Limnol. Oceanogr., 36(7), 1991,133s1345 0 199 1, by the American Society of Limnology and Oceanography, Inc.

The importance of zooplankton-protozoan trophic couplings in Lake Michigan

Hunter J. Carrick’ and Gary L. Fahnenstiel Cooperative Institute for Limnology and Ecosystem Research, Great Lakes Environmental Research Labor -atory, NOAA, 2205 Commonwealth Blvd., Ann Arbor, Michigan 48105

Eugene F. Stoermer and Robert G. Wetzel2 Center for Great Lakes and Aquatic Sciences, University of Michigan, Ann Arbor 48109

Abstract The importance of the zooplankton-protozoan trophic coupling was determined experimentally by measured changes in protozoan growth rates with increasing zooplankton biomass. In five of six experiments conducted in Lake Michigan, a significant inverse relationship between protozoan growth and zooplankton biomass was observed (avg r2 = 70%), Zooplankton clearance rates on protozoan assemblages (range, 1.0-6.2 ml (pg dry wt)-I d ‘1 were comparable to those previously measured for phytoplankton which suggested that protozoa are important prey for zooplankton. Clearance rates on individual protozoan taxa [O-15.6 ml (Fg dry wt)-I d-l] were size-dependent. Rates were greatest for taxa ~20 grn in size (mainly nanoflagellates and small ciliates). In contrast to findings for phytoplankton, no evidence emerged for grazer resistance nor growth enhancement by planktonic protozoa in response to grazing. The high flux rates for macrozooplankton on heterotrophic nanoflagellates observed in all experiments (0.2-6.0 pg C liter-’ d-l) provided evi- dence that a large fraction of picoplankton C may be directly transferred to higher trophic levels via a picoplankton-flagellate-zooplankton coupling.

Debate continues as to whether macro- Moreover, the harvesting efficiency of mac- zooplankton (rotifers, crustacean zooplank- rozooplankton on picoplankton cells varies ton) can effectively graze on components of among age classes and species of zooplank- the microbial food web (picoplankton, pro- ton consumers (Pace et al. 1983). Because tozoa), and thus, whether a significant most grazing on picoplankton is by proto- amount of C is transferred from picoplank- zoa (Sanders et al. 1989; Fahnenstiel et al. ton to higher trophic levels (e.g. Sherr et al. 199 1), without information on the fate of 1987). Recent studies suggest that pico- protozoa (Weisse et al. 1990) it remains dif- plankton-sized cells can be ingested by mac- ficult to assess whether picoplankton pro- rozooplankton; however, picoplankton pro- duction is an important trophic link to mac- duction is not efficiently grazed or grazing rozooplankton. may be intermittent (Pace et al. 1990). Information is limited on macrozoo- plankton predation on protozoa (seeStoeck- ’ Present address for correspondence: Department of er and Capuzzo 1990). Of the studies that Fisheries and Aquaculture, University of Florida, 7922 do exist, most are laboratory investigations NW 71st St., Gainesville 32606. which have demonstrated that single co- 2 Present address: Department of Biology, Univer- pepod species can actively graze protozoa sity of Alabama, Tuscaloosa 35487. and that a diet composed entirely of flag- Acknowledgments ellated or ciliated protozoa can support We thank the crew of the RV Shenehon: B. Burns, growth and reproduction of the macrozoo- J. Grimes, and D. Morse. Also, J. Cavaletto, M.J.P.G.R. McCormick, and T. Patton provided assistance in the plankton species tested (e.g. Stoecker and field. M. Brenner, W. S. Gardner, C. L. Schelske, and Egloff 1987; Sanders and Porter 1990). There H. A. Vanderploeg gave criticisms on the manuscript. is also some indication that copepods prefer Discussions with J. T. Lehman concerning expcrimen- protozoa over algae and that macrozoo- tal design proved helpful. plankton taxa fed seston ingest particles in Contribution 736, Great Lakes Environmental Re- search Laboratory. the nano- to microplankton size range (e.g. This paper is dedicated to the memory of Hunter Nival and Nival 1976). Little is known, Joseph Carrick, Sr. however, about predation by intact zoo- 1335 1336 Carrick et al.

