The Eastern Oyster Crassostrea Virginica

MARINE ECOLOGY PROGRESS SERIES Vol. 78: 285-301,1991 1 Published December 23 Mar. Ecol. Prog. Ser. - I P P P - --

Omnivorous feeding by planktotrophic larvae of the eastern oyster Crassostrea virginica

Brad S. Baldwin, Roger I. E. Newel1

Center for Environmental and Estuarine Studies. Horn Point Environmental Laboratory, University of Maryland System, PO Box 775, Cambridge, Maryland 21613, USA

ABSTRACT: In order to better understand the particle diet of planktotrophic larvae of the eastern oyster Crassostrea virginica (Gmelin) we measured their ingest~on of naturally occurring food organisms. By using a dual radioisotope (3Hand I4C)label~ng technique in conjunction w~thplankton size fractionation procedures we demonstrate that oyster larvae feed upon bacteria, phagotrophic protozoans and phototrophs present in the diverse summer plankton assemblages of Chesapeake Bay, USA. Prodissoconch I1 oyster larvae cleared 0.2 to 30 pm I4C-labeledplankton (primarily phototrophs) at a rate of 0.0825 n~llarva-' h-', and 0.2 to 30 pm 3H-labeled plankton (heterotrophlc bacteria and phagotrophic protozoans) at a rate of 0.0017 m1 larva-' h-' This calculated clearance rate for 0.2 to 30 pm heterotrophs was low due to the predominance of small (0.2 to 0.8 pm), poorly retained bacteria in thls size class. Oyster larvae consumed a wide size range of food part~cles(0 2-0.8 pm to 20-30 pm) and selectively ingested 20 to 30 pm organisms. In other feeding experiments, oyster larvae cleared laboratory cultured heterotrophic flagellates (12 pm) at a rate of 0.0640 m1 larva-' h-' and cultured heterotrophic ciliates (12 X 20 pm) at a rate of 0.1093 m1 larva-! h" The inclusion of heterotrophic food organisms In the diet of C. virginlca may enhance its growth and development by providing energy and nutrients that supplement those of ingested phytoplankton. We suggest that because oyster larvae ingest non-phytoplankton cells, estimates of standing stocks of phytoplankton may not always be a reliable measure of food supply

INTRODUCTION Crassostrea virginica (Gmelin) fed upon a variety of phytoplankton taxa present in Delaware Bay, USA. To Planktotrophic larvae of benthic marine inverte- date, however, no attempt has been made to determine brates must acquire adequate nutrition from their whether bivalve larvae also ingest non-phytoplankton environment to support their development and food organisms (e.g. heterotrophic bacteria, phago- metamorphosis to the benthic life stage. For bivalve trophic protozoans) that are abundant in coastal molluscs, a dominant taxonomic and functional group systems (e.g. Ducklow & Kirchman 1983, Davis et al. in many benthic systems, little is known concerning 1985, Malone et al. 1986, Geider 1988, Stoecker et al. the natural particle diets of their dispersive larvae. 1989, Dolan & Coats 1990, Malone & Ducklow 1990, Consequently, current understanding of the effects of McManus & Fuhrman 1990). food availability and nutritional quality on larval Previous studies on the feeding and growth of growth, survival, and n~etamorphosishas been derived planktotrophic larvae of bivalves and other taxonomic largely from laboratory studies that have used simpli- groups have suggested that non-phytoplankton foods fied unialgal diets (for reviews see Bayne 1983, Webb may comprise part of their natural diet. For example, & Chu 1983, Pechenik 1987, Strathmann 1987). How- Crisp et al. (1985) considered the standing stocks of ever, In the natural environment, larvae encounter 2 to 10 pm long phytoflagellates to be insufficient to more complex and variable assemblages of food support the growth of larvae of the bivalve Ostrea particles. In one of the few studies that examined larval edulis. As a supplement to these algae, Crisp et al. feeding in natural particle assemblages, Fritz et al. (1985) speculated that 0. edulis larvae feed upon (1984) found that planktotrophic larvae of the oyster smaller algae, dissolved organic compounds, and

O Inter-Research/Printed in Germany 0171-8630/91/0078/0285/$03.00 286 Mar Ecol Prog.Ser.

possibly bacteria and detritus. In studies with echino- MATERIALS AND METHODS derm larvae, Olson (1987) and Olson et al. (1987) measured development and survival of the asteroids Preparation of feeding treatments. We conducted Acanthaster planci and Odontaster- vahdus reared in preliminary feeding experiments using fluorescent situ in coastal waters of Australia and Antarctica, microspheres (Polysciences, Inc.) to determine the respectively. In both study sites, standing stocks of range of particle sizes ingested by larvae of the eastern phytoplankton were considered to be ~nsufficientto oyster Crassostrea virginica (Gmelin) and, hence, to support larval food requirements. However, Olson establish the size range of natural plankton to offer (1987) found that A. planci developed at near maximal larvae in feeding experiments (described below). rates and Olson et al. (1987) reported 92.2 % survivor- Microspheres of diameter 0.21, 0.52, 0.94, 1.97, 3.46, ship for 0. validus. Based on these results, the authors 4.80, 8.2, 15.8, 21.1, or 27.4 pm were tested in separate speculated that dissolved organic compounds and non- feeding trials. Microspheres were suspended in phytoplankton food particles may have been utilized deionized water, sonicated to break apart aggregates, by these larvae. Utilization of non-phytoplankton foods and added to glass vials containing larvae and filtered has been confirmed in other studics. For example, (0.2 pm) autoclave-sterilized estuarine water (SEW). Rivkin et al. (1986) found that Antarctic species of After the feeding trials, larvae were examined using an echinoderm, polychaete, and nemertean larvae fed epifluorescent microscope for the presence of ingested extensively on natural populations of bacteria. More microspheres. In general we could detect ingestion of recently, Qian & Chia (1990) reported that 3 species of microspheres 5 4.80 pm in feeding experiments lasting polychaete larvae grew on detritus collected from I 10 min; however, experiments lasting up to 4 h were seagrass beds and concluded that detritus may play a necessary for microspheres 2 8.2 pm because lower supplementary role in larval nutrition. microsphere concentrations were used. Negligible Many holoplanktonic suspension feeders are reaggregation of microspheres in SEW was detected known to feed as omnivores, even jn environments under our experimental conditions. These experiments where phytoplankton is abundant. For example, demonstrated that larvae up to ca 300 pm shell length several species of copepods ingest attached bacteria, could ingest microspheres as small as 0.21 um. We also detritus, and microzooplankton in addition to phyto- found that the maximum particle sizes ingested by plankton (Berk et a1 1977, Landry 1981, Roman larvae ca 150, 180, and 300 ,um in shell length were 1984a, b, Conley & Turner 1985, Stoecker & Sanders 15.8, 21.1 and 27.4 pm, respectively. 1985, Stoecker & Egloff 1987, Gifford & Dagg 1988, Plankton samples from surface (i 1 m) waters of the Kleppel et al. 1.988). Omnivory appears to confer Choptank River sub-estuary of Chesapeake Bay were several potential benefits, including: (1) ingestion of collected in 20 l carboys and returned to the laboratory. more energy and nutrients per unit of feeding effort Based on the particle size ingestion experiments (Gifford & Dagg 1988); (2) acquisition of sufficient above, we filtered plankton samples through a 30 pm ration despite a fluctua.ting balance between auto- Nitex screen using reverse flow gravity filtration in trophic and heterotrophic food organisms In the order to provide the most 1ikel.y size range of larval natural environment (Landry 1981, Kleppel et al. food organisms and to exclude metazoan grazers. The 1988); and (3) enhanced growth on a mixed diet of < 30 pm filtrate wa.s placed into several l 1 poly- plant and detrital (Roman 1984a) or plant and animal carbonate bottles and dual-labeled (Roman & Rublee foods (Stoecker & Egloff 1987). 1981) with filter sterilized (0.2 pm) "C-bicarbonate In this study we examine the natural diet of (NaH14C03,speclfic activi.ty = 58 mCi mmole-l, JCN Crassostrea virqinica larvae in the mesohaline (ca 7 to Biomedicals, Inc.) and tritiated-thymidine (methyl- 15 ppt) portion of Chesapeake Bay, USA. Specific I3H]- thymidine, sp~ciflcactivity = 60 Cl mmole ', ICN questions we address are: (1) Do C. virginica larvae Biomedicals, Inc.),each at a final concentration of 0.5 feed as omnivores? (2)What is the range of cell sizes ~CIml-' The final concentration of thymidine was 8.3 ingested by oyster larvae? To answer these questions nM. Bottles were incubated for 120 min at 25 "C over we used a dual radioisotope labeling technique and fluorescent light at an intensity of 200 pE m-' s and plankton size fractionation procedures to differentially Itrere swirled every 10 to 15 min in order to maintain label a wide size range of naturally occurring photo- well-mixed suspensions. trop'hic and heterotrophic food organisms and to Our incubation procedure directly labeled photo- provide a direct measure of particle ingestion by trophic organisms (phytoplankton and mixotrophic larvae. Our results indicate that C. virginica larvae protozoans) with ''C-bicarbonate and heterotrophic feed as omnivores and that they can capture and ingest bacteria (free-living and attached forms) with %-thy- food organisms ranging in size between 0.2 to 0.8 and midine. Studies have shown that eucaryotic micro- 20 to 30 pm. algae (e.g. Sagan 1965, Swinton & Hanawalt 1.972, Baldw~n& Newell:Omnivory by oyster larvde 287

