The structure and function of the microbial food web of Lake Ontario was assessed at 15 stations distributed across 4 transects during the spring and summer of 2003. This was the first major binational study of Lake Ontario since the Lake Ontario Trophic Transfer initiative of 1990. The microbial loop (bacteria, autotrophic picoplankton, heterotrophic nanoflagellates (HNF) and ciliates) and phytoplankton, were enumerated microscopically in addition to measurements of chlorophyll a, size fractionated primary productivity (14C) and bacterial growth (3H). HNF dominated the total biomass in spring (≈300 mg m−3) and summer (≈1250 mg m−3). The size of the organic carbon pool increased from ≈90 mg C m−3 in spring to ≈270 mg C m−3 with HNF contributing 36% of the total organic carbon in the spring and 52% in the summer; however the net balance of the organic carbon pool shifted from autotrophic in the spring to heterotrophic in the summer. The available evidence suggests that HNF are a poor quality food resource for zooplankton and it is likely that the carbon sequestered by HNF is not available to higher trophic levels resulting in dietary stress for planktivores. The implications of high HNF for both organic carbon cycling and maintaining healthy fisheries needs further research. Independent observations show that oligotrophic conditions prevail as evidenced by low phosphorus, low chlorophyll a, low plankton and high water clarity. Such conditions have been generally regarded as the gold standard for managing healthy lakes. Lake Ontario is oligotrophic and healthy from a water quality perspective, but from a food web dynamics point of view, Lake Ontario appears to be unhealthy due to the dominance of HNF, low zooplankton and poor quality of food available to higher trophic levels. We hypothesize that the lake's poor health is attributable to inefficient energy transfer from lower to higher trophic levels. The traditional understanding of trophic state based mainly on water quality criteria needs to be broadened by the inclusion of food web and fisheries based metrics.
Lake Ontario is an important global freshwater resource due to its size; it is the world's 11th largest lake by volume (1640 km3) and the 13th largest by area (18,484 km2) (Reynolds et al., 2000), with a watershed that covers 82,990 km2. Located in the centre of the North American continent and at the lower end of the Laurentian Great Lakes chain, the Lake Ontario watershed is home to over 8 million people and serves as an industrial and commercial hub for Canada and the United States (Environment Canada and Environmental Protection Agency, 1995). As a consequence of anthropogenic activities, the lake has been subject to multiple stressors including eutrophication, toxic chemicals, invasive species (Mills et al., 2003; Munawar, 2003) and climate change. All of these stressors have been shown to have an immediate impact on the structure and function of the microbial-planktonic food web of the lake which has in turn resulted in major perturbations across all trophic levels (Mills et al., 2003; Munawar and Munawar, 2003; Stewart et al., 2009). That being said, understanding the structure and function of the base of the food web is critical to a broader understanding of the health and integrity of the Lake Ontario ecosystem.
The first comprehensive lakewide surveys of the phytoplankton and primary production of Lake Ontario were undertaken in 1970 in response to widespread public concern over eutrophication (Munawar and Nauwerck, 1971; Vollenweider et al., 1974; Munawar and Munawar, 1996). Fish yields in Lake Ontario were also found to be lower than expected given the lake's trophic state, and the need for a more detailed understanding of the system (i.e. comparative ecology) was identified (Leach et al., 1987). The first microbial loop (bacteria, autotrophic picoplankton, heterotrophic nanoflagellates and ciliates) surveys began in the late 1980s as concern shifted to the impact of contaminants on aquatic food webs and microbial communities proved to be robust and sensitive indicators of ecosystem stress (Munawar et al., 1987a; Munawar and Weisse, 1989). In the late 1980s and early 1990s, integrated microbial and planktonic food web research in Lake Ontario revealed important information concerning the disruptive effects of dreissenid grazing on the benthic and pelagic food webs (Munawar and Weisse, 1989; Fahnenstiel et al., 1998; Munawar et al., 2003).
