sign in sign up
<<
Home , Ciliate, Dinoflagellate, Gymnodiniales, Gymnodinium, Predatory dinoflagellate

MARINE ECOLOGY PROGRESS SERIES l Vol. 158: 75-86,1997 Published November 17 Mar Ecol Prog Ser

Prey size selection, and growth response of the small heterotrophic sp. and the Balanion comatum-a comparative study

Hans Henrik Jakobsen*,Per Juel Hansen

Marine Biological Laboratory, Strandpromenaden 5, DK-3000 Helsinger, Denmark

ABSTRACT. Prey selectivity, growth and feeding responses were studied in the cillate Balanion coma- turn (17 pm) and the ~cdinoflagellate Gjmnodinium sp. (7 pm). Almost identical prey size spectra were found for the 2 organisms. Optimum preysize was 8 pm, while the lower and upper hmits of prey capture were -4 and 10 pm, respectively. Maximum growth and ingestion rates of B. comatum were slightly higher than those of Gj~mnodinium sp. Threshold prey concentrationfor growth of B. comatum and Gymnodinium sp. was 11 and 17 pg C I-', respectively. At 15"C, both organisms needed to ingest approx. 1 to 2% h-' of their cell volume in order to sustain basic metabolic actlvlty. Max~mum specific clearance was 2 to 3 tlmes h~gherfor the ciliate compared to the dinoflagellate. Gymnodlnium sp. survived for a longer time than B. comatum when deprived of prey organisms. Gymnodinium sp. cells were not ingested by B. comatum, although they were of a size which is optimal for B. comatum.

KEYWORDS: Balanion comatum. Gyrnnodinium sp. . Prey size spectra Growth - Grazing . Swimming behavior

INTRODUCTION and (Jacobson & Anderson 1986, Gaines & Elbrachter 1987, Hansen 1991a, b, Jeong & Latz 1994), Heterotrophic and ciliates are quan- thus making large heterotrophic dinoflagellates poten- titatively important parts of the marine planktonic food tial competitors with and cladocerans for web (Smetacek 1981, Lessard 1991). Apart from miner- microplankton prey (Lessard 1991, Hansen 1992). alizing organic matter, they represent a link between Only a few papers have been published on the trophic prlmary production and metazooplankton (e.g.Beers & role of the small heterotrophic dinoflagellates (5 to Stewart 1967). Most marine planktonic ciliates feed on 20 pm). These studies suggest that small heterotrophic prey which is about l0 times smaller than themselves dinofl.agellates mainly feed on nanoflagellates, there- (Heinbokel 1978, Jonsson 1986, Verity 1991). How- by potentially competing with planktonic ciliates for ever, raptorial ciliates occur in the (Mon- prey (Bjarnsen & Kuparinen 1991, Strom 1991). How- tagnes et al. 1988, Nielsen & Kiarboe 1994). Het- ever, small dinoflagellates are within the size range of erotrophic dinoflagellates are raptorial feeders capable prey fed upon by their ciliate competitors. Thus, while of feeding on prey items of their own size (Jacobson small dinoflagellates may be competitors with ciliates & Anderson 1986, Hansen 1992). Large dinoflagellates for food, they are also potential prey for the ciliates. (>20 pm) have been found to feed on mainly chain- The aim of the present study was to compare the forming , dinoflagellates, other functional biology of the small heterotrophic dinofla- gellate Gymnodinium sp. [7 pm equivalent spherical diameter (ESD)]and the prostomat~d ciliate Balanion comatum Wulff (17 pm ESD) in order to answer the fol-

O Inter-Research 1997 Resale offull drticle not permitted 76 Mar Ecol Prog Ser 158: 75-86, 1997

lowing questions: (1) Are these organisms competitors Because of the irregular shape of Gymnodinium for nanoplankton prey? (2) Are there any differences in sp., the cell volume could not be estimated from the functional and numerical responses of the 2 organ- linear dimensions. Instead, Gj7mnodinium sp. cells isms when presented with the same prey? (3) 1s the cil- (n = 20 cells) fixed in Lugol's (final conc. l'%) were iate capable of feeding on the smaller dinoflagellate? recorded using a video camera connected to a moni- (4) Can the observed differences be extrapolated to tor and an Olympus" inverted . Cell account for the different trophic roles of ciliates and shapes of Gymnodinium sp. were drawn on a plastic small heterotrophic dinoflagellates in general? transparency covering the monitor screen. The trans- parency was digitized using a MOP videoplan (Key- tronic, Germany) and cell volumes were estimated by MATERIALS AND METHODS the algorithm of the MOP, which measures the maxi- mum length and width of a 2-dimensional ellipso~d We isolated Gymnodinium sp. and Balanion coma- and subsequently rotates the object calculating the tum from water samples from the north of the Ore- volume as a prolate ellipsoid. It is assumed that the sund, Denmark, in September 1995 at a temperature shrinkage of predators and preys are of equal magni- of 1S0C, a salinity of 28 psu, and a depth of 10 m. tude. Crude cultures were initially made by adding the Bioenergetics. Growth, ingestion and clearance cryptophyte Rhodomonas salina (Wislouch) Hill & rates were measured for Balanion comatum and Gym- Wetherbee to samples of natural sea water. After 1 to nodinium sp. fed Rhodomonas salina at prey concen- 3 wk, cells of B. comatum were transferred to a tration ranging from 200 to 7000 cells ml-'. Expen- 65 m1 culture bottle (Nunclon@),isolated using ments were carried out in 270 m1 tissue culture bottles a micropipette and fed R. salina. Cultures were not in triplicate as batch cultures under conditions de- axenic. scribed above. Prior to experiments, cultures of Gym- A range of prey was used (Table 1).The algae nodinium sp, and B, comatum were preincubated at used were all observed to swim with a constant speed the experimental prey concentration for 24 h. Growth of less than 100 pm S-'. Rhodomonas salina and rate was measured as the increase in cell number and Isochrysis galbana was supplied from the culture col- ingestion rate as the decrease in cells compared to con- lection of the Marine Biological Laboratory in Hels- trols without predators. Predator cells were added inglar, University of Copenhagen. The other algae to the experimental bottles in concentrations which were obtained from The Scandinavian Culture Center resulted in a decrease of the prey concentrations of 10 for Algae and , Dept of Algae and Fungi, to 20% during the experiment. Samples were fixed in Botanical Institute, University of Copenhagen, Den- Lugol's (final conc. l %) and at least 400 cells were mark. Algae were grown in B-medium (Hansen 1989) counted using a 25 m1 Utermohl chamber or in a based on Millipore filtered autoclaved sea water Sedgwick-Rafter chamber and an inverted Olympus@ (salinity 30 to 32 psu) at a temperature of 15 * 1°C. microscope. Algae were grown in aerated 250 m1 Erlenmeyer Growth rate was calculated assuming exponential glass flasks at an irradiance of 10 to 15 pE m-' S-' on a growth: 16 h:8 h 1ight:dark cycle. Stock cultures of Gymno- dinium sp. and Balanion comatum were fed R. salina, and maintained in 270 m1 transparent tissue bottles where No and NI are particle concentration at the (Nunclon@)mounted on a plankton wheel (1 rpm). beginning and end of the experiment, respectively, p Otherwise the physical conditions were as stated is the growth rate, and t is the duration of the experi- above. ment (h). Cell volumes of algae (n = 20) and Balanion comatum (n = 20) were esti- ~~bl~1. ~l~~~ used and their corresponding size (as estimated spherical dia- mated from the linear dimensions meter, ESD) of Lugol's fixed cells (final conc. 1 %) using an inverted OlympusQPmicro- Species ESD (pm) Algal scope. Algae cells were assumed to be prolate ellipsoids. Due to the shape of Isochrysis galbana Parke 4.0 Prymnesiophyte the oral apparatus of B, comatum, the Chroomonas vectensis Carter 6.1 Cryptophyte Plagioselmis prolonga Buchter 6.6 Cryptophyte cell volume was estimated using the Rhodomonas salina (Wislouch) Hill & Wetherbee 7.8 Cryptophyte formula: 0.1875 WL2, where W and L Teleaulax amphioxeia (W. Conrad) Hill 8.5 Cryptophyte are the width and length of cells, Rhodomonas marina (Dangeard) Lemmermann 10.3 Cryptophyte respectively (Edler 1979). Jakobsen & Hansen:Comparative study of a ciliate and a dinofldc~ellate 77

