Journal of Paleolimnology 30: 183–193, 2003. 183  2003 Kluwer Academic Publishers. Printed in the Netherlands. Differential responses of zooplankton populations ( Bosmina longirostris) to fish predation and nutrient-loading in an introduced and a natural sockeye salmon nursery lake on Kodiak Island, Alaska, USA 1,2, 2 * Jon N. Sweetman and Bruce P. Finney 1 Paleoecological Environmental Assessment and Research Laboratory ( P .E. A.R.L.), Department of Biology, 2 Queen’ s University, Kingston, Ontario K7 L 3 N6, Canada; Institute of Marine Science, School of Fisheries * and Ocean Sciences, University of Alaska Fairbanks, Fairbanks, Alaska 99775 -7220, USA; Author for correspondence (e-mail: sweetman@biology.queensu.ca) Received 31 October 2002; accepted in revised form 5 March 2003 Key words: Bosmina longirostris, Cyclops columbianus, Marine-derived nutrients, Oncorhynchus nerka, Preda- tion, Sockeye salmon, Trophic dynamics Abstract Stratigraphic changes in the remains of Bosmina longirostris from a lake with an introduced sockeye salmon population and a lake with a natural salmon run on Kodiak Island demonstrated markedly different responses to past fluctuations in salmon populations. In both lakes, there was a positive correlation between the density of Bosmina microfossils and the abundance of sockeye salmon. However, opposite size trends were observed in the two lakes. In Karluk Lake, which has a native sockeye salmon population, Bosmina mean carapace lengths were largest at high salmon densities, and mean mucro and antennule lengths were also large, suggesting strong predation pressure from cyclopoid copepods, and less intense pressure from juvenile sockeye salmon. As salmon-derived nutrients are important in driving primary productivity in this system, changes in zooplankton productivity track salmon escapement, but grazing pressure on Bosmina from juvenile salmon is less important than that from cyclopoid copepods. In Frazer Lake, a lake with an introduced salmon population, Bosmina morphologies were smallest during periods of high sockeye salmon in the lake, suggesting much stronger predation effects from sockeye salmon due to the suppression of Cyclops columbianus. Latent development of compensatory mechanisms and the delayed recovery of copepod populations to salmon introductions has resulted in zooplankton populations that are still recovering from shifts in fish populations that occurred decades earlier. The differential response of Bosmina populations between the natural and manipulated lakes suggests that care must be taken when attempting to extrapolate results from whole-lake manipulations and short-term experiments to natural systems. Introduction and Goldman 1997). Recent evidence suggests that these studies may be biased, as many ecosystem The role of predation and nutrient loading in fresh- responses may occur on time scales much greater than water ecosystems has received considerable attention the typical length of such experiments (Donald et al. in recent decades (Brett and Goldman 1997). Propo- 2001; Kratz et al. 1987; Pace et al. 1998; Ramcharan nents of top-down based models suggest that zoo- et al. 1995). Consequently, longer-term studies can be plankton, being close to the top of the food web, invaluable for assessing the importance of predation should be more strongly controlled by predation than and nutrient loading on zooplankton dynamics. by nutrient availability (Northcote 1988). However, Historical records of zooplankton communities in evidence for top-down control of zooplankton is most lakes, however, are relatively short, and long- primarily derived from short-term experiments (Brett term responses and recovery from impacts may not be