Responses of phytoplankton and epilithon during acidification and early recovery of a lake

1. Lake 302S in the Experimental Lakes Area of Canada was acidified from pH 6.7 (1981) to 5.1 (1986). The pH was further reduced to 4.5 in 1987 and held at that level until 1991. From 1992 to 1995, the pH was allowed to increase to a target value of 5.8.
2. The response of the phytoplankton community to decreasing pH from 6.0 to 5.1 was similar to that observed in another experimentally acidified lake (223) and in other atmospherically acidified lakes. Acidification affected species diversity of both the phytoplankton and epilithon. Phytoplankton diversity was positively correlated with pH. Epilithic algal diversity was more variable and did not correlate with pH.
3. Phytoplankton biomass was enhanced by acidification as the assemblage shifted from a dominance of chrysophytes to large dinoflagellates ( Gymnodinium sp. and Peridinium inconspicuum ). Epilithon biomass was unaffected, but dominance shifted from filamentous cyanophytes ( Lyngbya ) to acidophilic diatoms ( Tabellaria quadriseptata and Anomoeonis brachysira ).
4. The only taxon to be similarly affected in both the phytoplankton and epilithon was the cyanobacteria, being significantly reduced below pH 5.1. During early recovery (pH 5.5–5.8), cyanobacteria increased and species present prior to acidification recolonized both habitats.
5. In the early stages of recovery, planktonic and benthic assemblages remained more similar to acidified than natural assemblages, but more profound change began at pH > 5.5.     

Cascading Trophic Interactions in an Oligotrophic Species-poor Alpine Lake

Non-native brook trout (Salvelinus fontinalis) were eradicated from alpine Bighorn Lake, Alberta, Canada, to test whether strong cascading trophic interactions (CTI) can occur in oligotrophic, high seston C:P, species-poor lakes. Fishless alpine Pipit Lake was used as a reference ecosystem. Bighorn Lake zooplankton biomass increased from 0.14:1 relative to Pipit Lake before fish removal began in 1997 to 0.6:1 afterwards due to an increase in the abundance of adult cyclopoid copepods beginning in 1997 and the reappearance of Daphnia middendorffiana in 1998. Following the reappearance of Daphnia, Bighorn Lake total phytoplankton biomass fell from 64:1 relative to Pipit Lake to 0.9:1. Over the same periods Bighorn Lake:Pipit Lake chlorophyll-a ratios declined from 2.4:1 to 1.6:1, although the decrease was not statistically significant. Mid-summer Secchi disc depth in Bighorn Lake increased from 3.1 m before manipulation to 9.2 m, the maximum depth of the lake, in 2001 and 2002. Increased transparency was most likely due to increased filtration of suspended inorganic particles from the water column by higher abundances of large zooplankton. Post-manipulation increases in dissolved inorganic nitrogen (DIN), DIN:total dissolved phosphorus (TDP) ratio and declines in TDP in Bighorn Lake were not attributable to ecosystem manipulation, similar changes were observed in reference Pipit Lake. We conclude that strong pelagic CTI, expressed as change in total phytoplankton biomass and largely mediated by Daphnia, can occur in oligotrophic, high seston C:P, species-poor ecosystems. However, strong CTI responses in phytoplankton biomass may lag trophic manipulation by several years.     

Effects of different fish species and biomass on plankton interactions in a shallow lake

Thirty-six enclosures, surface area 4 m², were placed in Little Mere, a shallow fertile lake in Cheshire, U.K. The effects of different fish species (common carp, common bream, tench and roach) of zooplanktivorous size, and their biomass (0, 200 and 700 kg ha^–1) on water chemistry, zooplankton and phytoplankton communities were investigated. Fish biomass had a strong effect on mean zooplankton size and abundance. When fish biomass rose, larger zooplankters were replaced by more numerous smaller zooplankters. Consequently phytoplankton abundance rose in the presence of higher densities of zooplanktivorous fish, as zooplankton grazing was reduced. Fish species were also significant in determining zooplankton community size structure. In enclosures with bream there were significantly greater densities of small zooplankters than in enclosures stocked with either carp, tench and, in part, roach. When carp or roach were present, the phytoplankton had a greater abundance of Cyanophyta than when bream or tench were present. Whilst top-down effects of fish predation controlled the size partitioning of the zooplankton community, this, in turn apparently controlled the bottom-up regeneration of nutrients for the phytoplankton community. At the zooplankton–phytoplankton interface, both top-down and bottom-up processes were entwined in a reciprocal feedback mechanism with the extent and direction of that relationship altered by changes in fish species. This has consequences for the way that top-down and bottom-up processes are generalised.     