Pollard & Moriarty 1984, Rivkin 1986, Rivkln & Voytek ingestion of heterotrophic bacteria and phagotrophic 1986, Martinez et al. 1989) and protozoans (e.g.Plaut & protozoans. Sagan 1958, Scherbaum & Luoderback 1960) can also Following the 120 min prelabeling period, the incorporate %I-thymidine, but these results were contents of the incubation bottles were pooled and derived from experiments that used either unspecified serially size fractioned through filters in order to thymidine concentrations and incubation periods, or establish the following feeding treatments: (1) the 0.2 higher thymidine concentrations (20 to 150 nM) and/or to 30 pm size class (no filtration), (2) the 0.2 to 20 pm longer incubation periods (up to 48 h) than we size class, (3) the 0.2 to 10 pm size class, (4) the 0.2 to 3 employed in the present study. Incorporation of %I- pm size class. and (5) the 0.2 to 0.8 pm size class. thymidine into cultured cyanobacteria (under similarly Plankton was gravity filtered through 20 pm Nitex high thymidine concentrations) has been documented screens and 10 pm Nuclepore filters and vacuum by Martinez et al. (1989) but not In other studies (Cuhel filtered (1 100 mm Hg pressure) through 3 pm and & Waterbury 1984, Pollard & Moriarty 1984). 0.8 pm Nuclepore filters. The lower size limit of all Significantly, however, in experiments using natural feeding treatments was considered to be 0.2 pm given plankton assemblages and low concentrations of the ability of larvae to ingest particles of this size (see added %I-thymidine (5.0 nM), Fuhrman & Azam (1982) above). Immediately before use in feeding experi- showed that %I-thymidine incorporation is almost ments, aliquots were removed from each of the treat- completely specific to nonphotosynthetic bacteria. ments and filtered through 0.2 pm Nuclepore filters. Using microautoradiography, they demonstrated that Filters were rinsed with SEW and fumed over acetic < 1 % of exposed silver grains were associated with acid to remove residual l4C-bicarbonate as ''CO2. organisms other than bacteria in natural plankton Filters were then placed in vials with scintillation cock- samples that had been labeled for up to 12 h. Similar tail (EcoLite(+), ICN Biomedicals, Inc.) and radio- results have been obtained in a number of other activity of filtered particles was determined on a studies that employed microautoradiography to Packard Tri-Carb liquid scintillation counter (model examine 3H-thymidine incorporation into natural 2260XL) which used an external standard and quench plankton assemblages (Bern 1985, Fuhrman et al. correction curve to calculate dpm. Results were 1986, Douglas et al. 1987) and sedimentary micro- expressed as dpm ml-' of suspension for each of the organisn~s (Carman 1990). Although the likelihood above size classes. The distribution of particulate label that planktonic organisms will incorporate 'H-thymi- among small and large plankton cells was determined dine is undoubtedly species-, time-, and location- by taking the difference between size class radio- specific, on the basis of the results from the above activity. This provided estimates of dpm ml-' of autoradlographic studies, we assumed (and later suspension in the 0.2 to 0.8 pm, 0.8 to 3 pm, 3 to 10 pm. confirmed, see 'Discussion') that under our incubation l0 to 20 pm, and 20 to 30 um size fractions. conditions, negligible amounts of %thymidine would The particle composition of the above feeding treat- be incorporated by organisms other than heterotrophic ments was determined using epifluorescent micro- bacteria. scopy. Between 2 and 10 m1 of each size class were Phagotrophic protozoans can become indirectly preserved in 1 % gluteraldehyde (final concentration) labeled with both 14C (via the consumption of I4C- and stained (if needed) for 4 min. Samples were then labeled phototrophs) and 3H (via the consumption of filtered onto irgalan black-stained Nuclepore filters 3H-labeled bacteria). A 120 min prelabeling period and mounted on slides between drops of 011. was chosen to ensure that plankton food cells were of Heterotrophic (i.e. non-photosynthetic) bacteria and high specific activity, thereby facilitating detection of protozoans were viewed under blue light excitation ingested cells in short-term (10 min) larval feeding after one of 2 staining treatments: (1) acridine orange experiments (described below). In addition, on the staining for enumeration of bacteria (Hobbie et al. basis of reported ingestion rates for phagotrophic 1977), or (2) proflavine staining for enumeration of protozoans (e.g. Lessard & Swift 1985, McManus & protozoans (Haas 1982). For both of these organisms, Fuhrman 1988, Sherr et al. 1988, Sherr et al. 1989) we only recognizable and stained cells that did not auto- estimated that an incubation period of 120 min would fluoresce were counted. Heterotrophs were cate- be adequate to label phagotrophic protozoans present gorized as bacteria (free-living or attached), flagel- in the <30 pm filtrate. Based on our labeling proce- lates, or ciliates. Autofluorescent phototrophs (no dure, we interpret the ingestion of l%-labeled orga- stains used) were examined under blue or green light nisms by larvae to represent consumption of photo- excitation. Phototrophs were categorized as cyano- trophs and possibly phagotrophic protozoans. On the bacteria, large dinoflagellates (i.e. > 10 pm), or other other hand, we assume that the consumption of parti- phototrophs (i.e. all other autofluorescent cells). The culate forms of 3H by oyster larvae represents 'large dinoflagellate' category was used in order to 288 Mar Ecol. Prog Ser

document the abundance of the dinoflagellates Gyro- by larvae from each size fraction was calculated by dinium uncatenum and Gymnodiniurn sanguinium, difference. Ornnivory was judged as the ability of which were undergoing a population bloom at the time larvae to ingest both "C-labeled organisms (photo- of this experiment. Prepared slides were stored in the trophs and possibly phagotrophic protozoans) and dark at -20 OC to preserve pigment fluorescence and 3H-labeled organisms (heterotrophic bacteria and/or were viewed within 1 mo. phagotrophic protozoans). Cell counts were determined for the 0.2 to 0.8 pm, Clearance rates of food organisms. Larval clearance 0.8 to 3 pm, 3 to 10 pm, and 10 to 20 pm size fractions rates of the different sizes and types of planktonic by filtering samples from the 0.2 to 0.8 pm, 0.2 to 3 pm, organisms consumed in the above feeding experi- 0.2 to 10 pm, and 0.2 to 20 pm size classes through 0.2, ments were calculated according to an equation 0.8, 3, and 10 pm Nuclepore filters, respectively. Cells similar to that developed by Daro (1978) and modified counts for the 20 to 30 pm size fraction were deter- by Baars & Oosterhuis (1984): mined by difference. Here, samples of the 0.2 to 30 Km CR (m1 larva-'h-') = size class were filtered through 10 pm filters, cells were counted, and values from the 10 to 20 pm fraction (dprn, - dpmd at time t) (1 were subtracted away, thereby providing counts for dpm per m1 dpm per m1 the 20 to 30 pm fraction. of suspension + of suspension Test of omnivory. Preceding the radiolabeling at time 0 at time t procedure mentioned above, prodissoconch 11 larvae were acclimated to unlabeled < 30 pm plankton for 2 d. where dprn, = dprn larva-' held in labeled plankton; Larvae were reprovisioned with freshly collected dpmd = dprn larva-' of dissolved controls: t = time of plankton each day. After this acclimation period, and feeding experiment in hours. The denominator of the after plankton assemblages were labeled as described equation provides an 'average' specific activlty for the above, 10 of these larvae were placed in a bottle (3 suspension since planktonic food organisms continue replicate bottles per treatment) along with 100 m1 of a to incorporate label during the 10 min feeding period. glven labeled plankton size class. After a 10 min Radioactivity of larvae (dpm larva-') was determined grazing period both the larvae and plankton were as previously described and the radioactivity of the poured over 95 pm Nitex screens in order to separate labeled plankton suspension (dpm ml-') in each larvae from labeled plankton. Aliquots of the filtrate feeding treatment was measured before and after the were taken to determine dprn ml-' of suspension. In 10 min feeding period (see above). this way we could account for changes in the radio- Underestimation of CR due to defecation of con- activity of feeding treatments over the 10 min grazing sumed label was minimized by using a feeding time of period (see below). Larvae retained on screens were l0 min which is within the 5 to 10 min gut passage time rinsed with SEW and then fumed over acetic acid to of Crassostrea virginica larvae, as determined in preli- remove residual I4C-bicarbonate as I4COz. Larvae minary experiments. Similar experiments also con- were rinsed off the screens and placed into liquid scin- firmed that these larvae release ingested label in tillation vials where they were counted and measured soluble form after feeding periods of > 10 min; there- on an inverted microscope. The mean shell length of fore, underestimation of CR due to the loss of con- the prodissoconch I1 larvae was 179 pm (f 5 pm, SD). sumed label via excretion, respiration or 'leakage' Larval tissues were then homogenized via sonication (Pechenik 1979) was minimized as well. Overesti- (Branson Sonifier 450) and chemical digestion mation of CR was corrected using previously described (Packard Soluene 350). Radioactivity of samples was exper~ments controling for uptake of radiolabeled determined and expressed as dprn larva-'. Because dissolved compounds. feeding treatments contained both radiolabeled Ingestion of phagotrophic protozoans. We com- particles and dissolved compounds, it was necessary to pared larval ingestion of 2 size classes (0.2 to 3 pm and run simultaneous dissolved uptake experiments using 0.2 to 30 pm) of organisms prelabeled with 3H-thymi- the 10.2 pm filtrate from the labeled plankton suspen- dine for 2 periods of time (15 and 120 min) in order to sion In this way we accounted for adsorption of d~s- ascertain larval ingestion of natura.11~ occurring solved label onto larvae as well as absorption by phagotrophic protozoans. Experiments using 120 min larvae, their associated gut flora, and/or attached cells. labeled plankton are described above; those using 15 Estimates of particle ingestion by larvae were calcu- min labeled plankton were conducted in the same lated as the difference between total label uptake by fashion. Based on results of the published 'H-thymi- larvae in each plankton size class (uptake of particles dine labeling studies discussed above, we assumed and dissolved compounds) and uptake of dissolved that during incubations as short as 15 min, negligible compounds. Estimates of the partlculate label ingested amounts of 'H would be taken up by organisms other Baldwin & Newell: Omlnlvory by oyster larvae 289