plankton assemblages on naturally occur- hours). Water-column temperature profiles ring planktonic protozoa (Kleppel et al. were measured with an electronic bathy- 1988). Laboratory findings cannot be ac- thermograph. Ambient protozoan abun- curately extrapolated to the field because dances were determined from lake-water grazing is not equal among co-occurring samples transferred into 250-ml amber bot- macrozooplankton species (e.g. Paffenhiifer tles and preserved with either 1% Lugol’s and Knowles 1980). Moreover, although acid iodine (ciliate and microflagellate Sam-, grazer resistance is a common adaptation ples) or 1% glutaraldehyde buffered with 0.1 of phytoplankton (Porter 1976; Lehman and M sodium cacodylate (nanoflagellate Sam-, Sandgren 1985), the ability of protozoa to ples). Because of the large range in both cell withstand or avoid predation is largely un- size and abundance among protozoa, the known. abundance of microprotozoa (microflagel-n The specific objectives, of this study were lates and Ciliophora, most >20 and ~200 to determine macrozooplankton clearance pm in size) and nanoprotozoa (nanoflagel-, rates on protozoa, assess whether macro- lates ~20 I.crn in size) was measured sepa- zooplankton feed selectively on protozoa of rately. Moreover, potential trophic status of a particular size, and evaluate the quanti- nanoprotozoa was assessedby the presence tative importance of carbon flux from pro- (pho totrophic, Pnano) or absence (hetero- tozoa to macrozooplankton. We present re- trophic, Hnano) of pigmentation (see be-, sults based on six experirnents conducted in low). 1989 (March-October) in Lake Michigan. Nanoprotozoa were enumerated with epi-a These experiments are unique, in that graz- fluorescence microscopy from slides pre-, ing was measured on natural assemblages pared within 24 h of samphng. Subsamples of protozoa by a complete macrozooplank- ( 1O-20 ml) were filtered on to prestained (Ir- ton community while the concentration and galan Black) 0.8~pm pore-size Nuclepore fil- composition of both plankton components ters that were subsequently stained with were maintained at levels representative of primulin. Filters were mounted between a ambient lake conditions. We believe the re- microscope slide and cover-slip with im- sults of this investigation are probably rep- mersion oil (Caron 1983). Microprotozoan resentative of many unproductive fresh- biomass and community composition were water and marine systems because Lake determined with the Uterriiohl technique Michigan is a large (surface area, 58,000 whereby subsamples (25-50 ml) were set- km2; mean depth, 84 m), oligotrophic lake tled onto coverslips and systematically enu- with chlorophyll concentrations ranging merated with an inverted microscope from 0.3 to 3.0 pg liter-l and total P values (400 X ). Cellular volume estimates for Pnano varying between 4 and 8 pg liter-’ (Schelske and microflagellates were converted to C et al. 1986). Moreover, the composition and based on Strathmann ( 1967) conversion abundance of picoplankton (Scavia and factors; Hnano and ciliate cell volumes were Laird 1987; Fahnenstiel and Car-rick 199 1) converted to C with the conversion factor and protozoa (Car-rick and Fahnenstiel 1989, (cell volume X 0.15 g C ml-‘) of Laws et 1990) are similar to those in various fresh- al. (1984). C estimates were corrected for water and marine ecosystems. cell shrinkage due to preservation (Choi and Stoecker 1989). Protozoan systematics used Methods here conform to those presented by Lee et Ambient conditions--Sampling was con- al. (1985). ducted at a single offshore station in Lake Macrozooplankton abundance was deter - Michigan (43’1’ 11“N, 86’36’48”W; max mined from vertical net hauls (0.5-m ap- depth, 100 m) on seven occasions in 1989 erture, 153-pm mesh, Wisconsin-type net) (29 March, 19 April, 101May, 13 June, 10 through the photic zone (O-40 m). Collected July, 28 August, and 4 October). For all samples were then narcotized and preserved analyses, water was collected from the sur- with sugar Formalin. Macrozooplankton face mixed layer (5 m) with a 5-liter or 30- biomass was determined by enumerating liter PVC Niskin bottle at dusk (2000-2200 subsamples and subsequently converting Zooplankton-protozoan coupling 1337 abundances to dry weight biomass with tax- Lake Michigan (Schelske et al. 1986). TO on-specific conversion values for Lake explore the possibility that macrozooplank- Michigan zooplankton (Hawkins and Evans ton might be supplying protozoa with or- 1979). ganic compounds which could enhance Zooplankton grazing experiments-The growth by either direct uptake (Haas and impact of macrozooplankton (organisms Webb 1979) or by augmenting bacterial prey > 153 pm in size) grazing on protozoa in densities (Taylor and Lean 198 l), we added Lake Michigan surface waters was deter- glucose (0.09 PM final concn) to one 0 X and mined on six of the seven sampling dates one 1 x bottle on several dates (13 June, 10 (excluding 10 May) by experimentally ma- July, 28 August, and 4 October). nipulating macrozooplankton concentra- All bottles were incubated for 24 h at am- tions across a series of bottles and evalu- bient light and temperature in a shipboard ating changes in protozoan densities within incubator equipped with rotating racks. Ini- the bottles over time (Lehman 1980). Col- tial and final subsamples for nano- and mi- lected lake water (150 liters) was screened croprotozoa were removed from the bottles, through a 153-pm mesh-size Wisconsin-type preserved, and enumerated (as described zooplankton net and dispensed into a shad- previously) in order to estimate exponential ed 200-liter polyethylene tank, after which growth by an additional 90 liters of screened lake water was dispensed into a second 200-liter tank. Epilimnetic macrozooplankton were col- lected with a 10-m vertical haul via a solid where r is the rate of population growth (d-l), bucket (to avoid excessive mechanical N,, and Nt are initial and final cell densities, damage to the zooplankton) and carefully and t is the duration of incubation. At the added to the 90-liter tank by submerging end of each experiment, macrozooplankton the bucket into the collected lake water and abundances in each bottle were determined allowing the macrozooplankton to escape. by passing the entire contents of each carboy Macrozooplankton treatments were ad- through a 153-pm mesh-size screen and ministered by filling lo-liter carboys with counting the retained animals as described screened lake water (from the 150-liter sam- above. ple) and subsequently inoculating them with The relationship between protozoan subsamples from the 90-liter zooplankton growth (dependent variable) and zooplank- sample, so that concentrations in the car- ton biomass (independent variable) was as- boys were “lx, 1.5x, and 3~ ambient sessed with simple linear regression. The macrozooplankton concentrations. Some slope of this relationship provides an esti- bottles had no macrozooplankton added and mate of the weight-specific zooplankton served as the 0 x treatment. In experiments clearance rate on protozoa (pg dry wt liter-l conducted on 29 March, the 1 x and 3 x d- ‘) and the y-intercept is an estimate of the macrozooplankton treatments were repli- exponential growth rate (d-l) of the proto- cated (total number of experimental con- zoa (Lehman and Sandgren 1985). We cal- tainers, n = 6); on 19 April and 13 June, culated the flux of C from protozoa to mac- the 1 x , 1.5 x , and 3 x treatments were rep- rozooplankton (bg C liter-l d-l) by licated (n = 7); all four zooplankton treat- multiplying the clearance rate for the pro- ments were duplicated for the remaining tozoan group under question by the ambient three experiments (n = 8). macrozooplankton biomass and, in turn, To minimize the effects of nutrient re- multiplying this product by the ambient cycling via zooplankton excretion to pro- biomass of the protozoan group itself. tozoa (particularly phototrophic forms), we added phosphate (0.23 PM final concn) to Results all bottles after thermal stratification (June Ambient conditions-The abundance of through October, Scavia and Fahnenstiel planktonic protozoa and macrozooplank- 1987). We chose to add P to our bottles, ton varied temporally with biomass maxi- because it limits phytoplanktonic growth in mal during early to midstratification (Fig. 1338 Carrick et al.