Two mesocosm experiments investigating the control of summer phytoplankton growth in a small shallow lake

1. Mesocosm experiments were carried out to examine the relative importance of top down (fish predation) and bottom up (nutrient addition) controls on phytoplankton abundance in a small shallow lake, Little Mere, U.K., in 1998 and 1999. These experiments were part of a series at six sites across Europe. 2. In the 1998 experiment, top‐down processes (through grazing of large Cladocera) were important in determining phytoplankton biomass. The lack of plant refugia for zooplankton was probably important in causing an increasing chlorophyll a concentration even at intermediate fish density. Little Mere normally has abundant macrophytes but they failed to develop substantially during both years. Bottom‐up control was not important in 1998, most probably because of high background nutrient concentrations, as a result of nutrient release from the sediments. 3. In 1999 neither top‐down nor bottom‐up processes were significant in determining phytoplankton biomass. Large cladoceran grazers were absent even in the fish‐free enclosures, probably because dominance of cyanobacteria and high phytoplankton biomass made feeding conditions unsuitable. As in 1998, bottom‐up control of phytoplankton was not important, owing to background nutrient concentrations that were even higher in 1999 than in 1998, perhaps because of the warmer, sunnier weather. 4. The differing outcomes of the two experiments in the same lake with similar experimental designs highlight the importance of starting conditions. These conditions in turn depended on overall weather conditions prior to the experiments.     

Effect of young-of-the-year walleye (Percidae: Stizostedion vitreum) on plankton dynamics and water quality in ponds

To understand the impact of young-of-the-year (YOY) fish on food web dynamics and water quality, we stocked larval walleye (9 mm TL) (Stizostedion vitreum) in six experimental ponds using two fish densities (10 and 50 fish m^–3) with three replicates. At high fish density, the average abundances of cladocerans and copepods and the Secchi depth were lower whereas abundances of rotifers and algae, gross primary productivity (GPP), pH and total phosphorus concentration were higher than at low fish density. Fish impact on bacterial abundance, dissolved oxygen, nitrogen and phosphorus concentrations, however, was not significant. The within treatment measurements of all variables except GPP were significantly different over time. Our results indicate that YOY walleye predation at high density can affect plankton community by reducing large zooplankton biomass and water clarity, and increasing phytoplankton abundance. The impact of YOY piscivorous fish on plankton should be considered when biomanipulation is applied for improvement of water quality.     

Predator‐induced changes in dissolved organic carbon dynamics

The fate of dissolved organic carbon (DOC) is partly determined by its availability to microbial degradation. Organisms at upper trophic levels could influence the bioavailability of DOC via cascading effects on primary producers and bacteria. Here we experimentally tested whether the presence of fish in aquatic food webs can indirectly affect the composition of the DOC pool. We found that fish had strong positive effects on phytoplankton biomass that affected the dynamics of DOC composition. Specifically, fish increased protein‐like, algae‐derived DOC mid‐experiment, concurrent with the strongest fish‐induced increase in phytoplankton biomass. Fish also increased bacterial abundance, altered the community composition and diversity of bacteria, and temporarily increased DOC compounds with fluorescence properties indicative of microbially‐reprocessed organic matter. Overall, our experiment revealed that fish can positively influence the substrate (algae‐produced DOC) and the key players (bacteria) of the microbial carbon pump. Consequently, fish could contribute to carbon sequestration by stimulating both the production of bioavailable DOC and the microbial degradation of bioavailable to persistent DOC. We propose this as a novel mechanism whereby the loss of predators from global ecosystems could alter carbon cycling.     

Planktonic food web in marine mesocosms in the Eastern Mediterranean: bottom-up or top-down regulation?

A mesocosm experiment was conducted in order to studythe structure of the planktonic food web. The dynamicsof pico-, nano- and microplankton populations werefollowed during 40 days in four large (40 m³)enclosures. In three tanks a gradient of addednutrients (nitrogen and phosphorus) was applied, whilea fourth tank was used as a control. On day 14, thetop predator (sea bream Sparus aurata larvae)was introduced into the tanks and part of the watercolumn in each tank was isolated in a plastic bagwithout fish larvae, to act as a control forpredation. Physical parameters, chlorophyll aand nutrient concentrations, as well as planktonconcentrations were monitored. A diatom bloom wasobserved in all four tanks, in the first phase endingwith silicate depletion. Flagellate and dinoflagellateabundance subsequently increased, these organismsbeing limited by zooplankton grazing. The zooplanktonpopulations were controlled by both resources (mostlyflagellates) and predation (by fish larvae) asindicated by the results of the control experiments. This revised version was published online in July 2006 with corrections to the Cover Date.     