Ingestion rates of Balanion comatun? and Gymno- and 12 h (Gymnodinium sp.), the first sample was dinium sp. were calculated using an iterative model. taken. At this time no Rhodomonas salina cells were left in the cultures. The number of B. comatum and Gymnodinium sp. cells were counted and cell volume was estimated at the time points shown in Fig. 9. Locomotive pattern. The locomotive pattern of Gp- nodinium sp. and Balanion comatum was studied by The model assumes that the concentrations of pre- adding a cell suspension to a 2.7 m1 multidish, which dators (y)and prey (X)increase exponentially, with the was subsequently covered with a cover slip. The multi- growth rate constants p, and p,, respectively. The mor- dish was placed under a NikonmDIAPHOT microscope tality induced by the predator is Uy, and was fitted with a video camera, and cells were recorded at calculated iteratively on a computer with steps of a magnification of between x80 and x400. Cells were 0.01 h. Uis the per capita prey uptake per unit time. tracked for at least 1 S, yielding a minimum of 25 video Clearance (F) is a function of the per capita prey frames per tracked cell. At least 40 cells of each preda- uptake per unit time (U)and the average prey concen- tor were tracked. After recording, the video tape was tration (C): played frame by frame and the cell positions were marked on plastic transparencies covering the screen. Subsequently, the transparencies were scanned into a C is calculated according to Frost (1972) to estimate computer data file and digitized using the program the average cell concentration in an exponentially SigmaScanB(Jandel Scientifica, CA, USA). growing culture. Growth yield (Y) was calculated according to the equation (Fenchel 1982a): RESULTS

The ciliate was identified as Balanion comatum from observations made on protargol- and silver-stained where V,. and V, are the cell volumes of Balanion specimens in the light microscope (Fig. l) and with the comatum or Gymnodinium sp, and Rhodo~nonas use of a transmission electron microscope. The cell salina, respectively. body is cup shaped, with a flattened oral end. While Prey size selection. Balanion cornatum and Gymno- the cell volume depended on the food concentration dinium sp. were fed algae ranging in size from 4 to (see Fig. 5), the oral disc was of constant size (9 pm). 12 pm in order to study prey size preferences (Table 1). The oral disc is surrounded by oral dikinetids, with The algae were added at a constant biomass (cell cilia measuring 8 pm in length, and an inner circle of number X cell volume) equivalent to 900 Rhodomonas tentacles (length 12 pm); there is 1 tentacle per salina cells ml-' in the case of Balanion comatum and dikinetid. The average dimension of cells growing at 1200 R. salina cells ml-l in the case of Gymnodinium food saturation was approx. 20 X 15 pm. sp., a biomass which supports approximately 90% of Gymnodinium sp. (Fig. 2) is a spindle-shaped athe- the maximum growth rate for each of the 2 predators cate dinoflagellate with an average length of 8 pm and when fed R. salina. Otherwise the physical conditions a variable width depending on food concentration. The were as described above. food is located in the anterior end of the cell. Mixture experiment. This experiment was conducted Prey organisms are captured by use of a tow filament to investigate the interaction between Balanion coma- and engulfed directly. tum and Gymnodinium sp. when they CO-occur.Cul- Ingestion rates of both Balanion comatum and Gym- tures of Gymnodinium sp. and R. comatum were mixed nodinium sp. increased with prey concentration until a in a suspension of Rhodomonas salina. Controls were maximum level was reached (Fig. 3). However, data run in which B. comatum and Gymnodiniunl sp. were obtained from the 2 organisms were fitted to different fed R. salina in separate cultures. All experiments were equations due to differences in their feeding biology. carried o'ut in 750 m1 tissue culture bottles in triplicate. Balanion cornatum ingested about 22 cells before it At intervals of between 8 and 16 h, 50 m1 was sampled divided. Thus, the functional response can be consid- from each bottle and replaced with fresh medium. The ered as a Holling type 11functional response. However, experiments were run until food was depleted. Gymnodinium sp, engulfed only about a single prey Starvation experiment. Dense exponentially grow- prior to cell division and ingestion rate is solely based ing cultures of Gymnodinium sp. and Balanion coma- on predator-prey encounter. The functional response tun1 cells were diluted to a concentration of -500 pre- of Gymnodiniunl sp. can therefore be considered as a dators ml-' in triplicate culture. After 6 h (B. comatum) Holling type I response. 78 Mar Ecol Prog Ser 158: 75-86, 1997

