Summary1. While density-dependent mortality and emigration have been widely reported in stream salmonid populations, density-dependent growth is less frequently detected. A recent study suggests that density-dependent growth in stream salmonids occurs at low densities, whereas density-dependent mortality and emigration occur at high densities. 2. To test the hypothesis that density-dependent growth occurs primarily at low rather than at high densities, we examined the relationship between average fork length and population density of young-of-the-year (YOY) Atlantic salmon at the end of the growing season using a 10-year data set collected on Catamaran Brook, New Brunswick. We tested whether (1) average body size decreases with increasing density; (2) the effect of density on average body size is greatest at low densities; (3) growth rate will decrease most rapidly at low effective densities [ Σ (fork length) 2 ]; (4) density-dependent growth is weaker over space than over time; and (5) the strength of density-dependent growth increases with the size of the habitat unit (i.e. spatial scale) when compared within years, but not between years.3. There was a strong negative relationship between the average body size and population density of YOY Atlantic salmon in the autumn, which was best described by a negative power curve. Similarly, a negative power curve provided the best fit to the relationship between average body size and effective density. Most of the variation in average body size was explained by YOY density, with year, location and the density of 1+ and 2+ salmon accounting for a minor proportion of the variation. 4. The strength of density-dependent growth did not differ significantly between comparisons over space vs. time. Consistent with the last prediction, the strength of densitydependent growth increased with increasing spatial scale in the within-year, but not in the between-year comparisons. 5. The effect of density on growth was strongest at low population densities, too low to expect interference competition. Stream salmonid populations may be regulated by two mechanisms: density-dependent growth via exploitative competition at low densities, perhaps mediated by predator-induced reductions in drift rate, and density-dependent mortality and emigration via interference competition at high densities.

Literature data for 19 populations from 16 different studies of six species of stream-dwelling salmonids were analysed to test the hypothesis that density-dependent growth is stronger at low rather than at high population density. Fifteen of 19 populations showed evidence of a significant decrease in growth rate with increasing density. In 11 of these 15 populations, the pattern of density-dependent growth was better described by a negative power curve than by a linear regression (i.e. Akaike Information Criterion, AIC linear À AIC power > 2), whereas only one population was better described by a linear regression than by a negative power curve; three populations were adequately described by both models (AIC < 2). In 10 of the 11 populations that were best described by a negative power curve, most of the decrease in growth rate occurred at population densities <1 fish m À2, when space limitation is unlikely. This analysis provides broad support for the hypothesis that density-dependent growth in stream salmonids occurs primarily at low population densities, probably due to exploitative competition. # 2005 The Fisheries Society of the British Isles

In the field, juvenile brook charr Salvelinus fontinalis inhabiting high-velocity water were found to have larger caudal fins and more slender bodies than those inhabiting low-velocity water. Young-of-the-year S. fontinalis were reared in either a high-or low-velocity treatment for 16 weeks and their morphology was measured bi-weekly. From the second to fourth weeks of the experiment onwards, fish reared in the high-velocity treatment had larger maximum caudal fin heights and deeper caudal peduncles than fish reared in the low-velocity treatment. This study demonstrated that the morphological variation in caudal area exhibited by wild juvenile brook charr from microhabitats differing in water velocity could be a consequence of phenotypic plasticity in response to hydrological conditions.

Visibility is thought to affect the territory size of visually oriented animals but there have been few experimental tests of the hypothesis. We re-examined the relationship between visibility and territory size in juvenile salmonids to test the hypothesis that increasing habitat heterogeneity results in a reduction in territory size and consequently in higher population densities. Equal densities of young-of-the-year rainbow trout (Oncorhynchus mykiss) were stocked in two experimental treatments with low visibility and a control treatment with high visibility. Visibility was decreased by placing large stones or plywood dividers onto the substrate of experimental stream channels. As predicted, the size of individual territories decreased with decreasing visibility of the habitat. However, the treatments did not differ significantly in population density or growth rate of the fish. While this study confirms the inverse relationship between habitat visibility and territory size, the decrease in territory size did not produce an increase in population density of juvenile salmonids.Résumé : La visibilité affecte, croit-on, la taille du territoire chez les animaux qui s'orientent par la vue, bien qu'il y ait peu de vérifications expérimentales de cette hypothèse. Nous avons examiné de nouveau la relation qui existe entre la visibilité et la taille du territoire chez des jeunes salmonidés dans le but d'éprouver l'hypothèse selon laquelle une hétérogénéité accrue de l'habitat réduit la taille du territoire et, par conséquent, permet des densités de population plus élevées. Nous avons établi des densités égales de jeunes truites arc-en-ciel (Oncorhynchus mykiss) de l'année dans deux sites expérimentaux à visibilité réduite et un site témoin à visibilité élevée. De grosses pierres ou des panneaux de contre-plaqué placés sur le substrat des ruisseaux artificiels ont servi à diminuer la visibilité. Tel que prédit, la taille des territoires individuels diminue lorsque la visibilité dans l'habitat est réduite. Cependant, il n'en résulte pas de différences dans la densité de population, ni dans le taux de croissance des poissons. Bien que notre étude confirme l'existence d'une relation inverse entre la visibilité dans l'habitat et la taille du territoire, la réduction de la taille du territoire n'entraîne pas d'augmentation de la densité de population chez les jeunes salmonidés.[Traduit par la Rédaction] Imre et al. 309

Using semi-natural enclosures, this study investigated (1) whether adult sea lamprey Petromyzon marinus show avoidance of damage-released conspecific cues, damage-released heterospecific cues and predator cues and (2) whether this is a general response to injured heterospecific fishes or a specific response to injured P. marinus. Ten replicate groups of 10 adult P. marinus, separated by sex, were exposed to one of the following nine stimuli: deionized water (control), extracts prepared from adult P. marinus, decayed adult P. marinus (conspecific stimuli), sympatric white sucker Catostomus commersonii, Amazon sailfin catfish Pterygoplichthys pardalis (heterospecific stimuli), 2-phenylethylamine (PEA HCl) solution, northern water snake Nerodia sipedon washing, human saliva (predator cues) and an adult P. marinus extract and human saliva combination (a damage-released conspecific cue and a predator cue). Adult P. marinus showed a significant avoidance response to the adult P. marinus extract as well as to C. commersonii, human saliva, PEA and the adult P. marinus extract and human saliva combination. For mobile P. marinus, the N. sipedon washing induced behaviour consistent with predator inspection. Exposure to the P. pardalis extract did not induce a significant avoidance response during the stimulus release period. Mobile adult female P. marinus showed a stronger avoidance behaviour than mobile adult male P. marinus in response to the adult P. marinus extract and the adult P. marinus extract and human saliva combination. The findings support the continued investigation of natural damage-released alarm cue and predator-based repellents for the behavioural manipulation of P. marinus populations in the Laurentian Great Lakes.