Food web manipulation in Lake Zwemlust: Positive and negative effects during the first two years

The hypertrophic Lake Zwemlust, a small water body used as a swimming pool, was characterized by algal blooms in summer, reducing the Secchi disk transparency to less than 0.3 m. Since in The Netherlands a Secchi disk transparency of 1 m is obligatory for swimming waters, corrective measures were called for to improve the light climate of the lake. In March, 1987, as an experiment, the lake was drained by pumping out the water to facilitate fish elimination. Planktivorous and benthivorous fish species, which were predominant, were removed by seine- and electro-fishing. After the lake had refilled by seepage it was restocked by a new simple fish community comprising pike (Esox lucius) and rudd (Scardinius erythrophthalmus) only. Stacks of willow twigs (Salix) and macrophytes (roots ofNuphar lutea and seedlings ofChara globularis) were introduced into the lake as spawning grounds and refuges for the pike against cannibalism and as shelter for the zooplankton. The effects of this food web manipulation on the light climate, phytoplankton, zooplankton, fish, macrophytes, macrofauna and on the nutrient concentrations were monitored during 1987 and 1988. In summer 1987, despite of high nutrient concentrations, the phytoplankton density was low, due to control by zooplankton, causing a Secchi disk transparency of 2.5 m, the maximum depth. Chlorophyll-a concentrations were low (<5 g Chl.l^–1), blooms of cyanobacteria did not occur and a shift from rotifers to cladocerans took place. In 1988, however, also some negative effects were noticed. Macrophytes and filamentous green algae reached a much higher biomass (50–60% cover of the lake bottom) than in 1987; some species, growing through the entire water column, interfered with the lake's recreational use. Associated with the macro-vegetation and possibly with the absence of larger cyprinids, the diet of which also comprises snails, a large scale development of the snail population, among themLymnaea peregra var.ovata took place. This species is known to act as an intermediate host of the bird-parasitizing trematodeTrichobilharzia ocellata, the cercariae of which cause an itching sensation at the spot of penetration of the human skin, accompanied by rash (schistosome dermatitis or swimmers' itch); in July, 1988, about 40% of the bathers complained about this itching. A positive effect of the macrophytes and filamentous green algae was the high uptake of nitrogen, resulting in a low nitrogen concentration in the lake and growth limitation of the phytoplankton population by nitrogen in the summer of 1988. In 1988 the cladocerans were abundant in April only; and unlike in 1987, in the summer of 1988 there was a shift from cladocerans to rotifers. Therefore, only in early spring (April) zooplankton grazing controlled phytoplankton growth and in summer nitrogen limitation was the major controlling factor, keeping chlorophyll-a concentrations low.     

Responses to fish predation and nutrients by plankton at different levels of taxonomic resolution

1. To improve mechanistic understanding of plankton responses to eutrophication, a mesocosm experiment was performed in the shallow littoral zone of a south Swedish lake, in which nutrient and fish gradients were crossed in a fully factorial design. 2. Food chain theory accurately predicted total biomass development of both phyto‐ and zooplankton. However, separating zooplankton and algae into finer taxonomic groups revealed a variety of responses to both nutrient and fish gradients. 3. That both nutrients and fish are important for phytoplankton dynamics was seen more clearly when viewing each algal group separately, than drawing conclusions only from broad system variables such as chlorophyll a concentration or total phytoplankton biovolume. 4. In some taxa, physiological constraints (e.g. sensitivity to high pH and low concentrations of free CO₂) and differences in competitive ability may be more important for the biomass development than fish predation, grazing by herbivorous zooplankton, and nutrient availability. 5. We conclude that food chain theory accurately predicted responses in system variables, such as total zooplankton or algal biomass, which are shaped by the dynamics of certain strong interactors (‘keystone species’), such as large cladocerans, cyanobacteria and edible algae (<50 μm), whereas responses at finer taxonomic levels cannot be predicted from current theory.     