A Fig. 2. Gymnodinium sp. Nomarslu contrast interference . Cells are fixed under a hanging drop with 5 ?G OsO,. Scale bar = 5 pm

4 Fig. 1. Balanlon comaturn. Nomarski contrast interference microscopy Cells are fixed under a hanging drop of 5 % OsO,. Scale bar = 10 pm

p Balanion comatum had a maximum ingestion rate of A -2 Rhodomonas salina cells h-' (Fig. 3), corresponding

- to a maximum specific ingestion rate (volume of prey ingested/predator volume) of -15 % h-'. The maximum ingestion rate of Gymnodiniurn sp. was -0.043 R. salina cells h-' (Fig. 3), corresponding to a maximum specific ingestion rate of -6 % h-', which is 2 to 3 times Balanion comatum lower than that of B. comatum. I l l The growth rate of Balanion comaturn reached a maximum of 0.058 h-' at a prey concentration of -1000 Rhodomonas salina ml-' (Fig. 4), while Gymnodinium sp. reached a maximum growth rate of 0.039 h-' at a prey concentration of -1300 R. salinaml-' (Fig. 4). Due to differences in the functional biology of the cil- iate and the dinoflagellate, the threshold prey concen- tration for growth (defined as the prey concentration at which p=O) was determined differently. The threshold prey concentration for growth for Balanion comatum was determined by the formula:

Rhodomonas salfna rnl" Pmax(X - XO ) IJ = K +(x+x0) Fig. 3. Ingestion rate as a function of prey concentration. (A) Balanion comatum. The curve is fit to a Michaelis-Menten equation (Holling type 11): where p,,, is the maximum growth rate, xis the actual 2.027(~ 204) U(h-' ) = - prey concentration, xois the threshold prey concentra- 511+(X-204) tion for growth and Kis the prey concentration sustain- (B) Gyrnnodiniurn sp. Curve represents a fit to a Holling type I ing 0.5 p,,,,,. Data was iteratively fitted to the model functional response. U(h-') = 0.00165 + (2.71 X 10"x for using Sigma plotm (Jandel Scientific). Threshold prey X < 1536 Rhodomonas salina rnl", and U(h-l) = 0.043~for X > 1536 R. salina ml-l. Data points represent treatment concentration for growth of Gymnodiniurn sp, was esti- means i l SE mated as the intercept when p = 0. The threshold prey Jakobsen & Hansen: Comparative study of a c~liateand a dinoflagellate 79 P

A- m,' , - .----A-

Balan~oncomatum -0.02 I I I 0 1000 2000 3000 4000 5000 6000 7000 - B Rhodomonas sallna rnl-' , , (i l l 1 'I Fig. 4. Growth rate of Balanion comatum (0) and Gymno- a, dinium sp. (0) as a function of prey concentration (X). See 5 T Table 2 and text for details. Data points represent treatment 9 0 means *l SE p 0a, - C Gymnodrnrum sp concentrations for growth of B. comatum and Gymno- I I I dinium sp, were approximately 227 + 14 (21 SE) and 0 2000 4000 6000 382 +. 61 (+l SE) Rhodomonas salina ml-l, respectively. Rhodomonas salina ml-' Results of these fits are shown in Table 2. The cell volume of Balanion comatum increased Fig. 5 Cell volume as a function of prey concentration (X) (A)Balanion comatum from -1100 pm3 at low prey concentrations to - 2614(x+58)- -2500 pm3 at prey concentrations sustaining maxi- "(lJrn ) - 338 A (X + 58) mum growth rates (Fig. 5).The cell volume of Gymno- and (B) Gymnod~niumsp. 168(~- 286) v(pn3) = - dinium sp. increased from -70 pm3 at low prey con- -124 + (X- - 286) centrations to -160 lu-n3 at food saturation (Fig. 5). for prey concentrations >400 Rhodomonas salina cells ml-'. Gymnodinium sp. and Balanion comatum had a yield Data points represent treatment means 1 SE of -68 + 10 % (&l SE) and 32 + 8 % (+l SE), respec- tively, when cells were growing at maximum growth rates (Fig. 6). Yield decreased at prey concentrations When subjected to starvation, Balanion comatum less than -600 and -250 cells ml-' for Gymnodinium immediately decreased in number and cell volume sp. and B. comatum, respectively (Fig. 6). Maintenance (Fig. 9). After 50 h, only 10% of the initial concentra- requirements, defined as the specific ingestion rate at tion of B. comatum was left, and cells had shrunk to p = 0, were low (1 to 2 % h-'; Fig. 7) for both species. -400 pm3 (9 pm ESD). After 60 h, no B. comatum cells The maximum specific clearance of Balanion coma- were left in the culture. When subjected to starvation, tum was -2 times higher than that of Gymnodinium Gymnodinium sp. produced swarmer cells (small fast- sp., 17 X 10' and 8.3 X 105 body volumes h-', respec- moving cells) which had a cell volume of about 60 to tively (Fig. 8). The maximum absolute clearance of B. 70 pm3. One third of cells immediately underwent l comatum and Gymnodinium sp. was 2.8 and 0.053 p1 post feeding cell division and the average cell volume h-', respectively (data not shown). was reduced by 50% before cells died off. Most of the

Table 2. Balanion comatum and Gymnodinium sp. Values of maximum growth (p,,,) and ingestion (U,,,,) rates. Due to differ- ences in feeding biology, data on B. comatum are fitted to a Holling type I1 response, while data on Gymnodinium sp. were fitted to a HolIing type I response. X actual prey concentration; K prey concentration sustaining 0.5pm,,,