than heterotrophic bacteria. In contrast, we assumed illatum ml-'. Crassostrea virginica larvae (mean shell that during 120 min incubations, larger organisms such length k SD; 237 f 13 pm) were acclimated to as phagotrophic protozoans would also become unlabeled I. pap~llatum(2 X 104I. papillatum m]-') for labeled but not eucaryotic algae or cyanobacteria. 12 h and then added to the experimental vials at a final Also, microscopic examination showed that hetero- density of 1 larva ml-'. Larvae were allowed to feed on trophs in the 0.2 to 3 pm size class consisted primarily labeled I. papillaturn for up to 60 min. At 5, 15, 30, and of free-living bacteria although relatively small num- 60 min, 5 replicate vials were filtered through 95 pm bers of attached bacteria and flagellates were also Nitex screens and the larvae were processed as de- present (Table 1). Larger protozoans and most of the scribed above. Dissolved control experiments were run using the same volume of 50.2 pm filtrate from the Table 1.Cell composition of plankton size fractions (cells ml-') labeled I. papillaturn culture. Ingestion of I4C-labeled I. papillatum was calculated as larval uptake of I4C in Phototrophs vials containing labeled I. papillaturn and residual S~zefract~on Cyanobacteria Large Others dissolved 14C compounds minus larval uptake of I4C in (pm) (X 103) dinoflagellates (X 103) vials containing residual dissolved 14C compounds. 0.2-0.8 2.13 Larval ingestion rate (IR,cells larva-' h-') was calcu- 0.8-3 143.15 171 lated according to the least-squares regression equa- 3-10 23.70 9 38 tion fitted to the uptake of 14C-labeled I. papillatum by 10-20 3.19 2 7 0.69 20-30 819 0.56 larvae. Larval clearance rate (CR, m1 larva-' h l) was calculated as: CR = IR / I. papillatum ml-l. Heterotrophs Another timecourse ingestion experiment was Size Free Attached Flagellates Ciliates conducted using a bacterized culture of a 12 X 20 pm fraction bacteria bacteria heterotrich ciliate (clone Smcil, isolated from Chesa- (pm) (x103) (x103) (X 103) peake Bay by E. Lessard and tentatively identified as 0.2-0.8 57 10.60 either Diplogrnus sp. or Propygocirrus sp.). Smcil was 0.8-3 518.21 0.76 0.25 incubated with %-thymidine (final concentration 0.5 3-10 0.24 3.73 4.48 pCi ml-l) for 12 h. In this way, the ciliates became 10-20 0.18 4.01 0.09 7 labeled with 3H via phagocytosis of 3H-labeled bac- 20-30 0.12 1.86 0.02 4 teria. An aliquot of this culture was placed in a filter tower over a 5 pm Nuclepore filter with no vacuum attached bacteria were present in the 3 to 30 urn size applied. The suspension was rinsed 5 times with an fraction. This experimental design therefore compared equal volume of SEW in order to retain Q-labeled the following food treatments: (1) the 0.2 to 3 pm size ciliates while allowing passage of dissolved label and class incubated for 15 min, comprised of labeled free- labeled bacteria. The > 5 pm suspension was then living bacteria, small numbers of labeled attached removed and centrifuged for 10 min at 30 X g and the bacteria, and small unlabeled flagellates; (2) the 0.2 to supernatant, which contained residual dissolved label, 30 pm class at 15 min, comprised of labeled free and bacteria, and some ciliates, was removed. Remaining attached bacteria and unlabeled protozoans; (3) the 0.2 ciliates were then resuspended in SEW and the centri- to 3 pm class at 120 min, comprised of labeled bacteria fuge washing procedure repeated. Microscopic obser- and labeled flagellates; and (4) 0.2 to 30 pm class at 120 vation confirmed that nearly all of the remaining min, comprised of labeled free and attached bacteria ciliates were alive and swimming. The resulting ciliate and labeled protozoans. s.uspension was separated into whole water (3H-labeled We also tested the ability of larvae to capture and ciliates, residual bacteria, and residual dissolved label) ingest laboratory cultured heterotrophic flagellates and c 5 pm (residual bacteria and residual dissolved and chates. An axenic culture of the heterotrophic label) size fractions. Sufficient volume of the whole flagellate Isonema papillaturn (8 to 12 pm maximum water fraction was then added to 3 sets of triplicate length) was incubated with 14C-glucose (14C-D- 100 rnl bottles each containing 10 larvae (mean shell glucose, specific activity = 6.12 mCi mmol-', ICN length t SD; 273 t17 pm) and < 30 lm unlabeled Biomedicals, Inc.) at a final concentration of 0.5 pCi plankton such that a final density of 17 labeled ciliates ml-' for 12 h. The labeled culture was then centrifuged ml-l was achieved. The same volume of the c5 pm for 10 min at 30 X g, the supernatant containing fraction was added to a separate set of bottles contain- residual dissolved label was removed, and the culture ing larvae. Larvae were allowed to feed on these mix- was resuspended in SEW. Aliqiiots of the labeled, tures for up to 360 min. At 30,120,and 360 min a set of 3 washed culture were added to 20 m1 vials along with replicate bottles was filtered onto 95 wm Nitex screens SEW to achieve a final cell density of 2.5 X 104I. pap- and larvae processed as described above. Ingestion of 290 Mar. Ecol Prog. Ser

3H-labeled ciliates was calculated as larval uptake of overall the class was dominated by cyanobacteria. The 3H in whole water minus larval uptake of % in the c 5 overall composition of larger size classes was similar pm treatments. Larval ingestion and clearance rates although large numbers of dinoflagellates (primarily were calculated as for Isonema papillaturn. 20 to 30 pm in size), which were undergoing a popula- Statistical analysis. Statistical differences among tion bloom at the time of this experiment, were present treatments were tested using l-way ANOVA. Data in the 0.2 to 30 pm size class. were log-transformed to satisfy the assumptions of In contrast to 14C data, larvae ingested statistically ANOVA Where significant differences were indi- similar amounts of particulate3H from the 0.2 to 0.8pm, cated, log-transformed treatment means were then 0.2 to 3 pm, 0.2 to 10 pm, and 0.2 to 20 pm size classes compared using the Student-Newman-Keuls (SNK) (Fig. 2a). The 0.2 to 3 pm and larger size classes multiple range test. For data presentation, means have contained significantly (p c 0.05) more particulate 'H been back-transformed to the linear scale and 95 % than the 0.2 to 0.8 pm class but were not statistically confidence intervals calculated as estimates of relia- different from each other (Fig 2b). It is apparent from bility (Sokal & Rohlf 1981).All statistical analyses were these results that larvae ingested insignificant quan- conducted using SAS software. tities of Q-labeled 0.8 to 3 pm cells. The amount of particulate 3H that larvae ingested in the 0.2 to 30 pm size class was not statistically different from that RESULTS

Types and sizes of ingested food organisms

Crassostrea virginica larvae ingested I4C- and 3H- labeled organisms and took up dissolved radiolabeled compounds. Total label uptake (i.e. dissolved uptake plus ingestion of labeled particles) was significantly (p c 0.05) greater than dissolved uptake in all feeding treatments, thereby confirming that labeled cells were ingested. Uptake of dissolved radioactive compounds by larvae and/or their associated gut or epibiotic flora accounted for as little as 0.2 % and 5.3 % of total I4C and 3H uptake, respectively, by larvae in the 0.2 to 30 km feeding treatment. However, due to the presence of less particulate label (Figs. lb & 2b), dissolved uptake was as high as 31 and 26 % of total 'vand uptake, respectively, in the 0.2 to 0.8 pm treatment. Larval ingestion of ''C and 3H-labeled organisms (i.e.total label uptake m1nu.s dissolved label uptake) is illustrated in Figs. 1 & 2. Larger plankton size classes contained significantly (p c 0.05) greater amounts of particulate I4C per m1 (Fig. lb) and as a result, larvae generally ingested significantly (p i 0.05) more particulate I4C when fed larger classes of labeled plankton (Fig. la). An exception was the similar amount of ingested particulate I4C from the 0.2 to 3 pm and 0.2 to 10 pm feeding treatments, which suggests that insignificant amounts of 3 to 10 pm '"C-labeled cells wcre ingested. The vast majority (ca 87 %) of ingested ''C- ~~!l.23Q!L22Q~ carbon came from 20 to 30 pm cells (Fig la). This cell Size Class (pm) size fractlon contained a similar amount (22 %) of the Fig. 1. Crassostrea virginica. Larval ingestion and clearance of total particulate 14C as the 10 to 20 pm (25 %), 3 to 10 various size classes of I4C-labeled plankton. (A) dpm of 14C ,urn(22 %,), and 0.8 to 3 ,pm (20 %) fractions (Fig. 3a). ingested per larva; (B)dpm of ''C per m1 of plankton; (C] Cyanobacteria were the only phototrophs detected in clearance rates for plankton size classes. Columns represent means (n = 31, error bars are 95 41 confidence intervals. the 0.2 to 0.8 pm size class (Table 1).Phototrophs in the Columns and error bars have been back-transformed from 0.2 to 3 Lrn class included flagellates, diatoms and log-transformed data. Columns with the same lett-er are not coccoid forms (listed together as 'others' in Table 1) but significantly different (SNKmultiple range test, p < 0.05) lnlvory by oyster larvae 291

Plankton Cell Size (km)