Invasion of a Top Predator into an Epipelagic Ecosystem can bring a Paradoxical Top-Down Trophic Control

We apply mathematical modeling to explore different scenarios of invasion of a top predator (carnivorous zooplankton or planktivorous fish) into an epipelagic plankton ecosystem. We use a ‘minimal’ model of three nonlinear ordinary differential equations (nutrient–phytoplankton–herbivores) with the top predator density as a time-dependent parameter. The ecosystem shows different types of response, which can be described in terms of top-down trophic control. Our investigation indicates that under certain conditions the plankton ecosystem model demonstrates a surprising kind of response: in a wide range of realistic ecosystem parameters the invasion of the top predator leads to a prominent increase in the average density of zooplankton and to a resulting decrease of phytoplankton density. This phenomenon is opposite to the ‘typical’ top–down control when the carnivore pressure decreases zooplankton density which, in turn, increases phytoplankton biomass. We call the revealed type of top-down control ‘paradoxical’. Examples of such a response in natural aquatic ecosystems were reported earlier but no clear explanation has been provided hitherto. In this paper, we analyze possible mechanisms of ‘paradoxical top–down control’ and show that it can occur in eutrophic epipelagic ecosystems subject to high rate of cross-pycnocline exchange.     

The planktonic food web structure of a temperate zone estuary,and its alteration due to eutrophication

Hydrobiologia - Current conventional wisdom argues that human-induced excesses in nutrient loadings to estuaries often stimulate ‘excess’ algal production leading to hypoxia, via...     

Cascading Trophic Interactions from Fish to Bacteria and Nutrients after Reduced Sewage Loading: An 18-Year Study of a Shallow Hypertrophic Lake

The effects of major reductions in organic matter, total phosphorus (TP), and total nitrogen (TN) loading on the chemical environment, trophic structure, and dynamics of the hypertrophic, shallow Lake Søbygård were followed for 18 years. After the reduction in organic matter loading in 1976, the lake initially shifted from a summer clear-water state, most likely reflecting high grazing pressure by large Daphniaspecies, to a turbid state with extremely high summer mean chlorophyll a (up to 1400 μg L^{? 1}), high pH (up to 10.2), and low zooplankton grazing. Subsequently, a more variable state with periodically high grazing rates on phytoplankton and bacteria was established. Changes in zooplankton abundance and grazing could be attributed to variations in cyprinid abundance due to a fish kill (probably as a consequence of oxygen depletion) and pH-induced variations in fish recruitment and fry survival. The results suggest strong cascading effects of fish on the abundance and size of zooplankton and phytoplankton and on phytoplankton production. A comparatively weak cascading effect on ciliates and bacterioplankton is suggested. Due to high internal loading, only minor changes were observed in lake-water TP after a reduction in external TP loading of approximately 80% in 1982; net retention of TP was still negative 13 years after the loading reduction, despite a short hydraulic retention time of a few weeks. TN, however, decreased proportionally to the TN-loading reduction in 1987, suggesting a fast N equilibration. Only minor improvement in the environmental state of the lake has been observed. We suggest that another decade will be required before the lake is in equilibrium with present external P loading.     

Long-term Annual and Seasonal Changes of Meta- and Protozooplankton in Lake Müggelsee (Berlin): Effects of Eutrophication,Grazing Activities,and the Impact of Predation

Annual changes of rotifers, copepods, cladocerans, the ciliate Epistylis rotans, and larvae of Dreissena polymorpha were analysed for the period 1908–1990. Though food resources increased 6–10 fold in the course of eutrophication, only rotifers and Epistylis increased accordingly. Probably as a result of increased predation pressure crustaceans increased only twice. The seasonal pattern of metazoans and protozoans (flagellates, sarcodines, ciliates) were analysed for 12 and 3 years, resp. During winter and spring, large heterotrophic flagellates and ciliates dominated the zooplankton and were responsible for a pronounced - formerly underestimated - grazing pressure on phytoplankton. In early summer, metazoan filter-feeders were often able to cause a significant reduction of phyto- and protozooplankton. However, during some years, phytoplankton declined in the absence of a pronounced grazing pressure. Field data and experiments revealed that predators were able to regulate the density of cladocerans in early summer (mainly cyclopoids) and summer (mainly Leptodora, smelt and fish juveniles).     