Predator Fit type Equation r2 pmaxlUn,,, (h-') K (cells ml-'1

Growth

Balanion comatum Holling I1 ~(h-']= 0.058(x- 227)/[315 + (X- 22711 0.89 0.058 315 1 Gymnodinium sp. Holling I p(h-l) = -0 0143 + 4.06 X IO-~X X< 1318 Ingestion Balanion comatum Holling I1 U(h-l) = 2.027(x - 204)/[511 + (X- 204)] 0.96 2.027 Gymnodinium sp. Holling I U(h-l) = 0.00165 + 2.71 X lO-'x 0.59 0.043 X < 1536 8 0 Mar Ecol Prog Ser 158: 75-86, 1997

Balanion comaturn A

l : &+ --@ Balanion cornaturn

l l I I Gyrnnod~nfumsp. B - ,I, ? -

U@ cc,

Rhodornonas sal~namlF1 Rhodomonas salina ml-'

Fig. 6. Yield as a function of prey concentration for (A]Bal- Fig 8 Specific clearance (body volumes h-') as a function of anion comaturn and (B) Gylnnodinium SP. Data points rep- prey concentration for (A)Balanlon cornaturn and (B)Gymno- resent treatment means * 1 SE dinium sp. Data points represent treatment means * 1 SE

Gymnodiniunl sp. cells (80%)were left after 150 h limit of prey capture was 4 pm ESD and the upper limit (Fig. 9). Formation of resting cysts was not observed in close to 10 pm ESD. Generally there is a good agree- any of the investigated species. ment between the prey size spectrum based on growth The prey size spectrum for Gymnodinium sp. is rate and the one based on ingestion rates. However, in almost identical to that of Balanion comatum (Fig. 10). The optimum prey size was 8 pm (ESD). The lower

7, I --c 6 - 0 5 - 4- -X 3- - Balanion cornaturn 1400 F 2- 5 t 3 1- 80- + Balanron cornaturn

0 20 40 60 80 100 120 140 160 180 hours

0 2 4 6 8 10 12 14 16 18 20 Fig. 9. Starvat~on.(A) Balanion cornaturn. The number of Specific ingestion rate (% h-') predator cells (a) and the volume of the starving predator (---a---).(B) Gymnodinium sp. The number of predator cells (0) Fig. 7. Growth rate (p) as a function of specific ingestion rate and the volume of the starving predator ( .o. ). Data points (U)for (A) Balanion cornaturn and (B)Cyrnnodjnjurn sp represent treatment means + 1 SE Jakobsen & Hansen: Comparative study of a ciliate and a dinoflagellate 81

3 4 5 6 7 8 91011 0 10 20 30 40 50 60 70 tlme (h) ESD (pm) Flg 11 Mixture experiments. (A) Cell concentration of Bal- Fig. 10. Prey size spectra. (A) Growth rate (p) (h-') as a func- anlon comatum in mlxed cultures with Gymnod~nlumsp, and tion of prey size (ESD)and (B)ingestion rate as a function Rhodomonas salina (a)and in the control experiment where of prey size (pm3 cell-' h-') for Balanion comatum (0) and Balanion cornatum only CO-occurswith R. saljna (M). (B) Cell Gymnodiniurn sp. (0). Data points represent treatment means concentration of Gymnodinium sp. in mixed cultures with 21 SE Balanion comatum and Rhodomonas salina (0) and in the control experiment where it only CO-occurswith R salina (U). In both plots, dotted and solid curves refer to cell concen- tration of R. saLina in the mixed culture and in the control, one case (Gymnodinium sp. fed Plagioselmisprolonga) respectively. Data points referto cell concentration t l SE a very high ingestion rate was not reflected in the growth rate, indicating a low growth efficiency. The average growth rates of Gj7mnodiniurn sp. and other objects, Gyn~nodiniurn sp. cells tumbled and sub- Balanion conlaturn obtained when they were grown to- sequently performeda burst with a maximum speed of gether on Rhodomonas salina did not differ significantly 6600 pm S-'. Thereafter, swimming speed decreased from growth rates obtained in cultures where they were asymptotically to 195 pm S-' or to passive drifting. The grown alone on R. salina (t-: Gymnodiniurn sp.,p = distance traveled during a burst was up to 400 pm. 0.8106, t = 0.246; B. cornatum, p = 0.7451, t = 0.334),in- Balanion comaturn swam in helices at an average dicating that B, comatum is unable to feed on Gymno- speed of 375 t 22 pm S-' (t l SE) corresponding to 19 dinium sp. and vice versa (Fig. 11, Table 3). body lengths S-'. Bursts were observed with a maxi- Gymnodinium sp. cells either drifted passively or mum speed of 1100 pm s-'. The burst was stopped swam in an almost straight path with an average speed drastically and the cell almost stopped swimming for of 195 * 10 pm S-' (* 1 SE) corresponding to 25 body some time (0.5 to 1 S) after which the cell resumed its lengths S-'. Driftlng was never observed in starved original swimmlng speed. The distance traveled dur- Gymnodinium sp. cells. Upon making contact with ing bursts was often up to 1000 pm.

Table 3. Balanion co~natun~and Gymnodiniurn sp. Growth rate (h-') in the mix- tul-e experiment No significant differences in growth rates were observed DISCUSSION between control and mixture experiments (for test used see text) n: number of replicates Prey selection

( Species Experiment Growth rate (h) r2 n 1 Gymnodinium sp. (7 pm) and Bal-

I I anion cornatum (17> urn),, both feed on Gymnodiniunl sp. Control 0.039 autotrophic nanoflagellates4 to 10 pm Gymnodinjum sp. Mixture experiment 0.039 Balanion cornaturn Control 0.051 in diameter and may be considered