Within the pelagic zone, the microbial-planktonic food web—composed of bacteria, autotrophic picoplankton, heterotrophic nanoflagellates, ciliates and phytoplankton—provides both the major food resource for zooplankton and the mechanism for recycling nutrients (Pomeroy, 1974; Sherr et al., 1986; 1988; Munawar et al., 1999). The relative importance of autotrophic and heterotrophic micro-organisms to the transfer of energy within microbial food webs of aquatic ecosystems has generated some debate in the literature, particularly within the North American Great Lakes. Heterotrophs are widely considered to play a more prominent role in energy transfer of oligotrophic ecosystems (Biddanda et al., 2001; Heath et al., 2003; Heath and Munawar, 2004), however research on ultra-oligotrophic Lake Superior has indicated that the organic carbon pool is dominated by autochthonous sources (Cotner et al., 2004; Urban et al., 2005). Also, in nearshore waters, smaller lakes and embayments, inputs of allocthonous matter can be expected to support a net heterotrophic respiration regime (del Giorgio and Peters, 1994; Hiriart-Baer et al., 2008; Bocaniov and Smith, 2009). Some of the debate, we suspect, is a consequence of not incorporating both structural (e.g. microscopic identification and enumeration) and functional (e.g. physiological) assessments into the study design. Determination of “net autotrophy” and “net heterotrophy” requires a more holistic evaluation of both autotrophs and heterotrophs (Dodds and Cole, 2007).
In order to gain a better understanding of autotrophic and heterotrophic communities and their processes, Munawar et al., (2009) developed an integrated microbial-planktonic food web model for Lake Superior based on microscopic identification and enumeration of both autotrophic and heterotrophic communities combined with radioisotope tracer measurements of primary productivity and bacterial growth. Under this model, a strongly autotrophic system would be predicted to have a high proportion of autotrophic carbon within the microbial-planktonic food web coupled with high turnover rates of autotrophic organisms. At the other extreme, a strongly heterotrophic system would be predicted to have a high proportion of heterotrophic organisms coupled with higher turnover rates of heterotrophs.
For the first time, this current paper offers a detailed analysis of the structure and function of the microbial food web of Lake Ontario based on the Lake Superior model (Munawar et al., 2009). The importance of autotrophic versus heterotrophic communities within the food web will be assessed leading to a discussion of the role of autochthonous and allocthonous sources of carbon in sustaining the Lake Ontario microbial and planktonic food web.
Materials and Methods
Synoptic surveys of Lake Ontario were conducted during April and August of 2003. In each survey, 15 stations were sampled across 4 transects spanning the lake (Figure 1) with some minor changes in site selection from spring to summer. Water was collected from each site using an integrated sampler (Schroeder, 1969) to a maximum of 20 m or 2 m off the bottom at shallower sites. During stratified conditions, epilimnetic water was collected. Phytoplankton, microbial loop and ciliate samples were preserved immediately upon collection. Size fractionated primary productivity and bacterial growth experiments were conducted shipboard during the summer cruise at 11 stations along the central and eastern transects.
Microbial loop samples, including bacteria, autotrophic picoplankton and heterotrophic nanoflagellates, were fixed with 1.6% formaldehyde and enumerated using DAPI staining (Porter and Feig, 1980) under epi-fluorescence microscopy (Munawar and Weisse, 1989). Organic carbon was estimated as 200 f g C cell−1 for autotrophic picoplankton (APP), 10 f g C cell−1 for bacteria and 14 pg C cell−1 for heterotrophic nanoflagellates (HNF) (Sprules et al., 1999). Ciliate samples were preserved with Lugol's iodine (Lynn and Munawar, 1999) and enumerated following the Quantitative Protargol Staining technique (Montagnes and Lynn, 1993). Ciliate organic carbon was calculated as 11% of the fresh weight biomass (Turley et al., 1986).
Chlorophyll a, phytoplankton biomass and size structure
Chlorophyll a concentrations were determined by filtering up to 1 L of water through Whatman GF/C filters followed by cold acetone pigment extraction and spectrophotometric analysis (Strickland and Parsons, 1968). Phytoplankton samples were fixed immediately with Lugol's iodine. Enumeration and measurement followed the Utermöhl (1958) inverted microscope technique as described by Munawar et al. (1987b). Organic carbon was estimated for the various groups of phytoplankton using equations from the literature namely: Diatomeae (Strathmann, 1967), Chlorophyta, Chrysophyceae, Cryptophyceae and Dinophyceae (Verity et al., 1992), and Cyanophyta (Lee and Furman, 1987). Size structure was determined as Equivalent Spherical Diameter (ESD) as per the procedures given in Munawar and Munawar (1996, 2000).