Host length,age, diet,parasites and stable isotopes as predictors of yellow perch (<Emphasis Type="Italic">Perca flavescens</Emphasis>Mitchill), trophic status in nutrient poor Canadian Shield lakes

We used data from parasites and stable isotopes of yellow perch, Perca flavescens, to determine trophic status in four small Canadian Shield lakes as parasites allow the identification of both prey and non-prey dietary components in the host's community. Stable C isotope ratios for all perch ranged from –34 to –19 while stable N isotope ratios ranged from 4.5 to 12.5. These ranges are larger than those observed in many other fish species. Perch age was the most significant predictor of stable C isotope ratio and perch parasite fauna was the most significant predictor of stable N ratios. In particular, parasite fauna indicative of zooplanktivorous or piscivorous perch were most accurate for predicting fish trophic position and thus stable isotope ratio. Fish length and age showed fewer significant relationships with isotope ratios than parasite infracommunity or diet and suggests that trophic category for perch cannot always be predicted based on size.     

Fatty acid composition of consumers in boreal lakes – variation across species,space and time

1. Fatty acids (FAs) have been widely applied as trophic biomarkers in aquatic food web studies. However, current knowledge of inter‐ and intraspecific variation in consumer FA compositions across spatial and temporal scales is constrained to a few pelagic taxa. 2. We analysed the FAs of 22 taxa of benthic macroinvertebrates, zooplankton and fish collected from the littoral, pelagic and profundal habitats of nine boreal oligotrophic lakes over spring, summer and autumn. We quantified and compared the FA variance partitions contributed by species identity (i.e. an integrative effect of phylogenetic origin, life history and functional feeding guild of individual taxa), site and season using partial redundancy analysis both on all consumers and on benthic arthropods alone. 3. Species identity alone contributed 84.4 and 72.8% of explained FA variation of all consumers and benthic arthropods, respectively. Influences of site, season and all joint effects accounted for 0–11.3% only. Fatty acid profiles of primary consumers differentiated below class level, but those of predators were distinguishable only when they became more taxonomically distinct (i.e. among classes or higher). 4. Pelagic and profundal consumers showed stronger reliance on autochthonous resources than did their littoral counterparts as reflected by their higher ω3 to ω6 FA ratios. Polyunsaturated FAs (PUFAs) were increasingly retained with trophic levels, and saturated FAs (e.g. FA 16 : 0) gradually reduced. Ecologically, this trade‐off enhances the trophic transfer efficiency and confirms the importance of PUFA‐rich autotrophs in aquatic food webs. 5. Our findings indicate strong interspecific differences in FA requirements and assimilation among aquatic consumers from a wide range of taxonomic levels, habitats and lakes. Consumers were able to maintain homoeostasis in FA compositions across spatial and temporal changes in resource FAs, but consumer homoeostasis did not limit the effectiveness of FAs as trophic biomarkers.     

Factors influencing Gonyostomum semen blooms in a small boreal reservoir lake

Blooms of the nuisance alga Gonyostomum semen occurred in Lake 979 (Experimental Lakes Area), a small brown-water lake, that was subjected to several years of an experimental flooding regime. During periods of flooding, blooms of G. semen developed when light decreased below 100molm^–2s^–1 and total phosphorous concentrations increased to >30gl^–1. Gonyostomum semen biomass was significantly correlated with total P and DOC concentrations. In addition, G. semen abundance increased at times when Daphnia rosea had rapidly declined to<10 animals l^–1. Daphnia egg ratios suggest that declines in Daphnia abundance were the result of shifts in mortality and not causally linked to changes in G. semen densities. The results observed in Lake 979 were supported by a laboratory experiment where the appearance of G. semen from Lake 979 sediment was stimulated by altering chemical and biological variables. The stimulation of G. semen blooms appears to be dependant on multiple variables rather than a single variable.     

Plankton dynamics in a tropical floodplain lake: fish, nutrients, and the relative importance of bottom-up and top-down control

1. Two enclosure experiments were carried out in Laguna Bufeos, a neotropical várzea lake located in the floodplain of River Ichilo (Bolivia). The experiments aimed (i) to assess the relative importance of bottom‐up and top‐down control on the plankton community, (ii) to assess the relative impact of direct and indirect effects of planktivorous fish on the zooplankton, and (iii) to attempt to identify the mechanisms responsible for these effects. 2. During the first experiment, bottom‐up control seemed to dominate the planktonic food web. Compared with fishless enclosures, oxygen concentrations, chlorophyll a levels and the population densities of all cladoceran zooplankton taxa increased in enclosures with fish. Birth rates of Moina minuta, the dominant taxon, were substantially higher in the presence than in the absence of fish, whereas death rates did not differ between treatments. These results are the first to suggest that the positive effects of fish on crustacean zooplankton via effects on nutrient cycling and the enhancement of primary production can compensate for losses because of fish‐related mortality. 3. During the second experiment, the direction of control appeared to vary between trophic levels: the phytoplankton appeared to be bottom‐up controlled whereas the zooplankton was mainly top‐down controlled. Chlorophyll a concentrations were enhanced by both fish and nutrient additions. The majority of the zooplankton taxa were reduced by the presence of fish. Birth rates of most cladoceran taxa did not differ between treatments, whereas death rates were higher in the enclosures with fish than in the fishless enclosures. Bosminopsis deitersi reached higher densities in the presence of fish, probably because of a release from predation by Chaoborus. 4. We convincingly showed strong deviations from trophic cascade‐based expectations, supporting the idea that trophic cascades may be weak in tropical lakes.     

