Limnological features of different lakes may limit the density of, or even completely restrict, populations of the European zebra mussel, Dreissena polymorpha. We developed statistical models to predict the occurrence (presence or absence) and density (number per square metre) of Dreissena in lakes, based on multivariate correlations between the density of Dreissena populations and the limnological characteristics of the lakes they inhabit. We found that both occurrence and average density of Dreissena populations were highly correlated with water chemistry. Dreissena was not found in lakes with average pH values below 7.3 and concentrations of calcium ion below 28.3 mg∙L−1. Above these thresholds, Dreissena density was negatively related to concentration of the algal nutrients PO4 and NO3.

Time trends in abundance, body size, species richness, and species composition indicate that crustacean zooplankton communities of southern Canadian Shield lakes changed between 1980 and 2003. Total abundance did not decline despite reductions in total phosphorus, but all other metrics changed. Species richness declined in Harp Lake (Ontario, Canada) following its Bythotrephes invasion, but richness increased in three other lakes. Average cladoceran body length increased from 0.6 to 1.0 mm in seven of the lakes, as larger-bodied taxa replaced smaller ones. Correlations with water quality and fish metrics suggest that cladoceran size increases were attributable to many factors: a decline in food availability following declining phosphorus levels increasing the competitive advantage of larger herbivores, a decline in acidity favouring the larger, acid-sensitive daphniids, and reduced risk of planktivory linked to a rise in dissolved organic carbon levels and changes in predation regimes. Zooplankton communities on the Canadian Shield are changing, and these changes are best viewed in a multiple-stressor context. Key anthropogenic stressors have also changed and may do so again if Ca concentrations continue to decline.

The predatory c1adoceran Bythotrephes longimanus LEYDIG has a natural distribution that extends throughout much of the Palearctic region, and an ever increasing distribution in the Great Lakes basin of North America, where it was first observed in 1982. In this study we define characteristics of 55 waterbodies with and without Bythotrephes in Europe, and use these distributions to predict the species' occurrence in 49 lakes in the Great Lakes basin of North America. Lakes in Europe that supported Bythotrephes were significantly larger, deeper, had higher transparency and lower maximum bottom temperature during summer, and lower total chlorophyll concentration, than those that lacked the species. These patterns also were observed for lakes in North America, although differences between invaded and noninvaded ba sins were significant only for lake area and maximum depth. A discriminant fu nction model correctly predicted Bythotrephes occurrence in 91 % of study lakes in Europe, and was influenced most by Se cchi disk transparency and lake surface area. Applica tion of this model to North America correctly predicted occurrence of Bythotrephes in 82 % (1 8 of 22 ) of lakes in which the species has been recorded. However, the model incorrectly predicted Bythotrephes presence in 74 % (20 of 27 ) of lakes in which the species has not yet been observed. These fi ndings indicate that many of the study lakes in the Great Lakes basin may be vulnerable to invasion by Bythotrephes. Human activ-

Northern temperate lakes often have high water color because of high concentrations of dissolved organic carbon (DOC). Altered light, temperature, and oxygen profiles in these brown-water lakes should reduce the foraging abilities of planktivorous fish and reduce predation on zooplankton and invertebrate predators such as Chaoborus. Additionally, reduced diurnal vertical migration should limit exposure to cold temperatures and increase zooplankton growth rates. We hypothesized that, with increasing water color, Chaoborus would become more important, and this change would be followed by a shift in the zooplankton assemblage toward larger species. To test this hypothesis, we carried out a 2 ϫ 2 ϫ 2 factorial enclosure study to examine the effects of high and low color and the presence and absence of both fish and Chaoborus on zooplankton assemblages. We also analyzed the zooplankton community structure of two lakes with similar morphometry and fish composition but very different water color. Both studies showed that, in highly colored water, Chaoborus was more abundant and the zooplankton community shifted from small species, such as Bosmina and small copepods, to large species, such as Daphnia and Holopedium. Concurrently, the food web structure changed from top-down control to intraguild predation. Because not only the physical habitat differs between clear-and brown-water lakes, but also the predation regimes and food web structure, we conclude that brown-water lakes are a distinct lake type.The humic compounds of DOC are the major factor controlling water color in lakes (Thurman 1985;Pace and Cole 2002). DOC concentrations vary from 0.5 to 20 mg C L Ϫ1 , enough to span the range from crystal clear to darkly teacolored waters. Elevated levels of DOC (Ͼ5 mg L Ϫ1 ) are typical for many northern temperate lakes. A study including more than 1,000 Finnish lakes revealed that Ͼ80% had DOC concentrations Ͼ5 mg L Ϫ1 (Kortelainen 1993). The proportion of lakes with elevated DOC in the northeastern United States and Ontario, Canada, is about 40% (Carter at al. 1980; EMAP 1996).Attenuation of light by DOC could be detrimental to fish planktivory in two ways: low light could directly interfere with visual foraging by planktivorous fishes, and DOC could 1 To whom correspondence should be addressed. Björn Wissel, Coastal Ecology Institute, Louisiana State University, 1243 ECE Building, Baton Rouge, Louisiana 70803 (bwissel@lsu.edu).2 Present address: University of Washington, Joint Institute for the Study of Atmosphere and Oceans, Seattle, Washington 98195.3 Present address: Department of Biology, Laurentian University, Sudbury, Ontario P3E 2C6, Canada. AcknowledgmentsWe thank Joelle Young, Emily Hyfield, Amanda Logue, Shannon Allen, Zoraida Quiñones, Kenyatta Esters, and Michelle Menard for assistance in the field and laboratory. Howard Riessen, Norm Yan, Peter Dillon, and Mark Ridgway were helpful with the experimental design and in locating field sites. We are grateful to the Ontario Ministry of Natural Resources...

Summary 1. Previous studies have suggested that the occurrence of larval Chaoborus in lakes may be affected by fish predation, pH, elevation, temperature, nutrient level, water transparency and interspecific competition, but so far, a detailed statistical evaluation of these findings has not been performed. 2. The aim of this study was to apply regression and ordination techniques to a large data set of 56 lakes in order to test which variables related to lake morphology, water chemistry, and fish predation determine (1) the abundance of individual Chaoborus species and (2) their species composition. 3. Individual Chaoborus species were influenced by very different sets of environmental factors. Nutrient levels positively affected the largest species, Chaoborus americanus, which was restricted to fishless lakes. Abundance of the smallest and most transparent species, C. punctipennis, seemed to be controlled more by the larger Chaoborus species than by fish. Larger chaoborids required low water clarity in order to co‐exist with fish, probably to increase refuge availability. Generally, small lakes (for C. flavicans/C. trivittatus) and shallow lakes (for C. punctipennis) supported higher abundances of Chaoborus.