I Balanlon cornaturn M~xtureexper~rnent 0 053 0 80 7 1 as competitors for similar-sized food. However, to what degree can this ob- 82 Mar Ecol Prog Ser 158: 75-86, 1997

servation be used to generalise on the competition A prey size spectrum has not previously been estab- between small dinoflagellates and ciliates? lished for prostomatid ciliates. However, the fresh- Gymnodinium sp. ingests single cells by direct en- water species Balanion planctonicum (15 pm) preyed gulfment, which is a common feeding mechanism upon a 10 pm (Miiller 1991, Sommaruga & among heterotrophic dinoflagellates (Gaines & El- Psenner 1993). while a marine Balanion sp. (34 pm) brachter 1987, Hansen 1991b). We found that Gymno- grew best on the 17 pm dinoflagellate dinium sp. had an optimum prey size corresponding triqueta (Stoecker et al. 1986). We found that the 17 pm approximately to its own size, which in specific terms Balanlon comatum grew best on Rhodomonas salina of is similar to that obtained for Gyrodinium spirale, 8 pm, indicating a general predator:prey ratio of 2:l for which also feeds using this feeding mechanism species belonging to the genus Balanion. This is in (Hansen 1992). Other feeding mechanisms found contrast to aloricate ciliates, which have an among heterotrophic dinoflagellates include pallium optimum predator:prey size ratio of about 8:l (Jonsson feeding and feeding tubes. These feeding mecha- 1986, 1987), but similar to loricate oligotrichs (tintin- nisms also allow the dinoflagellate to ingest relatively nids), which are able to ingest prey of a size which cor- large prey. In summary, predator:prey size ratios responds to a predator:prey size ratio of approx. 2.5:l found among the dinoflagellates range between 0.4:l (Heinbokel 1978). Thus, it appears that small (c20 pm) and ?:l (see Table 4). Hence, Gymnodinium sp. does heterotrophic dinoflagellates compete with small not differ from the large majority of dinoflagellates (<20 pm) prostomatids and loricate and aloricate olig- described so far. Note, however, that Noctiluca scintil- otrichs (40 to 60 pm) for prey in the 4 to 10 pm size range. lans is an exception by being capable of ingestion of It is surprising that Balanion comatum apparently a much broader range of prey due its ability to glue cannot catch Gymnodinium sp. cells, even though prey items into large packages (see Gaines & Gymnodinium sp. is of an optimal prey size for B. Elbrachter 1987, Buskey 1995). comatum. We believe that the reason for this is that Some studies have suggested that heterotrophic Gymnodinium sp. performs a fast escape response dinoflagellates ingest particles of the size of when making contact with objects, while the algae (Lessard & Swift 1985, Strom 1991). The study of used in the experiment do not, thereby making it Lessard & Swift (1985) demonstrated uptake of thymi- impossible for the ciliate to catch the dinoflagellate. dine-labelled bacteria in heterotrophic dinoflagellates, While data on swimming behavior in heterotrophlc but their investigation did take into account the follow- dinoflagellates are available in the literature (Jacobson ing: (1) Thymidine may have accumulated in the food & Anderson 1986, Strom & Buskey 1993), this is, to our chain. Thus, bacteria may have been fed upon by knowledge, the first time burst swimming has been heterotrophic nanoflagellates and ciliates, which sub- reported in dinoflagellates. We have observed burst sequently were fed upon by the heterotrophic dinofla- swimming in other small gymnodinoid dinoflagellates gellates. (2) Bacteria may be parts of aggregates of a (unpubl. obs.), indicating that this may not be an much larger size, making them available for hetero- unique trait for this species. trophic dinoflagellates. Strom (1991) documented feed- ing of a Gymnodinium sp. (12 pm) on the cyanobacte- ria Synechococcus sp. (1.2 X 2.4 pm) when presented In Bioenergetics mixture with the prymnesiophyte Isochrysis galbana (4.5 pm). However, Synechococcus sp. is a rather large The growth rate of the ciliate Balanlon comatum was autotrophic bacteria with a volume which is 20 to 30 higher than that of the dinoflagellate Gymnodinium times that of marine heterotrophic bacteria in natural sp. at prey concentrations which supported balanced environments (Lee & Fuhrmann 1987, Simon & Azam grovv.th (Figs. 3 & 4), indicating that the ciliate may 1989). potentially out-compete the dinoflagellate in natural An important observation is that the presently in- environments. The maximum growth rate of Gymno- vestigated Gymnodinium sp. has a lower prey size dinium sp. is close to the expected value which can limit of about 4 pm, suggesting that small dinoflagel- be calculated from published relationships between lates (

(Fenchel & Finlay 1983, Montagnes 1996 and refer- centration is being depleted during incubation In fact, ences therein). However, the maximum growth rate of often only the initial prey concentration is measured. B. cornatum is comparable to or even higher that that Thus, at present it is impossible to tell if there are sig- of other planktonic (and larger) ciliates feeding on nificant differences in threshold prey concentration for nanoplankton (Table 4). growth between clliates and dinoflagellates. Gymnodinium sp. and Balanion comatum have So far most laboratory experiments have been car- almost similar threshold prey concentrations for ried out under steady state conditions. However, growth, 18 and 11 pg C 1-', respectively (Table 4, steady state conditions are rarely found in nature, Fig. 4). Values of threshold prey concentrations for because pelagic environments are indeed heteroge- growth taken from the literature are shown in Table 4. neous in time and space (Andersen & Snrensen 1986, The threshold prey concentrations for growth of Gym- Owen 1989, Franks 1995). Also, selective by nodinjum sp. and B, cornatum are both at the lower metazooplankton may affect populations of ciliates end of reported values within the 2 groups. Consider- and small dinoflagellates differently. able variation is found in the estimates of the threshold prey concentrations for growth among both ciliates and heterotrophic dinoflagellates. However, data for Adaptations to a heterogeneous environment ciliates and heterotrophic dinoflagellates fall within the same range. Consequently, no relationship be- that live in heterogeneous environments tween predator size and threshold prey concentration have evolved adaptations to cope with fluctuations in is found in either of the groups (Table 4). The high vari- food availability, a phenomenon often referred to as a ability may reflect different 'strategies' within the feast and famine existence. Such adaptations can groups, but may equally be explained by the difficulty involve complex cycles (resting cysts and swarmer in obtaining reliable results. In fact, some of the ob- formation) and the ability to regulate metabolism when tained threshold values are extremely high (200 to food conditions change. 300 pg C 1-l). Laboratory data may overestimate the Formation of resting cysts is widespread among cili- threshold prey concentration if (1) the prey is of a sub- ates and heterotrophic dinoflagellates (e.g. Goodman optimal size or quality (see Table 4), (2) the prey is not 1987, Fenchel 1990). Resting cysts have not been evenly distributed during incubation or (3) prey con- reported among species in the genera Balanion and