Fig. 3. Percent of total particulate label distr~butedamong plankton size fractions. (A) 'v-labeled plankton; (B) 3H- labeled plankton

higher than 3H-labeled organisms (Figs. lc & 2c). As a general trend, larvae cleared particulate % and '" at greater rates when more and larger labeled food items were available. Larvae cleared 14C-labeled organisms ~~Q2lQe2ruz~ in the 0.2 to 30 pm size class at a significantly greater Size Class (pm) rate (p< 0.05) than smaller 14C-labeled classes (Fig. lc). Larval clearance rates for the 0.2 to 3 pm, 0.2 to 10 pm and 0.2 to 20 pm size classes were similar but all were Fig. 2. Crassostrea vlrginica. Larval ingestion and clearance of various size classes of 'H-labeled plankton. (A) dpm of 'H significantly greater (p i 0.05) than clearance of the 0.2 ingested per larva; (B)dpm of 'H per m1 of plankton; (C)clear- to 0.8 pm class. Clearance of %-labeled organisms in ance rates for plankton size classes. Other details as in Fig. 1 the 0.2 to 30 pm size class was significantly greater (p < 0.05) than all other size classes. No significant ingested in the 0.2 to 10 pm or 0.2 to 20 pm classes but differences in clearance rates were found among any it was significantly (p < 0.05) greater than that ingested of the size classes 20 um. in the 0.2 to 0.8 ,pm or 0.2 to 3 pm classes (Fig. 2a). This indicates that larvae ingested 3 to 30 pm heterotrophs. Cells 3 to 30 pm contained only 8 % of the total parti- Ingestion of protozoans culate 3H whereas the vast majority was contained in 0.2 to 0.8 pm (70 %) and 0.8 to 3 pm (22 %) cells (Fig. In experiments where larvae were fed plankton that 3b). Microscopic observations showed that the 0.2 to had been incubated with 3H-thymidine for 15 min the 0.8 pm class was primarily free-living bacteria while amounts of particulate 3H ingested were not signifi- larger classes included attached bactena, flagellates, cantly different between the 0.2 to 3 pm and 0.2 to 30 and ciliates (Table 1). Unattached bactena were also pm size classes (Fig. 4a) despite the fact that the 0.2 to found in > 3 pm size fractions, presumably due to 30 pm class contained significantly (p i 0.05) more dislodgement from detritus. particulate 3H (Fig. 4b). Both plankton size classes contained significantly (p < 0.05) more particulate 3H after 120 n~inthan after 15 min of incubation (Fig. 4b) Clearance rates of 14C- and 3H-labeled organisms and as a result larvae ingested significantly (p i 0.05) more particula.te 3H from each size class (Fig. 4a). Larval clearance rates, calculated using Eq. (l), show Larvae ingested significantly (p < 0.05) more parti- that I4C-labeled organisms were cleared at rates culate 3H from the 120 min labeled 0.2 to 30 pm size 292 Mar. Ecol. Prog. Ser. 78: 285-301, 1991

papillatum over the initial 30 min of the timecourse feeding experiment (Fig. 5a). Between 30 and 60 min the rate of uptake declined, suggesting that the flux of particulate 14C to and from the larval gut had reached steady state and that assimilation of I4C alone accounted for the subsequent slower rate of Increase in dpm larva-' (Fig. 5a). The ingestion rate of I. papillatum by oyster larvae was based on the least-squares regression for the linear portion (0 to 30 min) of the 14C uptake curve: 14C uptake (dpm larva-') = 0.33492(X) + 0.71021, where X= time (min). Specific activity of prelabeled I. papillatum was 0.013 dpm cell-'. The clearance rate of I. papillatum was 0.0640 m1 larva-' h-'. Larvae ingested the labeled ciliate Smcil but uptake was exponential (r2= 0.843, p = 0.0001) over the 360 min feeding timecourse (Fig. 5b). This pattern of uptake may be the combined result of high larval assimilation of 3H-labeled Smcil and an increase in the specific activity of Smcil over the 360 min timecourse. The specific activity of Smcil could increase since both residual dissolved 3H-labeled compounds and 3H- labeled bacteria were present in whole water samples taken from the washed Smcil culture. Therefore, in the presence of the <30 pm plankton Smcil may have ingested newly labeled natural bacteria as well as prelabeled cultured bacteria. The < 5 pm control experi-

m xxmlio Size Class (pn)

Fig. 4. Crassostrea virginica. Larval ingestion and clearance of 3H-labeled plankton slze classes at 15 and 120 min. (A) dpm of 'H ingested per larva; (B) dpm of 3H per m1 of plankton; (C) clearance rates for plankton size classes. Other detalls as in Fig. 1 class than from the corresponding 0.2 to 3 pm class (Fig. 4a). Larval clearance rates for both size classes prelabeled for 15 min were similar to the clearance of the 0.2 to 3 ,urn class prelabeled for 120 min but clearance of the 0.2 to 30 pm class prelabeled for 120 min was significantly greater (p < 0.05) than all other treatments (Fig. 4c). In the experiments using cultured heterotrophic flagellates and ciliates, total label uptake by Crass- ostrea virginica larvae (i.e. uptake of dissol.ved radiolabeled compounds and labeled cells) was significantly greater (p < 0.05) than uptake in control experiments containing dissolved label (flagellate experiment) or Minutes dissolved + bacterial label (ciliate experiment), thereby Fig. 5. Crclssostrea virginica. Larval ingestion of cultured confirming that labeled flagellates and ciliates were l:c-labeled ,sonc,n7r, papillaturn and ingested. Larvae showed a linear rate of uptake (B).E~ch point rc,presents the mean dprn per larva per (r2=0.641, p = 0.0006) of the labeled flagellate lsonerna replicate vial Baldwin & Newell: Om nivory by oyster larvae 293

ment would not have accounted for such an increase in As a result, we feel confident that the ingestion of specific activity of Smcil since Smcil were not present particulate 3H by C. virginica larvae represents in < 5 pm samples taken from the washed Smcil ingestion of heterotrophic bacteria and phagotrophic culture. Similar patterns of uptake by grazing zoo- protozoans. plankton have been demonstrated in experiments Results from our feeding experiments using dual- where grazers feed on prey organisms in the presence labeled natural plankton samples and labeled proto- of radiolabeled substrate (e.g. Roman & Rublee 1981). zoan cultures indicate that planktotrophic larvae of The ingestion rate of Smcil by oyster larvae was based Crassostrea vjrginica feed as omnivores, consuming on the least-squares regression fitted to the entire (0 to heterotrophic bacteria and phagotrophic protozoans in 360 min) 3H uptake curve: 3H uptake (dpm larva-') = addition to phototrophs. A previous study by Fritz et al. 10°~00"H7~-ulX 5.2730, where X = time (min). Specific (1984)demonstrated that C. vjrginica larvae could feed activity of prelabeled Sincil was 4.8 dpm cell-'. The on a variety of phytoplankton taxa found in natural clearance rate of Smcil was 0.1093 m1 larva-' h-'. plankton assemblages of Delaware Bay, USA. In the present study, ingestion of 14C-labeled organisms by oyster larvae also indicates their ability to ingest nat- DISCUSSION urally occurring phytoplankton and other phototrophic organisms (possibly rnixotrophic protozoans). While it Diet composition is possible that oyster larvae ingested 14C-labeled phagotrophic protozoans (which could include mixo- One of our primary goals was to determine whether trophic and obligate heterotrophic protozoans) it is Crassostrea vjrgjnjca larvae consume non-phytoplank- likely that at least some, and probably most, of the ton foods from the natural environment. Hence, a ingested particulate 14Cwas in the form of phototrophs critical assumption for our study held that particulate given the abundance of these organisms in the < 30 pm 3H uptake by oyster larvae represented the ingestion plankton (Table 1). of 3H-labeled heterotrophic bacteria and/or 3H-labeled Laboratory studies have indicated that bivalve phagotrophic protozoans. Following the 120 min incu- larvae can feed upon cultured bacteria (e.g. Hidu & Tu- bation of <30 pm Choptank River plankton with biash 1963, Martin & Mengus 1977, Douillet 1991). 3H-thymidine we found that of the total particulate 'H However, it is generally thought that these larvae are retained on 0.2 pm filters, 70.3 and 92 % passed incapable or extremely inefficient at capturing par- through 0.8 pm and 3 pm filters, respectively (Fig. 3b). ticles as small as most natural free-living bacteria (<0.5 Over 98.9 and 99.8 ?,0 of the heterotrophic bacteria pm; Ducklow et al. 1988). Our results ~ndicatethat pro- were counted in the 0.2 to 0.8 pm and 0.2 to 3 pm size dissoconch I1 oyster larvae are capable of capturing classes, respectively (Table 1) Furthermore, hetero- and ingesting free-living bacteria as small as 0.2 to 0.8 trophic bacteria accounted for over 99.9 and 97.6 'X, of pm but larval clearance rates for these cells are low the total cells enumerated in the 0.2 to 0.8 pm and 0.2 (0.0016 m1 larva-' h-') (Fig. 2c). Larval ingestion of to 3 pm size classes, respectively (Table 1). These larger (i.e. 0.8 to 30 ,pm) forms of bactei-ia, such as bac- findings confirm that heterotrophic bacteria were terial aggregates or bacteria attached to detritus, was labeled with 3H-thymidine and strongly suggest that not detected in this study. For example, on the bask of they dominated particulate matter labeled with 3H. feeding experiments using 15 min prelabeled and size Similar correspondence between size fractioned fractioned plankton, we found that oyster larvae con- particulate 3H and bacterial abundance has been sumed equal amounts of particulate 3H (i.e. bacteria) documented in other studies using 3H-thymidine (e.g. from both the 0.2 to 3 pm and 0.2 to 30 pm size classes, Fuhrman & Azam 1980, Fuhrman et al. 1986). Further- despite the fact that the 0.2 to 30 pm class contained more, one of these studies (Fuhrman et al. 1986) also more particulate 3H than the 0.2 to 3 pm class (Fig. used microautoradiography to confirm the nearly 4a,b). These results indicate that while 3 to 30 pm exculsive uptake of 3H-thymidine and 3H-adenine by bacteria (predominately attached cells, Table 1) were heterotrophic bacteria. Based on the dominant role of labeled with 3H-thymidine, only 0.2 to 3 pm bacteria 13 pm heterotrophic bacteria in the uptake of 3H- (predominately free-living cells, Table 1) were inges- thymidine in our study, and similar findings from pre- ted. Assuming that larvae do not discriminate between viously discussed autoradiograghpy studies (Fuhrman free and attached forms of bacteria, this result may & Azam 1982, Bern 1985, Fuhrrnan et al. 1986, Douglas indicate that larvae selected against the particles to et al. 1987, Carman 1990), we further conclude that which bacteria were attached. 3 to 30 pm Q-labeled organisms were attached bac- Our results also demonstrate that oyster larvae teria and, to a lesser extent (see below), phagotrophic ingest phagotrophic protozoans. For example, in the protozoans which had consumed 3H-labeled bacteria. above experiments using natural plankton, larvae 294 k1~1rEcol Prog. S