Size fractionated primary productivity
Size fractionated primary productivity was estimated for three size categories of phytoplankton (<2 μm, 2–20 μm and >20 μm) by the 14Carbon technique following the standard protocol of Munawar and Munawar (1996). Whole water samples were spiked with Na14CO3, incubated for 4 hours at surface temperature and exposed to a constant light level of 240 μE s−1 m−2. Because light and temperature levels are constant in these experiments, the results should be interpreted as potential rather than actual. After incubation, size classes were determined by filtration of the sample through polycarbonate filters, all filters were rinsed with 0.5 N hydrochloric acid to remove excess 14C and radioactivity was determined by liquid scintillation counting.
Bacterial Productivity (3H-Leucine incorporation)
Bacterial growth rates were estimated by 3H-Leucine incorporation into bacterial proteins following the protocol of Jørgensen (1992). Radioactivity was determined by liquid scintillation counting. Detailed procedures are available in Heath and Munawar (2004).
Microbial loop biomass during April 2003 is shown in Figure 2. Total microbial biomass ranged from 80–1225 mg m−3 with a mean of 352 ± 83. Of the components, bacteria ranged from 20–72 mg m−3 (40.4 ± 3.3); autotrophic picoplankton (APP) ranged from 0.5–9.3 mg m−3 (2.3 ± 0.6); heterotrophic nanoflagellates (HNF) ranged from 39–1179 mg m−3 (305.8 ± 82.6) and ciliates ranged from 0 (at 4 sites) to 17 mg m−3 (3.5 ± 1.4). The spring phytoplankton community is shown in Figure 3. Chlorophyll a concentrations varied from 0.8–3.0 μg l−1 (1.3 ± 0.1) and phytoplankton biomass ranged from 40–1241 mg m−3 (239 ± 82) (Figures 3a, b). With respect to the size of the phytoplankton cells, the weighted mean equivalent spherical diameter (ESD) (by biomass) ranged from 1.6–17.7 μm (Figure 3c) and the nanoplankton size fraction (2–20 μm) was typically dominant (Figure 3d).
During August 2003, microbial loop biomass ranged from 700–3800 mg m−3 (1558 ± 217). The structure of microbial loop is shown in Figure 4. Bacteria ranged from 80–250 mg m−3 (185 ± 11); APP ranged from 26–218 mg m−3 (88 ± 13), HNF ranged from 340–3370 mg m−3 (1249 ± 211) and ciliates ranged from 7–122 mg m−3 (36 ± 7). The phytoplankton community is shown in Figure 5. Chlorophyll a concentrations fluctuated from 0.8–6.8 μg l−1 (1.9 ± 0.4) and phytoplankton biomass ranged from 100–1400 mg m−3 (286 ± 82). The mean ESD at each station fell between 4.8 and 45.1 μm and typically net plankton were the major size class.
Size fractionated primary productivity experiments showed rates of 0.5–14.5 mgCm−3h−1 for net plankton, 1.0–6.9 mgCm−3h−1 for nanoplankton and 0.2–7.1 mgCm−3h−1 for picoplankton (Figure 6a–c). Bacterial growth rates ranged from 0.01–0.17 mgCm−3h−1. (Figure 6d) P/B quotients expressed as carbon turnover rates are shown logarithmically in Figure 7. Actual values were 0.3–15.9 d−1 for net plankton, 0.1–7.1 d−1 for nanoplankton and 0.1–9.3 d−1 for picoplankton. Bacterial carbon turnover rates ranged from 0.002–0.23 d−1.
The structure of the microbial food web is given in Figure 8 expressed as both freshweight and organic carbon. The pie charts show a relative % based on the lakewide mean for each season. The mean biomass for all components was 590.8 mg m−3 in the spring and 1844.1 mg m−3 in the summer. Likewise, the mean organic carbon for all components was 92.9 mg C m−3 in the spring and 266.8 mg C m−3 in the summer. Heterotrophic nanoflagellates were the largest single component in both spring and summer contributing 55% and 69% of the biomass respectively. However, phytoplankton contributed 58% of the organic carbon in the spring compared to 36% for HNF. In the summer, HNF contributed 52% of the organic carbon compared to 36% for phytoplankton.
The first lakewide study of the microbial food web of Lake Ontario was conducted in 1990 as part of the LOTT (Lake Ontario Trophic Transfer) initiative (Munawar et al., 2003; Munawar and Munawar, 2003). To give a sense of how the lake may have changed since that time, we provide a comparison of the microbial loop, chlorophyll a, phytoplankton biomass and size structure and primary productivity between August 1990 and August 2003 in Figure 9 at comparable stations. Significant increases in bacteria (from 85.2 ± 5.5 mg m−3 to 180.7 ± 16.9 mg m−3) and HNF (from 118.2 ± 22.6 mg m−3 to 1144.2 ± 215.0 mg m−3) were observed. Significant reductions in phytoplankton biomass (from 1.9 ± 0.3 g m−3 to 0.6 ± 0.4 g m−3) were observed primarily from a decline in nano-plankton. Chlorophyll a and size fractionated primary productivity did not show significant changes.