Exploration of a Submerged Sinkhole Ecosystem in Lake Huron

Dissolution of the Silurian-Devonian aquifer in the Lake Huron Basin has produced several karst formations in the bedrock (sinkholes), through which groundwater emerges onto the lake floor. During September 2003, we explored a recently discovered submerged sinkhole ecosystem (55 m × 40 m × ∼1 m) located at a depth of 93 m with a remotely operated vehicle (ROV) equipped with a conductivity-temperature-depth (CTD) system, an acoustic navigational system, a video camera, and a water sampling system. In addition to two morphotypes of benthic mats, a 1–2 m thick visibly cloudy near-bottom nepheloid-like layer (sinkhole plume) with a strong hydrogen sulfide odor prevailed just above the seepage area of clear water. Relative to lake water, water samples collected within the sinkhole plume were characterized by slightly higher (by 4°C) temperatures, very high levels of chloride (up to 175 mg l⁻¹) and conductivity (1,700 μS cm⁻¹), as well as extremely high concentrations of sulfate (1,400 mg l⁻¹), phosphorus (3 mg l⁻¹) and particulate organic matter (400 mg C l⁻¹). Compared to background lake water, sinkhole plume water was characterized by approximately twofold lower C:N ratios and tenfold higher levels of dissolved organic carbon, bacterial biomass as well as heterotrophic bacterial production. Significant uptake of ¹⁴C-bicarbonate in dark incubations provided preliminary evidence for occurrence of chemosynthesis, possibly mediated by specialized Bacteria and Archea present in this submerged sinkhole ecosystem in the Laurentian Great Lakes.     

Sustainability of the Lake Superior Fish Community: Interactions in a Food Web Context

The restoration and rehabilitation of the native fish communities is a long-term goal for the Laurentian Great Lakes. In Lake Superior, the ongoing restoration of the native lake trout populations is now regarded as one of the major success stories in fisheries management. However, populations of the deepwater morphotype (siscowet lake trout) have increased much more substantially than those of the nearshore morphotype (lean lake trout), and the ecosystem now contains an assemblage of exotic species such as sea lamprey, rainbow smelt, and Pacific salmon (chinook, coho, and steelhead). Those species play an important role in defining the constraints and opportunities for ecosystem management. We combined an equilibrium mass balance model (Ecopath) with a dynamic food web model (Ecosim) to evaluate the ecological consequences of future alternative management strategies and the interaction of two different sets of life history characteristics for fishes at the top of the food web. Relatively rapid turnover rates occur among the exotic forage fish, rainbow smelt, and its primary predators, exotic Pacific salmonids. Slower turnover rates occur among the native lake trout and burbot and their primary prey—lake herring, smelt, deepwater cisco, and sculpins. The abundance of forage fish is a key constraint for all salmonids in Lake Superior. Smelt and Mysis play a prominent role in sustaining the current trophic structure. Competition between the native lake trout and the exotic salmonids is asymmetric. Reductions in the salmon population yield only a modest benefit for the stocks of lake trout, whereas increased fishing of lake trout produces substantial potential increases in the yields of Pacific salmon to recreational fisheries. The deepwater or siscowet morphotype of lake trout has become very abundant. Although it plays a major role in the structure of the food web it offers little potential for the restoration of a valuable commercial or recreational fishery. Even if a combination of strong management actions is implemented, the populations of lean (nearshore) lake trout cannot be restored to pre-fishery and pre-lamprey levels. Thus, management strategy must accept the ecological constraints due in part to the presence of exotics and choose alternatives that sustain public interest in the resources while continuing the gradual progress toward restoration. Received 10 December 1999; accepted 13 June 2000.