ingested similar amounts of particulate 3H from the larvae cleared 0.2 to 3 pm cells (i.e. picoplankton size 0.2 to 3 pm and 0.2 to 30 pm size classes incubated for organisms) at similar rates as 0.2 to 20 pm cells. While 15 min but larvae ingested more particulate 3H from this may be explained In terms of the larval functional the 0.2 to 30 pm class than from the 0.2 to 3 lm class response, it may also indicate that picoplankton size after a 120 min prelabeling period (Fig. 4a). This indi- cells are cleared as efficiently as larger nanoplankton cates that 3 to 30 pm organisms were ingested by size cells. Efficient clearance of picoplankton slze cells larvae but ingestion was detected only after these was also reported by Gallager (1988) in a study of par- organisms had been prelabeled for > 15 min. Since 3 ticle manipulation by 100 pm and 234 pm larvae of the to 30 pm organisms were labeled at 15 min (mainly bivalve Mercenaria mercenaria. In those experiments, attached bacteria) but not ingested (Fig. 4a, b), the 3 to both sizes of larvae cleared the cyanobacterium 30 pm organisms ingested after 120 min most likely Synechococcus spp. (1 X 0.5 pm) at rates greater than were phagotrophic protozoans that had consumed the alga lsochrysis aff. galbana (4.5 pm diameter) particulate 3H.The ability of oyster larvae to capture when fed mixtures of these cells (each at 3 X 104 cells and ingest heterotrophic protozoans was confirmed in ml-l). Together, these results suggest that relatively grazing experiments using cultured flagellates and small organisms, such as the picophytoplankton ciliates (Fig. 5). assemblages that are highly abundant in coastal The inclusion of heterotrophs in a mixed particle diet marine systems (e.g Davis et al. 1985, Geider 1988, may serve to enhance the growth and devel.opment of Ray et al. 1989), may be important food so'urces for Crassostrea virginica larvae. Numerous laboratory oyster and other bivalve larvae. studies have demonstrated increased larval growth Our results also demonstrate that prodissoconch I1 and survival when fed mixed algal diets (for review see oyster larvae ingest relatively large cells (20 to 30 pm, Webb & Chu 1983) and other work has demonstrated Figs. la & 2a) and fluorescent microspheres (21 .l pm). enhanced growth of C. virginica and other bivalve Mackie (1969) also found that Crassostrea virginica larvae when fed certain mixed diets of algae and larvae could ingest cells in this size range. Although bacteria (Hidu & Tubiash 1963, Douillet 1991). Due to planktonic food organisms of this size generally are the small size of most bacteria and the low clearance numerically less abundant than smaller cells (Table 1; rates oyster larvae demonstrated for these cells, bac- Ray et al. 1989), larvae may still acquire significant teria may contribute more in terms of essential nutri- amounts of energy and nutrition from these organisms ents (e g. amino acids or B-complex vitamins; Phillips given their large cell volume. As an illustration, we can 1984) than energy. Phagotrophic protozoans may also compare estimates of the total cell volume (= biomass) contribute nutrients and energy to oyster larvae. In a of phototrophs ingested by larvae from the 0.8 to 3 pm study of other grazing zooplankton Stoecker & Egloff and 20 to 30 pm plankton size fractions. Briefly, in (1987) demonstrated enhanced egg production in the Table 2 we use data on cell abundance and larval copepod Acartia tonsa when fed mixed diets including ingestion of particulate I4C and 3H to estimate the heterotrophic ciliates. Phillips (3984) suggested that numbers of cells ingested from different size fractions protozoans may provide important nutrients such as (see discussion below regarding assumptions used in sterols and polyunsaturated fatty acids. We have found these calculations). Based on these ingestion rates, and that the cultured ciliate used in our experiments con- assuming [l) a mean equivalent spherical diameter of tains n3-polyunsaturated fatty acids (Baldwin. & Ne- 3 pm (a liberal estimate) and cell volume of 10.6 ,um3 well unpubl.), which are thought to be required for cell-' for 0.8 to 3 pm cells and (2) a mean equivalent growth and development of certain bivalve larvae spherical diameter of 21 pm (a potentially conservative (Helm & Laing 1987, Whyte et al. 1989). estimate) and cell volume of 3637 pm"el1-' for 20 to 30 In addition to the types of naturally occurring orga- pm cells, we estimate ~ngestionrates of 1.21 X 105 pm7 nisms that prodissoconch 11 oyster larvae ingest, our cell volume larva-' h-' and 9.13 X 10' pm%ell volume results also demonstrate that these larvae can ingest larva-' h-' for 0 8 to 3 pm and 20 to 30 pm cells, respec- 0.21 to 21.1 pm fluorescent microspheres and a wide tively. Thus, although 0.8 to 3 pm cells were 105 times size range of food particles (0.2 to 0.8 pm to 20 to 30 pm as abundant as 20 to 30 pm cells and ingested at 4.5 cells). In general, larvae cleared 0.2 to 0.8 pm plankton tim.es the rate for 20 to 30 pm cells (in terms of cell at lower rates than larger classes (Figs. lc & 2c). This number), larvae may have ingested 75 times more bio- may reflect a depressed clearance rate under relatively mass from 20 to 30 pm cells. Selective ingestion of 20 to low food biornass concentrations (i.e. a functional 30 pm cells (see below) contributed to this high response as described by Lam & Frost 1976, Gallager estimate; hence, it is unclear whether C. virginica 1988) and, to a certain extent, poor retention of these larvae typically ingest large cells in this quantity. Even cells on the velum (see Walne 1965, Riisgdrd et al. so, the above analysis suggests that relatively large 1980, Sprung 1984, RiisgArd 1988). We also found that but scarce cells should not be overlooked when con- LLOO'O L 100'0 OE-Z'O OC-Z'O 6200'0 9100'0 oz-Z'O oz-2'0 SZOO'O 9100'0 01-2'0 OI-Z'O ozoo'o 9200'0 C-Z'O C-2'0 9100'0 S IOO'O 8'0-2'0 8'0-2'0

PZ9 PPP I IL'O LOOZ 9ZP2 06-OZ I15Z 286 11 ILL'P 08E2 P8S9 OC-OZ C bZ 91 8PI'O 86Z P 8C9 OZ-02 86s 8SLI 9C6' 1 916E 18SL OZ-01 26 99 8lL'O PSP8 ZL09 01-C 80L 8C L S61'0 08OCC ZPP9 01-E I L2 Z L OZP'O CE8ZS Of LZZ C-8'0 SLPLI R9P I PO'O 098PbI OC6S E-8'0 69L8 Ptl CIO.0 00901LS Z98lL 8'0-2'0 ZS ZL SLC'I OClZ 6Z6Z 8'0-Z'O

uos!~eduro~loj papnpu! ale (l),b3 6u!sn palelnsles se sale1 a:,uel~al3.sale1 asuerea[s z poqlaw jo uo!lelnsles U! pasn suo!~dmnssejo uo~leueldxaloj lxal aas .sse[s azls ysea U! alqel!elze sllas pue palsabu! sllas uo paseq palelnsles sale1 asueleals 1enJeT ,e3!u!6rfir!near1sosse.y .Z alqeL 296 Mar. Ecol. Prog. Ser 78: 285-301, 1991