On a lakewide basis, microbial biomass increased dramatically from a mean of 352 mg m−3 ± 83 in the spring to 1558 mg m−3 ± 217 in the summer, with every component (Bacteria, APP, HNF, Ciliates) showing a significant rise. Heterotrophic nanoflagellates were the largest component in both seasons showing an increase from 306 to 1250 mg m−3. By comparison, autotrophic picoplankton was the smallest component in the spring whereas ciliates were least during the summer. The significance of HNF has been already alluded to in earlier studies of Lake Ontario. For example, Munawar and Nauwerck (1971) observed large numbers of “colourless chrysomonads” in their samples, but did not quantify their relative abundance or biomass. Similarly, Pick and Caron (1987) observed that “heterotrophic nanoplankon” abundance (≈103 cells ml−1) could exceed that of phototrophic plankton during the summer. However, microbial food web studies of Lake Ontario conducted in the late 1980s and early 1990s (Munawar and Weisse, 1989; Fahnenstiel et al., 1998; Munawar et al., 2003) did not show such large numbers of HNF suggesting that there may be considerable inter-annual variability in HNF abundance.
Lakewide, phytoplankton biomass did not change significantly from spring (239 mg m−3 ± 82) to summer (286 mg m−3 ± 82) and neither did phytoplankton size structure. ESD was 9.8 μm ± 1.7 during spring and 11.2 μm ± 2.7 during summer. Nanoplankton was predominant at most stations during spring and summer. However, there was considerable variability with phytoplankton size at individual stations. For example during spring, ESD exceeded 12 μm at 7 stations and was less than 6 μm at the remaining sites. The summer data were also skewed by the high ESD recorded at station 43 (45.1 μm) and station 81 (22.5 μm), otherwise mean ESD was below 8 μm. In fact the observed low biomass (<300 mg m−3 in spring and summer) was well below the ≈1000 mg m−3 reported in Lake Superior for 2001 (Munawar et al., 2009). These data are reflective of ultra-oligotrophic conditions in Lake Ontario according to the trophic classification of Munawar and Munawar (1982). Rapid oligotrophication of the lake has been reported previously by others (Mills et al., 2003; Munawar and Munawar, 2003) and the present study confirms these observations.
Size fractionated primary productivity and bacterial growth rates were estimated during the summer cruise on the eastern half of the lake. Mean primary productivity (all size fractions) was 5.8 mgCm−3h−1 with none of the three size fractions: net plankton (2.0 mgCm−3h−1 ± 1.3), nanoplankton (2.1 mgCm−3h−1 ± 0.5) and picoplankton (1.7 mgCm−3h−1 ± 0.6) showing significant differences in the rate of carbon assimilation. The lack of differential rates of primary production amongst the various size fractions is unusual since higher nanoplankton and picoplankton productivity relative to net plankton productivity have been reported in oligotrophic lakes such as Lake Superior (Munawar et. al., 1978; Munawar and Munawar, 2009).
Bacterial growth rates (0.1 mgCm−3h−1 ± 0.02) were 1–2 orders of magnitude lower than primary production rates for each size class of phytoplankton (i.e. net, nano- and pico-). Virtually all bacterial production is expected to be consumed by micro-zooplankton in oligotrophic environments and therefore serve as an important vector for energy exchange (Heath et al., 2003). However, these rates are low compared to the 0.9 mgCm−3h−1 reported in other offshore waters of the Great Lakes (ibid) and may help to explain the suppressed zooplankton community observed in Lake Ontario during 2003 (Holeck et al., 2008).
Carbon turnover rates (P/B ratios) were also estimated for bacteria, picoplankton, nanoplankton and net plankton during the summer cruise. Net plankton P/Bs (5.9 d−1 ± 2.0) were significantly higher than nanoplankton (0.9 d−1 ± 0.6), picoplankton (2.7 d−1 ± 0.8) and bacteria (0.1 d−1 ± 0.02). The carbon turnover of total phytoplankton was about 30 times greater than bacteria. These findings indicate that phytoplankton, specifically net plankton, was the dominant vector of energy transfer in Lake Ontario. These results are somewhat unexpected, since other studies on oligotriophic waters in the Great Lakes have suggested that picoplankton (Munawar et al., 2009; Ivanikova et al., 2007) and even heterotrophic bacteria (Heath and Munawar, 2004; Haffner et al., 1988) were likely the dominant vectors for energy transfer.