1). These temporal patterns seem to be typ- ical for these planktonic components in Lake Michigan based on previous studies (Car- rick and Fahnenstiel 1989, 1990; Dorazio et al. 1987). The Hnano assemblage (Fig. 1A) was composed of colorless chrysomonads, cryp- tomonads, and choanoflagellates whose C concentrations ranged from 5 to 15 pg liter’ throughout the course of this study and reached a maximum during midstratifica- tion (temp. > 15°C July-September). Pnano C (range, 16-63 pg liter-l) was dominated by chrysomonads and cryptomonads and reached maximal concentrations during ini- tial (temp. >4” < 15”C, June) stratification (Fig. 1B). Microflagellate C was composed entirely of dinoflagellates (range, 1.2-9.8 I,cg liter-l) and also increased during initial stratification, yet remained high until Oc- tober when C mass fell to levels common for spring isothermal (temp. <4”C, March- May) assemblages (Fig. 1C). Ciliate C (Fig. 1D) was also low prior to thermal stratifi- cation (< 5 pg liter-‘) and increased during initial stratification (33.9 pg liter-‘) when large spirotrichs were dominant. During midstratification, C declined to an inter- mediate level (5-l 0 pg liter’) and was com- posed of smaller spirotrichs and prorodon- tids, and this assemblage remained throughout the rest of our sampling. IV Lastly, macrozooplankton C was low dur- 30 ing the spring isothermal period and in- creased sharply during initial stratification 20 (Fig. 1E). C mass was maximal during mid- stratification, as levels near 50 pg liter-’ were observed in July, and remained relatively IO high thereafter (-25-30 bg liter-‘). Mac- rozooplankton communities were domi- * nated by copepods (juvenile and adult Diap- tomus spp.) in spring, while Daphnia (mainly D. galeata) constituted a significant fraction of total zooplankton biomass after thermal stratification (avg, 5 1.2%). Zooplankton grazing rates- Both nano- and microprotozoan growth rates were in-