Table 4 Maxiinum growth rates and threshold prey concentration taken from the literature for growth of heterotrophic dino- flagellates and planktonic ciliates. Literature values were converted to carbon assuming 0.12 pg C pm-3 (Strathrnann 1967). A conversion factor from volume to carbon of 0.2 pg C pm-3 was used to convert number of Rhodomonas salina to carbon in the present experiment (data not shown). Data on growth rates are adjusted to 15°C using Qlo = 2.8. nd = not determined

Species Predato~ Max Threshold Prey species ESD Source ESD growth (v) rate P (h 'l (~gC 1 '1 (pm) Dinoflagellates Gymnodinil~msp 7 0 043 18 Rhodomonas salina 8 This study Gyninodinlun? sp 12 0043 10 Isochrys~sgalbana/Synechococcus sp 4 5/1 5 Strom (1991) Gyrodinlu~nsplrale 27 0 027 246 Heterocapsa tnquetra 16 Hansen (1992) Oblea rotunda 23 0 017 20 Dltylum bnghtwelll~ 23 Strom & Buskey (1993) Oblea rotunda 23 0 010 50 DunabeUa tertiolecta 6 5 Strom & Buskev (1993) Protoper~drniumcrasslpes 73 0 008 247 polyedra 37 Jeong et a1 (1994) Protopendinlum cf d~vergens 61 0 013 262 Gonyaulax polyedra 37 Jeong et a1 (1994) Protoper~dlnlllrnhuben 48 0017 10 Dltylum bnghtwell!] 23 Buskey et a1 (1994) C~llates Eutlntinnus pectlnls 31 0054 14 lsochrysis galbana/Monochrys~sluthen nd He~nbokel(1978) Fd vella azonca nd 0 054 17 Heterocapsa tnquetra 16 Karniyama (1997) Fa vella ehrenbergil 57 0 034 6 Heterocapsa triquetra 16 Hansen (1995) Favella tara~kaens~s nd 0 048 10 Heterocapsa triquetra 16 Karnlyarna (1997) Tlnt~nnopslsacumlnata 24 0 043 15 Isochrys~sgalbana 5 Verity (1985) Tintinnopsls vasculurn 51 0050 15 D~cratenalnornate nd Venty (1985) Lohrnanlella splral~s 31 0037 8 Pyrarnirnonas sp 6 Jonsson (1986) Strob~hdiumneptunl 60 0 059 325 Chroomonas sahna 8 Montagnes (1996) Strob~bdumvenlhae 56 0 026 75 Isochrys~sgalbana Chroomonas sahna (1 l] 5 5 8 Montagnes (1996) Stromb~d~nopsischeshlr~ 63 0 037 6 Thalasaonra pseudonana 4 Montagnes et a1 (1996) Strornb~d~urnret~culaturn 42 0 049 8 Pyrarn~monassp 6 Jonsson (1986) Stl ombidium s~culm 38 0 023 16 4 Montagnes (1996) Bdlanlon comatum 17 0058 11 Rhodomonas sallna 8 Th~sstudy 84 Mar Ecol Prog Ser 158: 75-86, 1997

Gymnodinium and none of our isolates formed resting a non-tumbling swimming pattern, and a survival time cysts. However, we cannot totally exclude cyst produc- of 150 h. However, due to the production of swarmers tion in these genera, because cyst formation can be during starvation, the '' of the cell almost clone specific and even selected against within a doubles the potential distance travelled (200 m). In few generations (Fenrhol 1989). the case of Balanion comatum, the maximum distance Production of swarmers was only observed in travelled is about 75 m assuming a speed of 376 pm SF' Gymnodinium sp. in the present study. Reports on the and a survival time of 55 h. Thus, Gymnodinium sp. existence of swarmer cells is very rare in heterotrophic has potentially a competitive advantage compared to dinoflagellates. To our knowledge only the naked het- B. comatum when food is patchy in time and space. erotrophic dinoflagellates kofoidii (Morey- Gaines & Ruse 1980) and Gymnodinium fungiforme (Spero & Moree 1981) have been described to include Metazooplankton grazing on ciliates and swarmers in their life cycles. Production of swarmers small dinoflagellates is unknown in planktonic ciliates, but is common in benthic ciliates (e.g. Fenchel 1990). Grazing by metazooplankton on ciliates has been The ability to slow down metabolism to a minimum reported to range from insignificant to important is known among protists (e.g. Fenchel 1982b, 1989, (Stoecker & Sanders 1985, Wiadnyana & Rassoul- 1990). Some ciliates can survive for a period corre- zadegan 1989, Fessenden & Cowles 1994, Nielsen & sponding to 40 times their own minimum generation Kiarboe 1994). Field and laboratory experiments have time (Fenchel 1990). However, the few planktonic cili- shown that the greatest grazing impact on the ciliate ates studied so far are not able survive for very long stock by copepods is when concentra- when starved, only approx. 2 to 3 minimum generation tions are low and dominated by small phytoflagellates times (Fenchel 1989, Montagnes 1996). In this respect, (e.g. Jonsson & Tiselius 1990, Nielsen & Kierboe 1994, Balanion comatum is no exception; it is able to starve Atkinson 1996). However, the swimming behavior of for -4 times the minimum generation time. Reports on the ciliates also plays a role. Some ciliates, like Strobi- the ability of heterotrophic dinoflagellates to reduce lidium spp, and Mynonecta rubra, are able to escape metabolism when starved are sparse. The hetero- by burst swimming (3000 to 7000 pm S-') when trophic dinoflagellate Gyrodinium spirale can prolong attacked by metazoan predators (Jonsson & Tiselius survival by reducing ~tsmetabolism (Hansen 3.992). 1990, Gilbert 1994). Data on metazooplankton grazing However, the question of how long they are able to on small heterotrophic dinoflagellates are lacking. survive was not addressed. Jeong & Latz (1994) found However, laboratory experiments on prey size selec- that the large planktonic heterotrophic dinoflagellate tion by copepods suggest that particles the same size Protoperidinium divergens survived for at least 9 times as Gymnodinium sp. are retained with an efficiency its minimum generation time, while Strom (1991) which is less than 10 % of that obtained on particles the reported on a Gymnodinium sp. which was able to same size as Balanion comatum (Frost 1972, Nival & starve more than 30 times its minimum generation Nival 1976, Berggreen et al. 1988). In conclusion, the time. Our Gymnodinium sp. was able to starve for grazing impact on ciliates by metazooplankton might more than 10 times its own minimum generation time. under some conditions be much higher than that on Thus, although the information on the ability of plank- small heterotrophic dinoflagellates. tonic ciliates and heterotrophic dinoflagellates to pro- long survival is limited, the data suggest that hetero- Acknowledyements We are grateful to the Scandinavian trophic dinoflagellates may be able to cope better with Culture Collection, Botanical Institute, Department of Fungi and Algae, Copenhagen University, which provided most of starvation than planktonic cilidtes. the algal cultures. We are indebted to Christine Mwgelvang What are the benefits of prolonging survival for the Nielsen and Birgit Brander for technical assistance. We thank dinoflagellate? Generally, prolonged survival will give Tom Fenchel for suggesting the ingestion rate calculations the the organism more time to encounter patches in and Bent Vismann for setting up the computer model used in the calculation of feeding rates. We also thank David Mon- time or space. A compilation of swimming speed data tagnes and Cathenne Bernard for constructive criticism. This on ciliates and dinoflagellates presented in Buskey et work was funded by the Danish Natural Sclence Research al. (1993) and Hansen et al. (in press) suggest that cili- Council Contract No. 9502163-28808. ates generally swim 2 times faster than dinoflagellates, although the variation is large. The maximum distance travelled for an organism is a function of swimming LITERATURE CITED speed, time and tumbling frequency. In the case of Andersen P, Serensen HM (1986) Population dynamics and Gymnodinium sp., a cell subjected to starvation is able trophic coupling in pelagic in eutrophic to travel about 100 m assuming a speed on 195 pm S-', coastal waters. Mar Ecol Prog Ser 33:99-109 Jakobsen & Hansen: Comparative study of a ciliate and a dlnoflagellate