The relative importance of heterotrophs and autotrophs in the organic carbon pool reveals interesting information about potential energy transfer in Lake Ontario. We observed that on a lakewide basis, 58% of the organic carbon was autotrophic during the spring and 60% of the organic carbon was heterotrophic during the summer. While the structure of the carbon pool shifted from autotrophic in spring to heterotrophic in the summer, HNF contributed 36% of the organic carbon during the spring and 52% of the organic carbon during the summer. Despite the importance of HNF to the organic carbon pool of Lake Ontario, little is known about its dynamics. Our work on the Bay of Quinte (north eastern Lake Ontario) has also shown the microbial food web to be dominated by HNF (Munawar, unpublished data). Other researchers have shown that HNF can achieve relatively high growth rates. Within the Great Lakes basin, Carrick et al. (1992) identified Chromulina sp. and Katablepharis ovalis as the dominant HNF species in the oligotrophic waters of Lake Michigan with maximum growth rates of 1.0 d−1 exceeding those of autotrophic phytoflagellates. Similarly, in mesotrophic Lake Constance (Austria, Germany, Switzerland), Weisse (1997) found that average seasonal HNF production ranged from 1.4–3.5 mg C m−3 d−1. The relatively high growth rates of HNF observed in Lakes Michigan and Constance would appear to indicate that a high biomass of HNF could be attained in the absence of grazing pressure.
Zooplankton, particularly larger cladocerans and rotifers, have been shown to reduce the overall abundance of HNF in oligotrophic environments (Tadonléké et al., 2004). In Lake Ontario, we would argue that the relative dearth of zooplankton (3.8 mg m−3 in spring and 23.7 mg m−3 in summer: Holeck et al., 2008) has allowed an opportunity for HNF to thrive. While HNF are known to consume bacteria and picoplankton, the most likely food resource for HNF in Lake Ontario is picoplankton given the higher carbon turnover rates (2.7 d−1) which would be indicative of HNF grazing pressure. By way of comparison, bacterial carbon turnover rates were only 0.1 d−1. Picoplankton carbon turnover rates are roughly 4 times greater than those observed in Lake Superior during the summer (0.7 d−1) where the proportion of HNF is considerably lower (Munawar et al., 2009). It seems that HNF have a very important role in organic carbon cycling of Lake Ontario. Specifically, it appears that HNF are acting as a sink for a large pool of energy which is potentially sequestering the autochthonous carbon and making it unavailable to zooplankton and higher trophic levels.
This sequestering of autochthonous carbon would be predicted to have a bottom up effect on higher trophic levels (Brett et al., 2009; Fitzpatrick et al., 2008). In Lake Ontario, we estimate that only 55 mg C m−3 of carbon from autotrophs were available in the spring and 105 mg C m−3 were available in the summer. This in turn supported <2 mg C m−3 of zooplankton in the spring and roughly 11 mg C m−3 in the summer (Holeck et al., 2008; assuming zooplankton organic carbon is equivalent to 48% of the dry weight biomass after Anderson and Hesson, 1991). Indeed, Holeck et al. (2008) attributed the low zooplankton biomass to the poor food quality supplied by the phytoplankton. Similarly, Schlechtriem et al. (2008) observed that Mysis relicta from Lake Ontario had poor quality diets based on their fatty acid profiles at time of capture. With respect to the fisheries, M. relicta and other zooplankton are important food resources for various prey fish including alewife, rainbow smelt and slimy sculpin, which had faced declines historically and particularly during the study period (Walsh et al., 2008; O’Gorman et al., 2008). These observations are consistent with the view that a great amount of energy is taken out of the system by abundant HNF acting as a carbon sink.
The microbial food web assessment deployed in this study was originally applied to Lake Superior and it was concluded that Lake Superior was an ultra-oligotrophic environment (Munawar et al., 2009) with an efficient operating food web. In Table 1 (a,b), we compare a number of parameters of the microbial-planktonic food webs of Lake Superior and Lake Ontario. We observed that the biomass of heterotrophs (mainly HNF) was 3–7 X greater in Lake Ontario than in Lake Superior and that the biomass of autotrophs (e.g. phytoplankton) was much smaller in Lake Ontario (by a factor of 3). Also, carbon turnover rates for bacteria and phytoplankton in Lake Ontario were 1–2 orders of magnitude greater than in Lake Superior.