sidering the relative importance of naturally occurring treat cells wthin each size fraction in a similar fashion, cells in the diets of bivalve larvae. we would expect to find similar clearance rates among different size classes, regardless of the amount or na- ture of distribution of label in the various size fractions Food selection that constitute the different size classes. Our data support this conclusion. For example, clearance rates of Not only did we find ingestion of relatively large I4C-labeled organisms in the 0.2 to 3 pm, 0.2 to 10 pm, cells in this study, but it appears that larvae selectively and 0.2 to 20 pm size classes were not significantly ingested 20 to 30 pm cells. As shown in Figs. 1 & 2, different from each other (Fig. lc) despite the fact that larval clearance rates for 0.2 to 30 pm organisms were the fractions within these classes were non-uniformly much greater than for those in the 0.2 to 20 pm size labeled (Table 2). Based on these results, if oyster class. These results imply that the ingestion of 20 to 30 larvae captured and ingested 20 to 30 pm cells in a pm cells was the reason for the higher clearance of 0.2 manner simllar to that for < 20 pm cells, they would to 30 pm organisms. Over 87 and 61 % of the total have cleared a similar fraction of the total available particulate I4C and 3H,respectively, ingested by oyster particulate I4C or 3H,and therefore clearance rates as larvae when fed 0.2 to 30 pm plankton was in the form calculated using Eq. (1) would have been similar to of 20 to 30 pm cel.1~.Such a result might be expected if those of smaller size classes. However, since we found (1) larvae did not discriminate among cells in the 0.2 to a significantly higher clearance rate for the 0.2 to 30 30 pm class and (2) the 20 to 30 pm size fraction pm class, and because differences in the amount of contained a disproportionately high amount of the total label within size fractions would not have caused this, label present in the 0.2 to 30 pm class. However, the the feeding behavior of the larvae must be the under- amount of particulate '*C in all size fractions was lying cause of the high clearance rate This strongly similar and a very low percentage (1.4 %) of the total suggests that oyster larvae ingested the 20 to 30 pm particulate 3H was in the 20 to 30 pm fraction (Fig. 3) cells in a preferential manner as compared to < 20 pm These results suggest that oyster larvae must have cells. selectively ingested 20 to 30 pm organisms in order to Based on previous studies which show that retention acquire such a disproportionate amount of particulate efficiency in larval bivalves declines for particles label from these cells. Furthermore, even if the 20 to 30 greater than ca 4 pm (Walne 1965, Riisgdrd et al. 1980, pm fraction did contain a large amount of the total Sprung 1984, Riisgdrd 1988) we conclude that the high label in the 0.2 to 30 pm class, the clearance rates for clearan.ce of 20 to 30 pm cells is not due to enhanced this size class would not have been dramatically higher cell retention on the velum per se, but is the result of than that for the < 20 pm classes since, using Eq. (l), active selection by larvae. Gallager (1988) found that we compare the large amount of ingested label to the Mercenaria mercenaria larvae could discriminate large amount of label present. In other words, this among captured food cells both at the mouth and near clearance rate would be similar to rates calculated for the junction of the esophagus and stomach. While this the <20 pm size classes where relatively small remains unconfirmed, it is likely that Crassostrea virgi- amounts of ingested label were compared to small nica larvae possess similar discriminatory capabilities. amounts of label present. Clearance rates for the 0.2 to We are not certain of the type or types of cells that 30 pm class could only have been higher if the larvae were selectively ingested by oyster larvae in the pre- selectively ingested 20 to 30 pm cells. sent study. However, since larvae selectively ingested In addition, differences in the amount of particulate both I4C- and 3H-labeled 20 to 30 pm cells it is possible label among size fractions would not have led to the that larvae ingested organisms that were dual-labeled, high clearance rate for the 0.2 to 30 km size class, and cells that were labeled wlth "C or 3H,or a combination therefore this factor could not be mistaken for selective of all three. As discussed previously, it does not appear ingestion by larvae. It is evident that the amount of that larvae ingested attached bacteria. Thus, if 20 to 30 particulate label as well as the average specific activity pm heterotrophs other than attached bacteria ac- of cells is different for each plankton size fraction counted for the ingested particulate Q, and all 20 to 30 (Table 2). It is also likely that each size fraction con- pm bacteria were labeled at the same specific activity tained a different percentage of cells that were as 0.2 to 0.8 bacteria (0.013 dpm cell-'. Table 2), this labeled. Both cell-specific activity and percentage of would require that the ingested heterotrophic proto- cells labeled (as well as the n.u.mber of cells present) zoans from the 20 to 30 km slze fraction have a speclfic will contribute to the amount of label found in each activity of 58 dpm cell-' Clearly this is unrealistic size fraction. However, since we compare with Eq. (1) given that our 3H-labeled Smcil culture achieved only the amount of label ingested in a given size class to the 4.8 dpm cell-' after 12 h of incubation. For this reason amount of label present, assuming that oyster larvae we feel that much of the ingested 20 to 30 pm parti- Baldwin & Newell: Om~nivory by oyster larvae 297

culate label was derived from dual-labeled mixotrophs calculated clearance rates for the various plankton size (i.e. organisms categorized as 'phototrophs' in Table 1 classes using a second technique (Table 2) in order to given their possession of autofluorescent pigments). check for erroneously high clearance rates derived While some of these mixotrophs were probably cate- using Eq. (1). Here, we use data on cell abundance gorized as 'other' phototrophs (Table l),we believe the (Table 1) and larval ingestion of particulate I4C and 3H dominant mixotrophs in the 20 to 30 pm size fraction (Figs. la & 2a) to estimate the numbers of cells were the blooming dinoflagellates Gyrodjniun~ ingested from d~fferentsize fractions. Using these uncatenum and Gyrnnodjn~urnsanguinium. Both these estimates we can then calculate larval clearance rates species have been found to consume particulate foods for phototrophs and heterotrophs present in each size (Bockstahler & Coats 1990); hence, these large class (made up of one or more size fractions). This mixotrophic cells could have incorporated high approach provides clearance rate estimates based on amounts of "C-bicarbonate (Table 2) and they could cell number as opposed to rates based on the level of have phagocytized heterotrophic organisms that were radiolabel in the larvae and plankton size class, as labeled with 3H. If these organisms are included in calculated using Eq. (1). We will hereafter refer to estimates of the specific activity of ingested 3H-labeled Table 2 estimates as 'Method 2' clearance rates. cells from the 20 to 30 pm size fraction, we obtain more For ease of calculation in Table 2 we have assumed realistic values of 1.66 dpm cell-'. that: (1) all ingested particulate 14C was derived from Our finding of selective ingestion for 20 to 30 pm phototrophs and all ingested particulate 3H was cells by Crassostrea virginica larvae appears to conflict derived from heterotrophs, (2) all enumerated with the results of Fritz et al. (1984) who found that C. phototrophs and heterotrophs were labeled, and (3) all virginica larvae selectively ingested < 10 pm versus cells within a given size fraction were uniformly > 10 pm phytoplankton collected from Delaware Bay. labeled (i.e. of equal specific activity). Because it is Very likely, this discrepancy can be attributed to the possible that these conditions were not strictly met, use of different-sized larvae and/or different experi- actual values of ingested cell number and clearance mental techniques. For instance, we used larvae with a rates may differ somewhat from our estimates. How- mean shell length of 179 pm whereas Fritz et al. (1984) ever, rates calculated using Method 2 are in general used 3 sets of larvae, the largest with a mean shell agreement with those determined using Eq. (1) (Table length of 136 pm. We have found that larvae of shell 2) and both sets of data are in general agreement with length ca 150 pm can ingest fluorescent microspheres clearance rates reported for other bivalve larvae (Table up to 15.8 pm diameter whereas larvae of shell length 3).Two exceptions are the I4C- and 3H-based rates for ca 180 pm can ingest micropheres up to 21.1 pm Given the 0.2 to 30 pm size class as calculated using Eq. (1). these differences in particle ingestion and the fact that Clearly, our I4C-based rate is much higher (nearly many of the 20 to 30 pm cells In our study were oval or 6-fold) than the corresponding Method 2 rate (Table 2), cylindrical in shape with minumum hear dimensions and it is also much higher than rates reported for other of < 20 pm, it is conceivable that the larger larvae used bivalve larvae (Table 3). Although our 3H-based rate in our study were capable of ingesting larger plankton using Eq. (1) is not unusually high compared with cells than the larvae used by Fritz et al. (1984).Another literature values (Table 3) it is over 4 times higher than contributing factor may simply be the difference in the the Method 2 rate (Table 2). We therefore conclude plankton assemblages used in each study. Also, Fritz et that the Eq. (l)-based clearance rates for 0.2 to 30 pm al. (1984) based their results on the disappearance of organisms overestimate the actual clearance rate of phytoplankton (determined n~icroscopically)whereas these organisms and, in agreement with Baars & our results are based on the direct ingestion of labeled Oosterhuis (1985),we assume these rates are the direct cells, which include heterotrophic as well as photo- result of larval selection for high specific activity cells trophic organisms. present in the 20 to 30 pm size fraction. For the 0.2 to 30 pm size class then, we feel that Method 2 clearance rates are more realistic and conservative. We do not Larval clearance rates think Eq. (l)-based rates for the other size classes are in question since they agree with Method 2 rates Baars & Oosterhuis (1985) point out that clearance (Table 2) and other clearance rates found in the litera- rates as determined using radioisotope labeling tech- ture (Table 3). niques based on the model of Daro (1978) may over- Further examination of clearance rates generated in estimate actual clearance rates if the grazing organism this study indicates that I4C-labeled cells were cleared feeds selectively on a labeled food cell of high specific at higher rates than 3H-labeled cells (Table 2). While activity. Since we found such selective feeding by this may suggest that oyster larvae feed primarily on Crassostrea virginica larvae in the present study, we I4C-labeled organisms (see Roman & Rublee 1981, 298 Mar. Ecol. Prog. Ser. 78: 285-301, 1991

Table 3. Clea.rance rates of bivalve larvae reported in the literature

Species Shell length Temp. Food Clearance Method Source (Pm) ("C) rate (m1 larva ' h l)

Crassostrea 179 25 <30 pm natural plankton 0.0017-0 0825 Radioisotope labeling Thls study virginica 237 22 Isonerna papillaturn 0.0640 Radioisotope labeling This study 273 22 Cultured ciliate Smcil 0.1093 Radioisotope labeling This study

300-376 22 Isochrysis galbana 0.0541 Coulter Counter Wlddows et al (1989) (

Crassostrca 89-294 25 Isochrysis galbana + 0.0023-0.0935 Coulter Counter Gerdes (1983) gig& Chaetoceros calcitrans

Crassostl-ea 2 13 21 Tetraselmls suecica + 0.0017-0 0105 Hemacytometer Crisp et al. (1985) gigas Nannochloris sp.