--. c Fig. 1. Carbon mass among five groups of plankton sampled approximately monthly (March-October) from MONTH the surface waters of Lake Michigan in 1989. Zooplankton-protozoan coupling 1339 versely related with zooplankton biomass Table 1. Summary of regression analyses assessing from the bottle experiments on nearly all the relationship between the growth of nano- (Hnano and Pnano) and micro- (microflagellates and ciliates) dates (Table 1; Figs. 2, 3). On 19 April a protozoa and increasing macrozooplankton biomass marginally significant regression was ob- (pg liter-l dry wt) in Lake Michigan. served for nanoprotozoa (P < 0. lo), while Clcar- microprotozoan growth varied indepen- ante dently from zooplankton biomass. Zoo- Imlhg Flux Proto- dry- bg c plankton biomass accounted for 58-84% of zoan w--l lit&’ the variation in protozoan growth and 1989, temp. group d-l] d ‘) r* Prob. yielded estimates of zooplankton clearance 29 Mar, 2.O”C Nano 2.6 0.99 68.3 0.043 rates on protozoa. Micro 1.4 0.14 70.1 0.038 The intercept from these analyses (Figs. 2, 19 Apr, 2.9”C Nano 4.5 2.43 50.0 0.076 3) showed growth rates for nano- and micro- Micro 2.0 0.16 8.0 0.538 protozoa that were lower (range, - 0.17-0.47 13 Jun, lO.O”C Nano 2.1 8.03 84.4 0.003 d-l) than estimates with other techniques Micro 1.0 2.36 58.2 0.046 10 Jul, 19.O”C Nano 4.2 24.20 66.9 0.013 (Carrick 1990). These conservative growth Micro 4.7 8.28 83.7 0.001 estimates are probably a result of chemical 28 Aug, 21.S’C Nano 6.2 11.31 74.3 0.006 manipulations inside our experimental bot- Micro 4.8 4.24 85.1 0.001 tles due to the added P; however, this pro- 4 Ott, 13.8”C Nano 1.6 3.57 68.6 0.011 cedure was necessary to minimize potential Micro 1.2 0.82 75.9 0.005 differences in nutrient recycling from zoo- plankton to protozoan prey at differing mac- rozooplankton abundance (Lehman 1980). The reasonably strong relationship derived wt)-1 d-l] (Table 2). On four of six experi- from these experiments demonstrates the ment dates, protozoan taxa in the <7-~m validity of this assumption, despite conser- size range (Chromulina sp. I, Chromulina vative estimates of growth. Lastly, neither sp. 2, Ochromonas sp., and Katablepharis nano- or microprotozoa were affected by ovalis) were cleared by macrozooplankton adding glucose to 0 x and 1 x bottles, which at greater rates than other protozoa, while suggests that the growth of these organisms larger (> 30 pm) protozoa (Strombidium vir- is not currently limited by this organic com- ide, Strobilidium velox, and Ceratium hi- pound. This conclusion is tentative because rudinella) were grazed only during thermal it is difficult to ascertain growth responses stratification, which coincided with in- to nutrient addition with such short incu- creased zooplankton biomass. Moreover, bation periods (Healy 1979). zooplankton clearance rates decreased with The range in weight-specific clearance increasing cell size; the average clearance rates for macrozooplankton on nano- and rates for 16 common protozoan prey (Table microprotozoa varied from 1.6 to 6.2 and 2) were negatively correlated with cell size from 1.0 to 4.8 ml &g dry wt)-1 d-l (Table (Fig. 4). Variation in the average clearance 1). Average clearance rates on Hnano were rate for these 16 taxa on each date did not nearly twofold higher than those on Pnano, correlate with changes in macrozooplank- while clearances of microflagellates and cil- ton biomass or composition. Consistent with iates were similar (Table 2). In addition, this observation, only one taxon (Vorticella both nanoprotozoan and microprotozoan sp.) was unaffected by increasing macro- clearance rates increased during thermal zooplankton biomass on each of the three stratification and were correlated with in- dates that it was abundant in the water col- creasing macrozooplankton biomass (r = umn (Table 2). 0.87, n = 6, P < 0.01; and r = 0.83, n = 6, P < 0.01, respectively). Discussion Zooplankton selectivity-A great range Macrozooplankton grazing on proto- was observed in the weight-specific clear- zoa -Few studies have measured macro- ance by macrozooplankton on abundant zooplankton grazing on protozoa in the field. protozoan populations [O-l 5.6 ml (pg dry The present study is unique because we were 1340 Carrick et al.

NANOPROTOZOA 0.2 . T- D. IO July

n - P 6 -

60 June

-0.2 i I -0.3 1 -0.4h--7-d -0.61 I I-1 0 20 4C 60 80 100 120 0 too 200 300 ZOOPL,ANKTON BIOMASS (pg liter-l&y wt) Fig. 2. Nanoprotozoan growth rates regressed onto increasing zooplankton biomass for surface communities in Lake Michigan sampled in 1989. (Bottles with 0.09 PM glucose added-E.) Regression statistics arc presented in ‘Table 1. able to assessmacrozooplankton predation ments [3.53 and 2.52 ml (pg dry wt))’ d-l] on protozoa, while main taming reasonably were at the high end of the range for rates natural composition and concentration of reported on epilimnetic phytoplankton in both plankton components. Our results in- Lake Michigan [range, O-2.6 ml (pg dry wt-’ dicate that nano- and microprotozoa are d-l, Scavia and Fahnenstiel 19871. Also., consistently ingested by macrozooplankton clearance rates of smaller Hnano cells (avg., in Lake Michigan, and that zooplankton 4.1; SD = 1.9) by macrozooplankton de- biomass explained >70% of the variation termined here were almost twofold higher in protozoan growth (on average and ex- than clearance rates measured for phyto- cluding the nonsignificant regressions on 19 plankton both in surface and deep regions April). of Lake Michigan (Scavia and Fahnenstiel ln contrast to other studies that estimated 1987; Fahnenstiel and Scavia 1987). macrozooplankton grazing on components Laboratory studies indicated that mac- of the microbial food web (e.g. Pace et al. rozooplankton have higher ingestion rates 1983, 1990), we found that protozoa were for ciliates than for algae (e.g. Stoecker and consistently importan’t prey items for Egloff 1987). Some field studies showed that epilimnetic macrozooplankton in Lake a natural assemblage of microzooplankton Michigan. Average estimates of macrozoo- (ciliates and zooflagellates) comprised a large plankton clearance rates on nano- and mi- portion of the C ingested by estuarine co- croprotozoa determined from our experi- pepods (Gifford and Dagg 1988) and that Zooplankton-protozoan coupling 1341

