Using data from bottom trawl surveys conducted each September in the southern Gulf of St. Lawrence, Canada, we examined variation in the temperature distribution of Atlantic cod Gadus morhua from 1971 to 1991. Distribution and temperature selection were descnbed using the cumulative distribution functions of temperature and of catch of each age class in relation to temperature. Temperature distribution varied widely among years. For example, the median temperature occupied by Age 5 cod ranged from 0.3 to 7S°C and the 97.5th percenhle of occupied temperatures ranged from 4.6 to 16.6OC. Median occupied temperature decreased as age increased. Temperature selection was highly variable in both extent and direction, varying from cases of no selection to statistically sigmhcant positive or negative selection, depending on year and age class. Cod tended to select warm temperatures relative to those available m 1972 to 1976, 1978 and 1991 and relatively cool temperatures in 1980 to 1990. We used multiple regression to test for an effect of cod abundance on temperature distribution and selection, controlling for variation in available temperatures. When abundance was high, cod tended to occupy colder water, both in absolute terms (the 97.5th, 50.0th and 2.5th percentiles of cod temperature distribution) and in relative terms (the measure of temperature selection). The 9725th and 2.5th percentiles of occupied temperature increased with increasing available temperatures; conversely, temperature selection was inversely related to available temperature. These patterns provide the first field support for the hypothesis that a density-dependent decrease in food availability should result in increased preference for habitats with lower density-independent costs and provide a possible explanation for density-dependent shifts m spatial and bathyrnetnc patterns in this population

We ~nvestigated patterns of distribution and foraging by young-of-the-year pollock Pollachius virens in the rocky intertidal zone. Pollock were sampled by beach seine in fucoid macroalgae and in open habitats at all stages of the tide, day and night throughout the summer. Their presence in shallow water at the high tidal stages indicated that at least part of the pollock population migrated across the full width of the intertidal zone (150 m) each tide. Densities in shallow water were much higher at low than at high tidal stages suggesting that a large influx of pollock moved in from the subtidal zone at low tidal stages and then dispersed into intertidal habitats at high tidal stages. There were few differences in pollock densit~es between algal and open habitats but abundances likely increased in the algal habitat at higher tidal stages when changes in habitat availability are taken Into account. Densities were higher at night and there was an order of magnitude decline in pollock densities from early to late summer. In another study we showed that piscivorous birds are a probable cause of pollock summer mortality. Pollock fed on invertebrates from intertidal algae relatively continuously. The tidal migrations of juvenile pollock observed in this study and their use of macroalgae as a foraging and possibly a refuging habitat strongly suggests that the rocky intertidal zone may be an important fish nursery area.

Fucoid macroalgae in the rocky intertidal zone are a potentially important foraging and refuging habitat for juvenile fishes. A dominant feature of this habitat is that its availability changes with the tides. Vegetated habitat availability changes in many other systems as well, yet little is known about the effects of these changes on animal distributions. We addressed this problem by studying young-of-the-year pollock Pollachius virens using visual transect surveys in the rocky intertidal zone. We examined tidal effects on pollock distribution, their depth and habitat preferences and the relationship between habitat use and patterns of aggregation. Nearly all the pollock population in our study area moved into the intertidal zone and alternated between aggregation in the open and dispersal in the algae. On rising tides, pollock moved from the subtidal zone to the open intertidal zone in large schools then dispersed among available depths and throughout algal habitats in small schools or as solitary flsh When in algae, pollock preferred the dense algal habitat over the sparse algal habitat. On falling t~d e s , pollock schooled In the open habitat In downshore intertidal and subtidal zones. The main seasonal difference, from early to late summer, was an overall distributional shift towards the downshore open habitat and thv subtidal zone and a preference for greater depths in the intertidal zone. These results support the hypothesis that pollock were using both refuging and schooling antipredator tactics during intertidal zone migrations. We suggest that rocky shores are important nurseries for juvenile pollock.