As discussed earlier, Lake Ontario is oligotrophic similar to Lake Superior. However, the two lakes differ dramatically when their microbial food webs are compared. The Lake Superior MFW was characterized by the dominance of autotrophs which contributed more than 90% to the organic carbon with picoplankton showing the highest P/Bs. These criteria confirm the healthy state of Lake Superior. By contrast, in Lake Ontario we found that over 50% of the organic carbon pool was composed of heterotrophs and that net plankton had the highest P/B quotients. Furthermore, Lake Ontario's microbial food web structure suggests that HNF may be a carbon sink inhibiting the flow of energy that would otherwise be available to zooplankton and passed up the food chain. Both Lake Superior and Ontario are oligotrophic and healthy based on water quality criteria, however they appear quite different when the microbial–planktonic community and trophic transfer are considered. Such observations have serious implications for evaluating lower to higher trophic level linkages. An attempt has been made here to suggest one such hypothesis about the role of HNF. Our data reveal a dichotomy in that healthy lakes from a water quality point of view may not harbour healthy fisheries, and Lake Ontario as observed in 2003 appears to be one such case.
The apparent dichotomy may be due to the fact that trophic state is a balance between autotrophic and heterotrophic processes rather than just a result of autotrophic activity (Dodds and Cole, 2007). The current state of knowledge is that the pelagic waters of Lake Ontario have been rapidly becoming more oligotrophic as evidenced by low chlorophyll a and low zooplankton especially Daphnia (Dove, 2009; Holeck et al., 2008). But, this conclusion does not account for the importance of the microbial loop to the lower food web and in particular the role of heterotrophic nanoflagellates (HNF). Daphnia have been shown to regulate HNF biomass (Porter, 1996; Pace, 1993), so the absence of Daphnia may help to explain the high HNF biomass observed in Lake Ontario. Furthermore, Daphnia and other Cladocerans as well as HNF have been shown to graze on bacteria (Hwang and Heath, 1997, 1999) so in Lake Ontario this potential food resource would be available almost exclusively to HNF. However, this raises the question of whether or not there is sufficient bacterial production to sustain the large biomass of HNF observed or are the HNF also grazing the primary production as well as detrital material? Very little is known about HNF therefore production and grazing experiments are clearly needed to understand their dynamics. In Lake Ontario, we hypothesize that autochthonous production is being sequestered by heterotrophic nanoflagellates and not passed on to zooplankton. The consequence would be that energy is not being transferred along a traditional grazing food chain and fisheries that depend on this energy vector (i.e. planktivores and piscivores) are necessarily imperiled. Integrated microbial food web studies that include structural and functional assessments of autotrophic and heterotrophic communities coupled with grazing experiments are essential for a robust and sensitive determination of ecosystem health.
Summary and Conclusions
We examined the structure and function of the microbial–planktonic food web of Lake Ontario during the spring and summer of 2003. Independent observations showed that Lake Ontario is oligotrophic as evidenced by low phosphorus, low chlorophyll a, low plankton biomass and high water clarity. From a management perspective, these conditions are the gold standard for achieving a healthy lake. However, we conclude from our analysis of the microbial–planktonic food web that Lake Ontario is unhealthy when compared with oligotrophic and healthy Lake Superior, due to the dominance of heterotrophic nanoflagellates (HNF), low zooplankton biomass and the poor quality of food available to higher trophic levels. We put forth the hypothesis that autochthonous energy is being sequestered by heterotrophic nanoflagellates, disrupting the normal trophic transfer to zooplankton in Lake Ontario as a framework to guide future research and management efforts. We recommend that management and action plans not be based on simple masurements of chlorophyll a as a proxy for the autotrophic community, but rather be expanded to include microbial-planktonic communities. It is suggested that the trophic classification of water bodies, largely based on water quality parameters, be broadened to include food web metrics for a more comprehensive assessment of ecosystem health. Future research in Lake Ontario and other lakes needs to be directed at the role of HNF in organic carbon cycling in order to understand the mechanisms of trophic transfer and to achieve sustainable fisheries.
We sincerely thank Drs. J.H. Leach, R.T. Heath and E.L. Mills for their constructive criticism which greatly improved the quality of this manuscript. Thanks are also due to Jennifer Lorimer for the graphics and Susan Blunt for the technical editing of this manuscript.