0stres 230 2 1 Tetraselmis suecica + 0.0025-0.0097 Hemacytometer Crisp et al. (1985) edulis Isochrysis galbana

Mercenaria 100-234 22 Isochrysis galbana 0.0000093-0.0994 Video micoscopy Gallager (1988) 93-164 28 Isochrysis galbana 0.01 15-0.0552 Coulter Counter Riisgird (1988)

Mytilus 120-250 18 Isochrysis galbana 0.017-0.052 Coulter Counter Sprung (1984) edulis 125-250 17-19 Isochrysis galbana + 0.01 1.8-0.0853 Coulter Counter Jespersen & ~Vfonochrysisl uthen Olsen (1982) 110-220 1.5 Isochrysis galbana + 0.0125-0.0967 Coulter Counter Riisgird & Monochrysis lutheli Randlov (1981)

Lessard & Swift 1985),the disparity in rates is at least Such rates are similar to those calculated for 14C- in part because 3H-basedclearance rates calculated for labeled cells (Table 2) and are ca 10 times higher than all size classes were kept low due to the presence of those calculated including 3H-labeled 0.2 to 0.8 pm abundant and poorly retained 0.2 to 0.8 pm hetero- cells. trophic bacteria.. In other words, in both 3H-based clearance rate determ~nati.ons(Eq. (1) and Method 2) CONCLUSION the denominator of the equation was dominated by abundant 0.2 to 0.8 pm bacteria; in terms of particulate In the present study we have determined the sizes 3H, these cells contained 70 ':'o of the total label ava~l- and general trophic groups of planktonic organisms able (Fig 3b) and in terms of cell num.bers these cells that comprise the natural particle diet of Crassostrea represented 99 % of the total (Table 2). Because these virginica larvae. Knowledge of the natural diets of such small cells were cleared so poorly (Fig. 2c) and the planktotrophic larvae is central to an understanding of larger 'A-labeled cells were relatively few in number the relationships between in situ food supply and (Table 1) and contained, so little 3H (Fig. 3b), the composition and larval growth, surv~val,and metamor- ingested number of large cells or forms of particulate phosis success. It is clear from our resuts that C. virgi- 3H was small relative to the total number of cells or nicalarvae do not feed exclusively as herbivores in the particulate a.vailable in each size class. As a result, natural environment. Even i.n phytoplankton-rich clearance rates were low. In contrast, because parti- estuaries such as Chesapeake Ray, it appears that culate 14C was more evenly distributed among size oyster larvae consume heterotrophic bacteria and fractions (Fig 3a) and small, less efficiently retained protozoans. Copepods are also known to consume "C-labeled cel.1~did not dominate total cell number heterotrophic prey in estuaries where standing stocks to the degree seen for 3H-labeled cells, 14C-based of phytoplankton are high (G~fford& Dagg 1988. White clearance rates were higher and were likely more 1991). Thus, it is clear that the relationships between realistic. It is interesting to note that if we assume larval nutrition and larval growth and survival in the oyster larvae are incapable of clearing Q-labeled 0.2 natural environment may not always be explained to 0.8 pm cells, i.e. we calculate clearance rates for 0.8 solely on the basis of phytoplankton abundance or to 30 ,urn cells, we derive rates as high as ca 0.0245 m1 composition. The relative importance of phototrophs larva ' h-' (Eq. 1) and 0.0167 m1 larva-' h-' (Method 2). and heterotrophs in the diet of oyster larvae is unclear Baldwin & Newell: Om~nivory by oyster larvae 299

but it is possible that heterotrophs play an important Bockstahler, K. R., Coats, D. W (1990). Mixotrophy in supplementary role in larval nutrition. This may be Chesapeake Bay dinoflagellates. J. Protozool. 38: 11A particularly true for phagotrophic protozoans (which (Suppl.)(Abstract) Carman, K R. (1990). Radioact~ve labeling of a natural have been largely overlooked as a food source for assemblage of marine sedimentary bacteria and micro- planktotrophic larvae) given their cell size and abun- algae for trophic studies: an autoradiographic study. dance. The nutritional importance of protozoans has Microb. Ecol. 19: 279-290 also been suggested for other taxonomic groups of Conlcv, W. J., Turner, J T (1985). Omnivory by the coastal marine copepods Centropages hamatus and Labidocera & grazing zooplankton (for review see Stoecker aestiva. Mar. Ecol. Prog. Ser. 21. 113-120 Capuzzo 1990). Crisp, D. J., Yule, A. B., White, K. N. (1985). Feeding by oyster Our experiments also suggest that a wide size range larvae: the functional response, energy budget and a com- of naturally occurring food organisms may be nutri- parison with mussel 1.arvae. J. mar b~ol.Ass. U.K. 65 tionally important to Crassostrea virginica larvae. In 759-783 Cuhel, R. L., Waterbury, J. B. (1984).Blochemica1 composition particular, highly abundant picoplankton assemblages and short term nutrient incorporation patterns in a and less abundant but large planktonic organisms may unicellular marine cyanobacterium, Synechococcus be more important than previously thought. It also (WH7803). Limnol. Oceanogr. 29: 370-374 appears that oyster larvae can selectively ingest Daro, M. H. (1978). A simplified 14C method for grazing certain naturally occurring plankton cells. Together, measurements on natural planktonic populations. Helgolander Meeresunters. 31: 241-248 omnivory, selective ingestion, and the utilization of Davis, P. G.. Caron, D. A., Johnson, P. W., Sieburth, J. McN. a wide size range of food organisms by C. virginica (1985). Phototrophic and apochlorotic components of larvae suggest that these larvae are well adapted to picoplankton and nanoplankton in the North Atlantic the complex food environments typical of coastal geographic, vertical, seasonal and dlel distributions. Mar Ecol. Prog Ser. 21. 15-26 marine systems. Dolan, J. R., Coats, D. W. (1990). Seasonal abundances of planktonic ciliates and microflagellates in mesohaline Chesapeake Bay waters. Estuar. coast. Shelf Sci. 31. Acknowledgements. We thank T Diamond and G. Williams 157-175 for technical assistance and Drs R. Rivkin, E. Lessard, M. Douglas. D. J., Novitsky. J. A., Fournier, R. 0. (1987). Micro- Roman, and V. Kennedy for helpful discussions and for radiography-based enumeration of bacteria with esti- improvements in an earlier draft of this manuscript. We also mates of thymidine-specific growth and production rates. thank D. Meritt for supplying oyster larvae. Dr E. Lessard pro- Mar. Ecol. Prog. Ser. 36: 91-99 vided the cultured ciliate Smcil used in this study and Drs D. Douillet, P A (1991).Beneficial effects of bacteria on the cul- Coats and G. Small identified the ciliate Sharyn Hedrick ture of larvae of the Pacific oyster Crassostrea gjgas identified the bloomlng dinoflagellates. The flagellate Iso- (Thunberg). Doctoral dissertation, Oregon State Univer- nema papjllatum was purchased from the Amer~canType sity, Corvall~s Culture Collection. This study was supported by Grant Ducklow, H. W, Kirchman, D. L (1983). Bacterial dyna- #NA86AA-D-SGOOG from the National Sea Grant Program, mics and d~stnbutionduring a spring diatom bloom in NOAA, to Roger I. E. Newell and Thomas J. Jones (University the Hudson River plume, USA. J. Plankton Res. 5: of Maryland), and a Research Assistantship for B.S.B. pro- 333-354 vided by Horn Point Environmental Laboratory. Ducklow, H. W., Peele, E. R.. Hill, S. M., Quinby. H. L. (1988). Fluxes of carbon, nitrogen, and oxygen through estuarine bacterioplankton. In: Lynch, M. (eds.) Understanding the estuary: advances in Chesapeake Bay Research. LITERATURE CITED Chesapeake Research Consortium Publication 129. Baltimore, Maryland, p. 51 1-523 Baars, M. A., Oosterhuls, S. S. (1984). Diurnal feeding Fritz, L.W., Lutz, R. A., Foote, M. A., Van Dover, C. L.,Ewart, rhythms in north sea copepods measured by gut fluores- J. W. (1984) Selective feeding and grazlng rates of oyster cence, digestive enzyme activity and grazing on labeled (Crassostrea v~rginica)larvae on natural phytoplankton food. Neth. J. Sea Res. 18: 97-119 assemblages. Estuaries 7: 513-518 Baars, M. A., Oosterhuis, S. S. (1985). Zooplankton grazing in Fuhrman, J. A., Azam, F. (1980). Bacterioplankton secondary natural water with high concentration of 14C bicarbonate: production estimates for coastal waters of British variable live controls and gut passage time. Hydrobiol. Columbia, Antarctica, and California. Appl. environ. Bull. 19: 71-80 Microbiol. 39: 1085-1095 Bayne, B. 1. (1983). The physiological ecology of marine Fuhrrnan, J. A.. Azarn, F. (1982).Thymidine incorporation as a molluscan larvae. In: Verdonk, N. H., van den Blggelaar, measure of heterotrophic bacterioplankton production in J. A. M,, Tompa. A. (eds ) The Molluscs, Vol. 111. Develop- marine surface waters: evaluation and field results. Mar. ment. Academic Press, New York, p. 229-343 Biol. 66: 109-120 Berk, S. G., Brownlee, D. C., Heinle, D. R., Kllng, H. J., Fuhrman, J. A., Ducklow, H. W., Kirchman. D. L., Hudak, J., Colwell, R.R. (1977).C~liates as a food source for marine McManus, G. B., Kramer, J. (1986). Does adenine incorpo- planktonic copepods. Microb. Ecol. 4: 27-40 ration into nuclelc acids measure total microbial produc- Bern, L. (1985).Autoradiographic studies of [methyl-3H]thymi- tion? Limnol. Oceanogr. 31: 627-636 dine incorporation in a cyanobacterium (Microcys6s Gallager, S. M. (1988).Visual observations of particle manipu- wesenbergi1)-bacterium associatioin and in selected algae lation during feeding in larvae of a bivalve mollusc. Bull. and bacteria. Appl. environ. Microbiol. 49: 232-233 mar. Sci. 43: 344-365 300 Mar. Ecol. Prog. Ser. 78: 285-301, 1991