MICROPROTOZOA 0.2 0.4 0 A. 29 March 0

-0 2’0 4’0 60 8’0 100 0 40 80 120 160 0.2 0.5 0 B. I9 April E. 28 August 0. I - 0 0 O- -O.I- 0 % -0.2 I I I I I 0. I a . 0 IO 20 3040 5060,,0 I C. I3 June F. 4 October I I

0 2.0 4.0 6’0 a0 lb0 @O 0 IO0 200 300 ZOOPLANKTON BIOMASS (pg liter-l,dry wt) Fig. 3. As Fig. 2, but for microprotozoan growth rates. heterotrophic flagellates could be important (Stoecker and Egloff 1987). Also, Cladocera food for copepods, particularly at lower such as Daphnia were found to grow and phytoplankton densities (Roman et al. reproduce on a diet composed entirely of 1988). In addition, a model of grazing con- heterotrophic flagellates (Sanders and Por- trol of phytoplankton in the open subarctic ter 1990), and comparatively, mixed nano- Pacific Ocean could be balanced only by flagellates enhanced the growth and survi- assuming that macrozooplankton are om- vorship of Daphnia over other algal diets nivorous and graze on both phytoplankton (DeBiase et al. 1990). and protozoa (Frost 1987). Although our Selective grazing by zooplankton on pro- data do not demonstrate macrozooplankton tozoa - Although the average clearance rates selection for protozoa over phytoplankton, on individual protozoan taxa showed con- they suggest that protozoa are significant siderable variation, rates were generally prey items in addition to phytoplankton in higher for taxa < 20 pm in equivalent spher- Lake Michigan. ical diameter. Other studies assessing mac- High macrozooplankton clearance rates rozooplankton predation on natural plank- on protozoa might be a function of their ton assemblages reported similar results in high food quality (Stoecker and Capuzzo that adult copepods feed most readily on 1990). Consistent with this idea, a diet of cells 3-7 pm in size (Nival and Nival 1976). ciliates similar to some of those occurring Moreover, freshwater plankton in the 3-30- in Lake Michigan (Strombidium and Uro- pm size range are preferred by Calanoid co- tricha) was shown to enhance egg produc- pepods (Vanderploeg 198 1) and filter-feed- tion of the Calanoid copepod Acartia tonsa ing Cladocera (Gliwicz 1980). The relative 1342 Carrick et al.

Table 2. Examples of weight-specific clearance rates [ml&g dry wt -‘)d ‘1 of macrozooplankton on 16 common protozoa from the surface wal.ers of Lake Michigan in 1989. ~---_____------~-I_-- --

Taxon ESD* 29 Mar 19 Apr 13 Jun 10 Jul 28 Aug 4 act Mean ~------__--__- --..------Hnano Chromulina sp. 1 2.5 3.3 10.3 3.2 4.2 8.0 2.1 5.2 Chromulina sp. 2 2.7 5.7 - 5.1 - 3.9 - 4.9 Ochromonas sp. 2.7 - 0.1 4.1 3.2 15.6 1.0 4.8 Katablepharis ovalis 6.2 - - 0 3.4 3.8 3.4 2.7 Pnano Chrysochromulina parva 3.7 - 0 1.8 4.2 11.8 1.9 3.9 Ochromonas sp. 3 6.3 - - 0 4.3 7.2 1.6 3.3 Rhodomonas minuta 7.0 1.9 3.5 0 6.0 6.3 2.5 3.4 M icroflagellates Gymnodinium varians 11.6 2.2 5.4 - 5.7 8.1 1.6 4.6 Gymnodinium helveti- cum 27.5 0 0 1.1 - - - 0.4 Ceratium hirudinella 44.8 - - - 0 1.7 0 0.6 Ciliates Strobilidium sp. 12.5 0.3 4.1 9.2 5.0 7.7 1.2 4.6 Urotricha sp. 14.0 0.3 1.5 5.9 6.3 4.1 1.9 3.3 Ilalteria sp. 18.0 - - 3.8 - - 1.0 2.4 Vorticella sp. 25.4 - - - 0 0 0 0 Strombidium viride 35.7 - - 2.6 11.9 - 0.7 2.5 Strobilidium velox 36.5 0 - - 7.6 - 0 2.5