Atkinson A (1996) Subantarctic copepods in an oceanic, low ciliate. Mar Ecol Prog Ser 53:105-116 environment: ciliate predation, food selectivity Hansen PJ (1991a) -a planktonic dinoflagellate and impact on prey populations. Mar Ecol Prog Ser 130: genus which act both as a prey and a predator of a ciliate. 85-96 Mar Ecol Prog Ser 69:201-204 Beers JR, Stewart GL (1967) M~cro- in the Hansen PJ (1991b) Quantitative importance and trophic role euphotic zone at five locations across the California of heterotrophic dinoflagellates in a coastal pelagial food current. J Fish Res Bd Can 24:2053-2068 web. Mar Ecol Prog Ser 73:253-261 Berggreen U, Hansen B, Kiarboe T (1988) Food size spectra, Hansen PJ (1992) Prey size selection, feedlng rates and ingest~onand growth of the tonsa during growth dynamics of heterotrophic d~noflagellates with development: implications for determ~natlonof copepod special emphasis on Gyrodjnium splrale. Mar Biol 114: production. Mar Biol99:341-352 327-334 Bjarnsen PK, Kuparinen J (1991) Growth and herbivory by Hansen PJ (1995) Grocvth and grazing response of a ciliate on heterotrophic dinoflagellates In the southern , stud- the dinoflagellate Gyrodiniurn aureolum in mono- ied by microcosm experiments. Mar Biol 109:397-405 culture and in mixture with a non-toxic alga. Mar Ecol Buskey EJ (1995) Growth and of NOC~J~UC~ Prog Ser 121:65-72 scintillans on varying algal diets. J Plankton Res 17(1): Hansen PJ, Bjarnsen PK, Hansen B (in press) Zooplankton 29-40 grazing and growth: scaling within the 2-2000 pm body Buskey EJ, Coulter CJ, Brown SL (1994) Feeding, growth and size range. Limnol Oceanogr bioluminescence of the heterotrophic dinoflagellate Proto- Heinbokel JF (1978) Studies on the functional role of tintin- pendlnium huberi. h4ar Biol 121 373-380 nids in the Southern California Bight. 1. Grazing and Buskey EJ, Coulter C, Strom S (1993) Locon~otorypatterns of growth rates In laboratory cultures. Mar Biol 47:177-189 microzooplankton: potential effects on food selectivity of Jacobson DM. Anderson DM (1986) Thecate heterotrophic larval fish. Bull Mar Sci 53(1):29-45 dinoflagellates feeding behavior and mechanisms. J Phy- Eccleston-Parry JD, Leadbeater BSc (1994) A comparison c01 22:249-258 of the growth kinetics of six marine heterotrophic nano- Jeong HJ, Latz M1 (1994) Growth and grazing rates of the flagellates fed with one bacterial species. Mar Ecol Prog heterotrophic dinoflagellate Protoperidinium spp. on red Ser 105:167-177 tide dinoflagellates. Mar Ecol Prog Ser 106:173-185 Edler L (ed) (1979) Recommendations on methods for marine Jonsson PR (1986) Particle size selection, feeding rates and biological studies in the Baltic sea. Phytoplankton and growth dynamics of marine oligotnchous ciliates (Cilio- chlorophyll. The Baltic Marine Biologist publication, phora: Oligotrichina). Mar Ecol Prog Ser 33:265-277 Malmo no. 5:l-38 Jonsson PR (1987) Photosynthetic assimilation of inorganic Fenchel T (1982a) Ecology of heterotrophic microflagellates carbon In manne oligotrich ciliates (Ciliophora, Oligo- 11.B~ocncrgetics and growth. Mar Ecol Prog Ser 8:225-231 tr~china).Mar Microb Food Webs 2 55-68 Fenchel l (lY82b) Ecology of heterotrophlc m~croflagellates Jonsson PR,Tisellus P (1990) Feeding behaviour, prey detec- 111. Adaptations to heterogeneous environments. Mar Ecol tion and capture efficiency of the copepod Acartla tonsa Prog Ser 9:25-33 feedlng on planktonic ciliates. Mar Ecol Prog Ser 60:35-44 Fenchel T (1989) Adaptations to a feast and famine existence Kamiyama T (1997) Growth and grazing response of tlntinnid in protozoa. In: Wieser W, Gnaiger G (eds) Energy trans- ciliates feeding on the toxic dinoflagellate Heterocapsa formation in cells and organisms. Georg Thieme Verlag, circularisquama. Mar Biol 128509-515 Stuttgart, p 290-295 Lee S, Furhman JA (1987) Relationships between biovolume Fenchel T (1990) Adaptive significance of polymorphic life and biomass of naturally derived marine bacterioplank- cycle in protozoa: responses to starvation and refeeding ton. Appl Environ Microbiol53(6):1298-1303 in two species of marine ciliates. J Exp Mar Biol Ecol 136: Lessard EJ (1991) The tropic role of heterotrophic dinoflagel- 159-177 lates in diverse marine environments. Mar Microb Food Fenchcl T, Finlay BJ (1983) Respiration rates in heterotrophic, Webs 5(1).49-58 free-llving protozoa. Microb Ecol 9:99-122 Lessard EJ, Swift E (1985) Species-specific grazing rates of Fessenden L, Cowles TJ (1994) Copepod predation on phago- heterotrophic d~noflagellatesin oceanic waters, mesaured trophic ciliates in Oregon coastal waters. Mar Ecol Prog with a dual-label radioisotope technique. Mar Biol 87. Ser 107:103-l11 289-296 Franks PJS (1995) Thin layers of phytoplankton: a model of Montagnes DJS (1996) Growth responses of planktonic cili- formation by near-interstitial wave shear. Deep Sea Res ates in the genera Strobihdiurn and Stromb~dium.Mar 42:75-91 Ecol Prog Ser 130:241-254 Frost BW (1972) Effects of size and concentration of food par- Montagnes DJS, Berger JD, Taylor FJR (1996) Growth rate of ticles on the feeding behavior of the marine planktonic the marine ciliate Strombidinopsis cheshiri Snyder and copepod Calanus pacificus. Limnol Oceanogr 172305-815 Ohman as a function of food concentration and interclonal Gaines G, Elbrachter M (1987) Heterotrophic nutrition. In: variability. J Exp Mar Biol Ecol 206:121-132 Taylor FJR (ed) The biology of dinoflagellates. Blackwell, Montagnes DJS, Lynn DH, Roff JC, Taylor W (1988) The Oxford, p 224-268 annual cycle of heterotrophic planktonic ciliates in the Gilbert JJ (1994) Jumping behavior in the oligotrich ciliate waters surrounding the Isles of Shoals, Gulf of Maine: an Strobllidlum velox and Halteria grandinella, and its sig- assessment of the11 tropic role. Mar Biol88:21-30 nificance as a defense against predators. Microb Morey-Gaines G, Ruse RH (1980) Encystment and reproduc- Eco127:189-200 tion of the , Polykrikos kofoidi Goodman DK (1987) Dinoflagellate cysts in ancient marine Chatton (). Phycologia 19:230-236 and modern marine sediments. In Taylor FJR (ed) The Miiller H (1991) Pseudobalanion planctonicum (Cdiophora, biology of dinoflagellates. Blackwell, Oxford, p 649-722 Prostomatida): ecological significance of an algivorous Hansen PJ (1989) The red tide dinoflagellate Alexandrium nanociliate in a deep meso-eutrophic lake. J Plankton Res taniarense: effects on behaviour and growth of a 13:247-262 86 Mar Ecol Prog Ser 158: 75-86. 1997