Geider, R. J. (1988). Abundance of autotrophic and hetero- Acanthasterplanci. Limnol. Oceanogr. 32: 895-904 trophic nanoplankton and the size distribution of micro- Olson, R. R., Bosch, I.,Pearse, J. S. (1987). The hypothesis of bial biomass in the southwestern North Sea in October antarctic larval starvation examined for the asteroid 1986. J. exp. mar. Biol. Ecol. 123: 127-145 Odontaster validus. Llmnol. Oceanogr. 32: 686-690 Gerdes, D. (1983). The Pacific oyster Crassostrea gigas. Part I. Pechenik. J. A. (1979). Leakage of ingested carbon by gastro- Feeding behav~ourof larvae and adults. Aquaculture 31: pod larvae, and its effect on the calculation of assimilation 195-219 efficiency. Estuaries 2. 45-49 Gifford D. J., Dagg. M. J. (1988). Feeding of the estuarine Pechenik, J. A. (1987). Environmental influences on larval copepod Acartia tonsa Dana: carnivory vs, herbivory in survival and development. In: Giese, A. C., Pearse, J. S., natural microplankton assemblages. Bull. mar. Sci. 43: Pearse, V. B. (eds.) Reproduction of marine invertebrates, 458-468 Vol. IX, General aspects: seeking unity in diversity. Black- Haas, L. W. (1982). Improved epifluorescent microscopy well Scientific Publications. Palo Alto, California, p. for observing planktonic micro-organisms. Annls. Inst. 551-608 oceanogr., Paris 58: 261-266 Phillips. N. W. (1984). Role of different microbes and sub- Helm, M. M., Laing, I. (1987). Preliminary observations on the strates as potential suppliers of specific, essential nutrients nutritional value of 'Tahiti Isochrysis' to bivalve larvae. to marlne detritivores Bull. mar. Sci. 35: 283-298 Aquaculture 62: 281-288 Plaut, W., Sagan, A. (1958). Incorporation of thymidine in the Hidu H.,Tubiash H. S. (1963). A bacterial basis for the growth cytoplasm of Amoeba proteus. J. biophys. biochem. Cytol. of antibiotic-treated bivalve larvae. Proc. natn Shellfish. 4: 843-847 ASS. 54: 25-29 Pollard, P. C., Mor~arty,D. J. M. (1984). Validity of the tritiated Hobbie, J. E., Daley, R., Jasper, S. (1977). Use of Nuclepore thymidine method for estimating bacterial growth rates: filters for countlng bacteria by fluorescence rnlcroscopy. measurement of isotope dilution durlng DNA synthesis. Appl. environ. Microbiol. 33: 1225-1228 Appl. environ. Microb~ol.48: 1076-1083 Jespersen, H., Olsen, K. (1982). Bioenergetics in veliger Qian, P.-Y., Chia, F.-S. (1990). Detritus as a potential food larvae of Mytjlus edulis L. Ophelia 21: 101-1 13 source for polychaete larvae. J. exp. mar. Biol. Ecol. 143: Kleppel, G. S., Frazel, D., Pieper, R. E., Holliday, D. V. (1988). 63-71 Natural diets of zooplankton off southern California. Mar. Ray, R. T., Haas, L. W., Sieracki, M. E. (1989). Autotrophic Ecol. Prog. Ser. 49: 231-241 picoplankton dynamics in a Chesapeake Bay sub-estuary. Lam, R. K., Frost, B. W. (1976). Model of copepod filtering Mar. Ecol. Prog. Ser. 52: 273-285 response to changes in size and concentration of food. Riisgbrd, H. U. (1988). Feeding rates in hard clam (Mercenar- Limnol. Oceanogr. 21: 490-500 ia rnercenaria) veliger larvae as a function of algal (Iso- Landry, M. R. (1981). Switching between herbivory and carni- chrysis galbana) concentration. J. Shell. Res. 7: 377-380 vory by the planktonic marine copepod Calanus pacificus. Riisgbrd, H. U., Randlev, A., (1981). Oxygen consumption and Mar. Biol. 65: 77-82 clearance as a function of size in Mytilus edulis L. veliger Lessard, E. J., Swlft, E. (1985). Specles-specific grazing rates larvae. Ophella 20: 179-183 of heterotrophic dinoflagellates in oceanic waters, meas- RiisgArd, H. U., Randlsv, A., Kristensen, P. S. (1980). Rates of ured with a dual-label radioisotope technique. Mar Biol. water processing, oxygen consumption and efficiency of 87: 289-296 particle retent~onin vellgers and young post-metamorphic Mackie, G. (1969). Quantitative studies of feeding in the Mylilus edulis. Ophelia 19: 37-47 oyster, Crassostrea virginica. Proc. natn. Shellfish. Ass. 59: Rivkin, R. B. (1986). Incorporation of tritiated thymidine by 6-7 (Abstract) eucaryotic microalgae J. Phycol. 22: 193-198 Malone, T. C., Ducklow, H. W. (1990). Microbial b~omassIn R~vkin,R. B., Bosch, I., Pearse, J. S., Lessard, E. J. (1986). the coastal plume of Chesapeake Bay: phytoplankton- Bacterivory: a novel feeding mode for asteroid larvae. bacterloplankton relationships Limnol. Oceanogr 35. Science 223: 1311-1314 296-312 Rivkin. R. B., Voytek, M. A. (1986). Cell divis~on rates of ~Malone, T C., Kemp, W. M., Ducklow, H. W., Boyton, eucaryotic algae measured by tritiated thymidine incorpo- W. R., Tuttle, J. H., Jonas, R. B. (1986). Lateral variation ration into DNA: coincident measurements of photo- in the production and fate of phytoplankton in a synthesis and cell div~s~onof Individual species of phyto- partially stratified estuary. Mar. Ecol. Prog Ser. 32: plankton isolated from natural populations. J. Phycol. 22: 149-160 199-205 Martin, Y P., Mengus, B. M. (1977). Use of selected bacterial Roman, M R. (1984a). Utilization of detrltus by the copepod strains for feeding larvae of biytilus galoprovincialis Lmk Acartia tonsa. Limnol. Oceanogr. 29: 949-959 in experimental rearing Aquaculture 10: 253-262 Roman, M. R. (1984b). Ingestion of detritus and rnicro- Martinez, J , Riera, M,, Lalucat, J . Vives-Rego, J (1989) heterotrophs by pelagic manne zooplankton Bull, mar. Thymidine incorporation into algal DNA from axenic sc1.35: 477-494 cultures of Synechococcus, Chlorella, and Tetraselmis. Roman, M. R., Rublee, P. A. (1981). A method to determine In Lett. appl. Microbiol. 8: 135-138 situ zooplankton grazing rates on natural part~cleassern- McManus, G. B.. Fuhrman. J. A. (1988). Clearance of blages. Mar. Biol. 65: 303-309 bacteria-sized particles by natural populations of nano- Sagan, L. (1965). An unusual pattern of tritiated thymidine plankton in the Chesapeake Bay outflow plume. Mar incorporation In Euglena. J. Protozool. 12: 105-109 Ecol. Prog. Ser. 42: 199-206 Scherbaum, 0. H., Louderback, A. L. (1960). Incorporation of McManus, G. B.. Fuhrman. J. A. (1990). Mesoscale and tritiated thymidine into heat-treated Tetrahymena. J. seasonal variability of heterotrophic nanoflagellate abun- Protozool. 7 (Suppl.):25 dance In an estuarine outflow plume. Mar. Ecol. Prog. Ser Sherr, B. F., Sherr, E. B.. Rassoulzadegan. F. (1988). Rates of 61: 207-213 digestion of bacteria by marine phagotrophic protozoa: Olson, R.R (1987). In situ culturing as a test of the larval Temperature dependence. Appl, environ. Microbiol. 54: starvation hypothesis In the crown-of-thorns starfish. 1091-1095 Baldwin & Newell. Omn~voryby oyster larvae 301

Sherr, E B., Rassoulzadegan, F., Sherr, B. F. (1989). Bacteri- California, p. 465-550 vory by pelagic choreotrichous ciliates in coastal waters of Swinton, D C., Hanawalt, P. C. (1972). In vivo specific the NW ]Mediterranean Sea. Mar Ecol. Prog. Ser 55, labeling of Chlamydomonas chloroplast DNA. J. Cell Biol. 235-240 54. 592-597 Sokal, R. R,Rohlf, J. F. (1981). Biometry, 2nd edn. \V. H Walne, P. R (1965). Observations on the influence of food Freeman and Co., New York supply and temperature on the feeding and growth of the Sprung, M (1984).Physiological energetics of mussel larvae larvae of Ostrea edulis. Fishery Invest., Lond. I124: 1-45 (Mytilus edulis). 11. Food uptake. Mar. Ecol. Prog. Ser. 17: Webb. K. L..Chu, F.-L. E. (1983). Phytoplankton as a food 295-305 source for bivalve larvae. In: Pruder, G. D., Langdon, Stoecker. D. K., Capuzzo, J. M. (1990). Predation on Protozoa: C., Conklin, D. (eds.) Proceedings of the Second its importance to zooplankton. J. Plankton Res. 12: International Conference on Aquaculture and Nutrition: 891-908 Biochemical and Physiological Approaches to Shell- Stoecker, D. K., Egloff, D. A. (1987). Predation by Acartia fish Nutrition. Louisiana State University, Division of tonsa Dana on planktonic ciliates and rotifers. J. exp. mar. Continuing Education, Baton Rouge, Louisiana, p. Biol. Ecol. 110: 53-68 272-291 Stoecker, D. K.,Sanders, N. K. (1985). Differential grazing by Widdows, J., Newell, R. I. E., Mann, R. (1989). Effects of Acartia tonsa on a dinoflagellate and a tintinnid. J. hypoxia and anoxia on survival, energy metabolism, and Plankton Res. 7: 85-100 feeding of oyster larvae (Crassostrea virginica, Gmelin). Stoecker, D. K., Taniguchi, A., Michaels, A. E. (1989). Abun- Biol. Bull. mar. biol. Lab., Woods Hole 177: 154-166 dance of autotrophic, mixotrophic and heterotrophic White, J. R. (1991). Seasonal study of zooplankton dynamics planktonic ciliates in shelf and slope waters. Mar. Ecol. in the mesohaline Chesapeake Bay. Doctoral dissertation, Prog. Ser. 50: 24 1-254 University of Maryland. College Park Strathmann, R. R. (1987). Larval feeding. In: Giese, A. C., Whyte, J. N. C., Bourne, N., Hodgson, C. A. (1989). Influence Pearse, J. S., Pearse, V. B. (eds.) Reproduction of marine of algal diets on biochemical composition and energy invertebrates, Vol. IX, General aspects: seeking unity In reserves in Patinopecten yessoensis (Jay) larvae. Aqua- d~versity. Blackwell Scientific Publications, Palo Alto, culture 78. 333-347

This article was submitted to the editor Manuscript first received: August 29, 1990 Revised version accepted: October 11, 1991