* Equivalent spherical diameter (pm). consistency in this pattern of protozoa mor- gut passage. Second, the size range of the tality in our experiments occurred despite protozoa encountered here was much small- large changes in macrozooplankton com- er than the distribution of ungrazed phy- munity structure (O-69% Daphnia domi- toplankton reported by Lehman and Sand- nance) and thus tends to downplay the gren ( 198 5) and conformed to the size and importance of consumer community com- shape of grazed phytoplankton from their position in determining the fate of micro- studies. Third, the protozoa observed here bial production, although such alterations did not possesselaborate morphologies, such in consumers are influential for macrozoo- as spines or extending appendages that are plankton grazing on bacteria in some eco- common among algal taxa (see Lee et al. systems (Pace et al. 1990). 1985). However, Vorticella sp. was one of Unlike investigations of macrozooplank- the few protozoan taxa resistant to grazing ton grazing on phytoplankton (e.g. Lehman on all dates that it was encountered. As has and Sandgren 1985), we did not observe a been reported elsewhere (Pratt and Rosen positive relationship between macrozoo- 1983; Carrick and Fahnenstiel 1990), this plankton biomass and growth for any pro- taxon can grow attached to colonies of Alz- tozoan taxon, and only one taxon was con- abaena flos-aquae (Lyngb.) deBribisson as sistently unaffected by zooplankton grazing. was noted here, which apparently provided These results suggest basic differences be- it refuge from potential grazers; colonies tween the two protistan groups with respect were large enough (> 200~pm diam) to pre- to morphological adaptations and nutri- vent handling by zooplankton (Lehman and tional considerations. Sandgren 198 5). First, unlike ‘many a.lgal species (Porter Last, many of the protozoa we encoun- 1976), most planktonic protozoa do not tered were bacterivorous flagellates and cil- possess thick gelatinous sheaths or persis- iates, whose growth may not be directly tent cell walls that might afford them viable augmented by nutrients excreted by mac- Zooplankton-protozoan coupling 1343 rozooplankton (Taylor and Lean 198 l), in contrast with phytoplankton (Lehman 1980). Thus, our results suggest that plank- tonic protozoa in Lake Michigan are dom- inated by small (<50+m diam) unicellular forms that are susceptible to zooplankton grazing in that they generally do not possess morphological adaptations to afford resis- tance. Although protozoa may possess be- havioral adaptations to avoid grazing, as 0 indicated by the habits of Vorticda sp., 0 10 20 30 40 50 those qualities were not assessedhere. Quantitative importance of the zooplank- EQUIVALENT SPHERICAL ton-protozoan trophic link-The rate of DIAMETER (pm) phototrophic (Pnano and microflagellates) Fig. 4. Correlation between the average zooplank- and heterotrophic (Hnano and ciliates) pro- ton clearance rates [ml (pg dry wt)-’ d-l] and the cell tozoan C consumption by macrozooplank- size of 16 dominant protozoan taxa. ton is similar to rates determined for Great Lakes phytoplankton (see Table 2). The ex- isting conceptual model of trophic struc- ton; second, the bulk of heterotrophic nano- ture, at least in Lake Michigan, must be flagellate C appears to be consumed by mac- modified to account for these observations. rozooplankton rather than smaller grazers Our estimates of heterotrophic protozoan C such as ciliates. flux to macrozooplankton (0.3-l 1.2 pg C The flux of C between heterotrophic pro- liter-l d-l) are comparable to flux rates of tozoa and macrozooplankton has important phytoplankton to zooplankton determined implications for the fate of bacterial pro- previously (from 2.1 to 15.8 pg C liter-’ d-l) duction in Lake Michigan. Heterotrophic in the surface waters of Lake Michigan (Sca- protozoan (Hnano and ciliates) production via and Fahnenstiel 1987). The flux of (mean 10.8 pg C liter-’ d-l, Carrick 1990) Hnano C alone accounted for > 50% of the can account for > 100% of bacterial pro- total heterotrophic transfer to zooplankton duction (mean, 25.7 pg C liter-’ d-l; Scavia on most dates (mean, 2.1; range, 0.3-6.0 pg and Laird 1987), if we assume that these C liter-l d-l) and seems to indicate a tight heterotrophic protozoa are bacterivores with coupling between Hnano and zooplankton a 30% growth efficiency. This observation, as these flux estimates are comparable (on in concert with the high average flux of het- average 75%) to Hnano productivity values erotrophic protozoa to macrozooplankton (mean, 2.7; range, 0.9-5.7 pg C liter-’ d-l, (mean, 3.9 pg C liter-’ d-l), indicates that Carrick 1990). 15% of bacterial production in the lake can In addition, phototrophic protozoan be consumed by macrozooplankton through (Pnano and microflagellates) flux rates (0. l- a simple two-step trophic coupling. This es- 24.4 pg C liter-l d-l) were similar to those timate is probably conservative, as it does for algae, which suggests that the bulk of not account for direct bacterivory by mac- phototrophic C consumed by macrozoo- rozooplankton or mixotrophy among fla- l plankton in Lake Michigan is composed of gellates, which both can be significant (e.g. flagellates, while a small fraction is com- Sanders et al. 1989). Although this com- posed of green algae, blue-green cyanobac- parison is preliminary, our results demon- teria, diatoms, and picoplankton. Thus, our strate the importance of the microbial food results differ from contemporary food-web web in Lake Michigan with implications for models in other freshwater systems (e.g. trophic structure in other aquatic systems. Weisse et al. 1990) in two ways. First, we The trophic importance of protozoa is il- observed comparable macrozooplankton lustrated further by the substantial fraction grazing on both protozoa and phytoplank- of potential macrozooplankton annual pro- 1344 Carrick et al.