Nielsen TG, Kiorboe T (1994) Regulation of zooplankton bio- Stoecker DK, Cucci TL, Hulburt EM, Yentsch C\.? (1986) mass and production in a temperate, coastal ecosystem. 2. Selective feeding by Balanion sp. (: Balanionidae) Ciliates. L~mnolOceanogr 39508-519 on phytoplankton that best support its growth. J Exp Mar Nival P, Nival S (1976) Particle retention efficiencies of a Biol Ecol 95.113-130 herbivorous copepod, Acartia clausii (adult and cope- Stoecker DK, Sanders NK (1985) Differential grazlng by Acar- podite stages): effects on grazing. L~mnolOceanogr 21. tia tonsa on a dinoflagellate and a tintinnid. J Plankton Res 24-38 7:85-100 Owen RW (1989) Microscale and finescale variations of small Strathmann RP (1967) Estimating the organlc carbon content plankton in coastal and pelagic environments. J Mar Res of phytoplankton from cell volume or plasma volume. 47:197-240 Limnol Oceanogr 12:965-977 Sherr EB, Sherr EB (1994) Bacterivory and herbivory: key Strom SL (1991) Grazing and growth rates of the herbivorous roles of phagotrophic protists in pelagic food webs. dinoflagellate Gymnodinium sp. from the subarctic Pacific M~crobEcol 28.223-235 Ocean. marEcol Prog Ser 78:103-113 Simon M, Azarn F (1989) Protein content and protein synthe- Strom SL, Buskey EJ (1993)Feeding, growth, and behavior of SIS rates of planktonic marine bacteria. Mar Ecol Prog Ser the thecate heterotrophic dinoflagellate Oblea rotunda. 51:201-213 Limnol Oceanogr 38:965-977 Smetacek V (1981) The annual cycle of protozooplankton in Verity PG (1985) Grazing, respiration, excretion, and growth the Kiel Bight. Mar Biol63:l-11 rates of . Limnol Oceanogr 30: 1268- 1282 Somrnaruga R, Psenner R (1993) Nanociliates of the order Verity PG (1991) Measurements and simulation of prey up- Prostomatida their relevance in the microbial of take by marine planktonic ciliates fed plastidic and aplas- a mesotrophic lake. Aquat Sci 55.179-187 tidic nanoplankton. L~mnolOceanogr 36.729-750 Spero HJ, Moree MD (1981) Phagotrophic feeding and its Wiadnyana NN, Rassoulzadegan F (1989) Selective feedmg unportance to the life cycle of the holozoic dinoflagellate of Acartla clausi and Centropages typicus on rnicrozoo- Gymnodinium fungiforme. J Phycol 17:43-5 1 plankton. Mar Ecol Prog Ser 53:37-45

Editorial responsibility: Otto Kinne (Editor), Submitted: July 14, 1997; Accepted: September 22, 1997 Oldendorf/Luhe, Germany Proofs received from author@): November 3, 1997