger, respectively, than estimates not involving Gilbert Bay cod. The differences between inshore cod from Gilbert Bay and Trinity Bay raise the possibility that other genetically distinguishable coastal populations may exist, or may have existed prior to the northern cod fishery collapse.Harvesting strategies for northern cod should recognize the existence of genetic diversity between inshore and offshore components as well as among coastal components. 2000 The Fisheries Society of the British Isles

Divergent natural selection affecting specific trait combinations that lead to greater efficiency in resource exploitation is believed to be a major mechanism leading to trophic polymorphism and adaptive radiation, We present evidence of trophic polymorphism involving two benthic morphs within Percichthys twhu, a fish endemic to temperate South America. In a series of lakes located in the southern Andes, we found two morphs of R hucha that could be distinguished on the basis of gill raker length and five other morphological measures, most of which are likely associated with the use of food resources. The differences were consistent across all lakes examined, and were correlated with habitat use and diet. Individuals with longer gdl rakers were more abundant in the littoral zone (littoral morph) while the short gill-raker morph was more abundant at 10 m depth and deeper (deep benthic morph). Both morphs fed primarily on benthic invertebrates, but the littoral morph fed more on larval Anisoptera than did the deep benthic morph. Phenotypic correlations among traits were high for the littoral morph, but low and non-significant for the deep-benthic morph. We suggest that gill raker length may influence the relative efficiency of suction feeding for the two morphs. This is the first evidence of trophic polymorphism in fishes from temperate South America. 0 IYY8 The Linnean Society or London ADDITIONAL, KEY WORDS:-trophic morphology -gill raker -competitive divergence -habitat use -endemics -Patagonia.

The collapse of various stock complexes of cod (Gadus morhua) in the northwest Atlantic has prompted a clarification of relationships among stock components. Here we examine the genetic composition of Ͼ2300 cod collected during 1994-1997 in the Gulf of St. Lawrence and its approaches to determine whether: (1) stock components can be genetically identified; (2) population structure is temporally stable; (3) components are always separated and, if not, where and when are they mixed; and (4) component contributions to mixtures can be estimated. We use polymorphism at six microsatellite DNA loci from cod collected on or near their spring and summer spawning grounds to examine structure and then employ maximum likelihood analyses to estimate contributions of each component to mixtures overwintering near the entrance to the Gulf. Estimates of genetic structure (F ST and R ST ) reveal significant differences among cod populations during stockseparated periods, and the structure appears to be temporally stable. Multidimensional scaling analysis of estimates of genetic distance (D A ) suggest that the structure results from differences among cod collected within the Gulf of St. Lawrence and those collected near the entrance to the Gulf on either side of the Laurentian Channel in the Cabot Strait, as well as among cod collected south of Newfoundland along the north side of the Channel. Weak genetic heterogeneity among seven regional mixed-stock collections during the overwintering period suggests that cod aggregations characteristically found in the overwintering region represent population mixtures that differ in the proportion of cod contributed to them by the various stock components. Maximum likelihood estimates indicate no significant temporal changes in component contributions to the mixed-stock samples between 1996 and 1997 when all of the winter mixed-stock samples were pooled. The combined contribution of cod from the southern and northern Gulf of St. Lawrence to the mixedstock samples ranged between 46% and 71% (expected 64%). More precise estimates of contributions from these two regions are precluded by the weak genetic differentiation detected in our samples. The contribution by cod from the Cape Breton Island region was small and estimated at 3%. Contributions by cod from the eastern Scotian Shelf, southwest Newfoundland and south-central Newfoundland were in the range of 13-14%, 4%, and 8%, respectively. Contributions by inshore cod from Placentia and Fortune Bays in south Newfoundland were small to negligible (ϳ3% each). The results indicate that future management could be designed around the spatial and temporal scale of the stock structure identified during the stock-separated period and around the spatially varying contributions to the overwintering mixed-stock fishery.