duction that can be supported by ingestion Huron and Michigan. Can. J. Fish. Aquat. Sci. 46: of protozoan C. If we assume production 1922-1928. -, AND -. 1990. Planktonic protozoa in estimates (Borgmann et al. 1984) for Lake Lakes Huron and Michigan: Seasonal abundance Ontario macrozooplankton (- 2.0 pg C li- and composition of ciliates and dinoflagellates. J. ter-l d-l) are in the range for production Great Lakes Res. 16: 3 19-329. expected in Lake Michigan (macrozoo- CHOI, J. W., AND D. K. STOECKER. 1989. Effects 01 plankton production has not been estimated fixation on cell volume of marine planktonic pro- tozoa. Appl. Environ. Microbial. 55: 1761-1765. for Lake Michigan) based on the similarities DEBIASE, A. E., R. W. SANDERS, AND K. G. PORTER. between lakes in terms of zooplankton com- 1990. Relative nutritional value of ciliate pro- position and abundance, grazing on proto- tozoa and algae as food for Daphnia. Microb. Ecol. zoa by macrozooplank.ton can support 19: 199-210. DORAZIO, R. RI., J. A. BOWERS, AND J. T. LEMMAN. >80% of annual macrozooplankton pro- 1987. Food-web manipulations influence grazer duction (assuming 15% zooplankton assim- control of phytoplankton growth rates in Lake ilation efficiency). These estimates indicate Michigan. J. Plankton Res. 9: 891-899. that protozoa are likely to be quantitatively FAHNENSTIEL,G. L.,AND H.J. CARRICK. 1991. Pho- significant prey for epilimnetic macrozoo- totrophic picoplankton in Lakes Huron and Mich- igan: Abundance, distribution, composition, and plankton in Lake Michigan. contribution to biomass and production. Can. J. Protozoa seem to be significant prey items Fish. Aquat. Sci. 48: In press. for macrozooplankton in Lake Michigan, as -- --, AND R. ITURRIAGA. 1991. Physio- clearance rates are similar to those previ- loiical characteristics and food-web dynamics of Synechococcus in Lakes Huron and Michigan. ously determined for phytoplankton, par- Limnol. Oceanogr. 36: 2 19-234. ticularly for protozoa < 20 pm in equivalent -, AND D. SCAVIA. 1987. Dynamics of Lake spherical diameter. Also, a large fraction of Michigan phytoplankton: The deep chlorophyll potential annual zooplankton production in layer. J. Great Lakes Res. 13: 285-295. the lake can be accounted for by zooplank- FROST, B. W. 1987. Grazing control ofphytoplankton slack in the open subarctic Pacific Ocean: A model ton grazing on protozoa. This observation, assessing the role of mesozooplankton, particu- as well as the fact that the flux of heterotro- larly the large Calanoid copepods Neocalanus spp. phic protozoan C to macrozooplankton is Mar. Ecol. Prog. Ser. 39: 49-68. comparable to that for phytoplankton, is GIFFORD, D. J., AND M. J. DAGG. 1988. Feeding of contrary to contemporary pelagic food-web the estuarine copepod Acartia tonsa Dana: Car- nivory vs. herbivory in natural microplankton as- models (e.g. Weisse et al. 1990). These find- semblages. Bull. Mar. Sci. 43: 458468. ings strongly imply that ,picoplankton C in GLIWICZ, Z. M. 1980. Filtering rates, food size selec- Lake Michigan may be more directly trans- tion, and feeding rates in cladocerans: Another ferred to higher trophic levels via a pico- aspect of interspecific competition in filter-feeding zooplankton. Am. Sot. Limnol. Oceanogr. Spec. plankton-flagellate-zooplankton coupling, Symp. 3: 282-29 1. New England. compared with the less direct picoplankton- HAAS, L. W., AND K. L. WEBB. 1979. Nutritional flagellate-ciliate-zooplankton coupling pro- mode of several non-pigmented microflagellates posed for oligotrophic marine systems from the York River estuary. J. Exp. Mar. Biol. (Sheldon et al. 1986). Ecol. 39: 125-134. HAWKINS, B. E., AND M. S. EVANS. 1979. Seasonal cycles of zooplankton biomass in southeastern Lake References Michigan. J. Great Lakes Res. 5: 256-263. BORGMANN, U., H. SHEAR,AND.~. MOORE. 1984. Zoo- HEALY, F. P. 1979. Short-term responses of nutrient- plankton and potential fish production in Lake deficient algae to nutrient addition. J. Phycol. 15: Ontario. Can. J. Fish. Aquat. Sci. 41: 1303-l 309. 289-299. CARON, D. A. 1983. Techniq,ue for enumeration of KLEPPEI,, G. S., D. FRAZEL, R. E. PIEPER, AND D. V. heterotrophic and phototrophic nanoplankton, us- HOLLIDAY. 1988. Natural diets of zooplankton ing epifluorescence microscopy, and comparison off southern California. Mar. Ecol. Prog. Ser. 49: with other procedures. Appl. Environ. Microbial. 231-241. 46: 49 l-498. LAWS, E. A., AND OTHERS. 1984. High phytoplankton CARRICK, H. J., JR. 1990. Planktonic protozoa in growth and production rates in oligotrophic Ha- Lake Michigan: Distribution, production, and fate. waiian coastal waters. Limnol. Oceanogr. 29: 116l- Ph.D. thesis, Univ. Michigan. 140 p. 1169. --- ,ANDG.L.FAI~NENSTIEI.. 1989. Biomass,size LEE, J.J .,S. H. HUTNER,AND E.C. BOVEE[EDS.]. 1985. structure, and composition of phototrophic and Illustrated guide to the protozoa. Sot. Protozool. heterotrophic nanoflagellate communities in Lakes LEHMAN, J. T. 1980. Release and cycling of nutrients Zooplankton-protozoan coupling 1345

