Commercially available multiple funnel traps have three potential limitations for trapping large wood-boring insects: (i) escape by captured insects from the dry collecting cup, (ii) low catches of insects that fall outside the trap, and (iii) poor visual orientation to the narrow funnel column. To test the importance of these limitations, we compared conventional multiple funnel traps to multiple funnel traps with water-filled collecting cups or large bottom funnels and to crossvane traps with a prominent silhouette. The experiment was conducted in a mill yard in the southern interior of British Columbia between 5 July and 2 October 2000. Differences in catch among different trap types indicated that the first and third of the three potential limitations were important for the capture of most target species. Crossvane traps captured significantly greater numbers of most Cerambycidae and Siricidae, and similar numbers of most Buprestidae, compared with the other traps. Of the two most abundant species, Xylotrechus longitarsus Casey was captured in consistently greater numbers in crossvane than in other traps, but Monochamus scutellatus (Say) showed little discrimination early in the flight season and much higher captures in crossvane traps late in the season. The change in behaviour of M. scutellatus may be related to a transition from maturation feeding to searching for oviposition sites.

1 Traps of four new designs were tested against the conventionally used multiplefunnel trap to determine whether trapping of large wood-boring insects can be improved in western Canada. All four new traps used a large collecting receptacle containing detergent-laced water, and three presented a prominent visual silhouette above the receptacle. 2 In total, 27 336 large woodborers were captured from 10 June to 30 September in an experiment in the southern interior of British Columbia, and 4737 from 6 June to 27 July in an experiment in northern Alberta. The woodborers captured in the British Columbia experiment were mainly beetles in the families Cerambycidae (79%) and Buprestidae (15%), and woodwasps in the family Siricidae (6%). Most woodborers, e.g. three Monochamus spp. and Xylotrechus longitarsus (the predominant cerambycids), were captured throughout the summer, with peak captures in August. 3 Cross-vane, pipe and stacked-bottomless-¯ower-pot traps were generally superior to pan and multiple-funnel traps for insects in nine taxa, but cross-vane traps were the most effective overall, trapping 32% of all insects captured. 4 The large number of target insects captured in a relatively small number of traps in the two experiments suggests that employment of an ef®cacious trap with a large vertical silhouette and a wide, escape-proof collecting receptacle could make mass trapping of large woodborers in timber processing areas operationally feasible. 5 Because the most effective traps were unstable in the wind, and the detergentlaced water captured unacceptably high numbers of small mammals, design mod-i®cations are necessary. We are currently developing a wind-®rm trap, with a prominent vertical silhouette, a wide collecting surface, and an escape-proof, but dry collecting receptacle.

Marine exit timing of sockeye salmon Oncorhynchus nerka populations on the Haida Gwaii Archipelago, British Columbia, Canada, is described, with specific focus on Copper Creek. Marine exit in Copper Creek occurs > 130 days prior to spawning, one of the longest adult freshwater residence periods recorded for any O. nerka population. Copper Creek presents an easy upstream migration, with mild water temperatures (7 to 14°  C), short distance (13·1 km) and low elevation gain (41 m) to the lake where fish hold prior to spawning. An energetic model estimates that <1% of the initial energy reserve is required for upstream migration, compared with 62% for lake holding and 38% for reproductive development. Historical records suggest that it is unlikely that water temperature in any of the O.nerka streams in Haida Gwaii has ever exceeded the presumed temperature threshold (19° C) for early marine exit. Although it is not impossible that the thermal tolerance of Copper Creek O.nerka is very low, the data presented here appear inconsistent with thermal avoidance as an explanation for the early marine exit timing in Copper Creek and in three other populations on the archipelago with early marine exit.

We examined a small population of sockeye salmon (Oncorhynchus nerka) that enters their natal stream, to hold in their natal lake, months (>130 days) prior to spawning. This effectively decouples the influence of spawn timing requirements and behaviours from river entry (alternately referred to as “marine exit”) timing and is therefore a good model to study the migration strategies specifically associated with marine exit. We found individuals with early marine exit had higher growth rates in the months prior to river entry, had greater lipid density, were more likely male, more likely of the 2.2 versus 1.2 age class, had smaller gonads, and (if female) had more and smaller eggs. Body size at river entry did not vary seasonally. These patterns are explained using a life history model proposing that individual fish exit the sea when the marginal fitness benefits of further growth are outweighed by the marginal fitness cost of further marine residence. This point is reached at different times depending on body size, sex, lipid reserves, and the growth rate.