between planktonic algae and herbivores. Limnol. AND G. A. LAIRD. 1987. Bacterioplankton in Oceanogr. 25: 620-632. dke Michigan: Dynamics, controls, and signifi- - AND C. D. SANDGREN. 1985. Species-specific cance to carbon flux. Limnol. Oceanogr. 32: 10 19- raies of growth and grazing loss among freshwater 1035. algae. Limnol. Oceanogr. 30: 34-46. SCHELXE, C. L., E. F. STOERMER, G. L. FAHNENSTIEL, NIVAL, P., AND S. NIV&. 1976. Particle retention of AND M. HAIBACH. 1986. Phosphorus enrich- an herbivorous copepod, Acartia cluusi (adult and ment, silica utilization, and biogeochemical silica copepodite stages): Effects on grazing. Limnol. depletion in the Great Lakes. Can. J. Fish. Aquat. Oceanogr. 21: 24-38. Sci. 43: 407-4 15. PACE, M. L., G. B. MCMANIJS, AND S. E. G. FINDLAY. SHELDON, R. W., P. NIVAL, AND F. RASSOULZADAGEN. 1990. Planktonic community structure deter- 1986. An experimental investigation of a flagel- mines the fate of bacterial production in a tem- late-ciliate-copepod food chain with some obser- perate lake. Limnol. Oceanogr. 35: 795-808. vations relevant to the linear biomass hypothesis. -, K. G. PORTER, AND Y. S. FEIG. 1983. Species- Limnol. Oceanogr. 31: 184-188. and age-specific differences in bacterial resource SHERR, E. B., B. F. SHERR, AND L. J. ALBRIGHT. 1987. utilization by two co-occurring cladocerans. Ecol- Bacteria: Link or sink? Science 235: 88-89. ogy 64: 1145-l 156. STOECKER, D. K., AND J. M. CAPUZO. 1990. Pre- PAFFENH~FER, G-A., AND S. C. KNOWLES. 1980. Om- dation on protozoa: Its importance to zooplank- nivorousness in marine planktonic copepods. J. ton. J. Plankton Res. 12: 891-908. Plankton Res. 2: 355-365. - AND D. A. EGLOFF. 1987. Predation by Acar- PORTER, K. G. 1976. Enhancement of algal growth tia’ tonsa Dana on planktonic ciliates and rotifers. and productivity by grazing zooplankton. Science J. Exp. Mar. Biol. Ecol. 110: 53-68. 192: 1332-l 334. STRATHMANN, R. R. 1967. Estimating the organic PRATT, J. R., AND B. H. ROSEN. 1983. Associations carbon content of phytoplankton from cell volume of species of Vorticellu (Peritrichida) and plank- or plasma volume. Limnol. Oceanogr. 12: 41 l- tonic algae. Trans. Am. Microsc. Sot. 102: 48-54. 418. ROMAN, M. R., AND OTHERS. 1988. Production, con- TAYLOR, W. D., AND D. R. S. LEAN. 198 1. Radio- sumption and nutrient cycling in a laboratory me- tracer experiments on phosphorus uptake and re- socosm. Mar. Ecol. Prog. Ser. 42: 39-52. lease by limnetic microzooplankton. Can. J. Fish. SANDERS, R. W., AND K. G. PORTER. 1990. Bacteri- Aquat. Sci. 38: 13 16-l 321. vorous flagellates as food resources for the fresh- VANDERPLOEG, H. A. 1981. Seasonal particle selec- water crustacean zooplankter Daphnia umbiguu. tion by Diuptomus sicilis in offshore Lake Mich- Limnol. Oceanogr. 35: 188-l 9 1. igan. Can. J. fish. Aquat. Sci. 38: 504-5 17. - - S. J. BENNETT, AND A. E. DEBIASE. WEISSE, T., AND OTHERS. 1990. Response of the mi- 1989. Seakonal patterns of bacterivory by flagel- crobial loop to the phytoplankton spring bloom lates, ciliates, rotifers, and cladocerans in a fresh- in a large prealpine lake. Limnol. Oceanogr. 35: water planktonic community. Limnol. Oceanogr. 78 l-794. 34:673-687. SCAVIA, D., AND G. L. FAHNENSTIEL. 1987. Dynamics of Lake Michigan phytoplankton: Mechanisms Submitted: 15 February 1991 controlling epilimnetic communities. J. Great Accepted: 7 May 1991 Lakes Res. 13: 103-l 20. Revised: 23 July 1991