Differential growth of Euplotes vannus fed fragmented versus unfragmented chains of Skeletonema costatum

The major food source of marine c~liates has traditionally been thought to consist of small nano-sized prey items (including diatoms), and bacteria. Larger prey, including chain-forming diatoms, have been considered important dietary items for certain ciliates, particularly larger benthic species, but may provide a source of nutrition for a wider range of ciliate species if they are first modified in nature by feeding or environmental turbulence-related fragmentation. Such fragmentation of diatom chains by microcrustaceans has been documented in the literature. To begin to examine what effects algal fragmentation might have on ciliate growth. several experiments were conducted in which Euplotes vannus were presented with equal amounts of fragmented and unfragmented Skeletonema costatum as food. In all cases, growth of the ciliates was enhanced in the presence of the fragmented food. It is known that ciliates, especially benthic forms, can ingest cells their own size and larger (Kahl 1935, Smetacek 1981). Ingestion of diatoms has been reported for hypotrich, heterotrich, gymnostome, trichostome, hymenostome, odontostome, oligotrich and choreotrich ciliates found associated with benthic sediments (Borror 1963, Fenchel 1968, 1969, 1987). For planktonic ciliates, Verity & Villareal (1986) found that tintinnid ciliates could ingest certain diatoms laclung spiny projections, and obtain levels of nutrition sufficient for growth. Smetacek (1981) observed Strombidium spp. with protruding ingested cells, at times larger than the ciliate predator. Strombidium sp. has also been observed with large numbers of diatoms in its food vacuoles (Fenchel 1987). Nonetheless, current marine microbial food web models rarely consider potential pathways between diatoms, particularly chain-forming species, and ciliates. This is primarily due to the differences in their relative sizes, with many diatoms or their composite chains being in size equal to, or larger than, many ciliates. Present address: Long Island Sound Taskforce (Oceanic Society), 185Magee Ave., Stamford, Connecticut 06902, USA O Inter-Research/Printed in F. R. Germany Various disruptive mechanisms, both biological and physical, exist in nature, by which natural food assemblages, including diatom chains, can have their size distribution altered toward a lower mean. For example, routine and episodic turbulent mixing of sediment in coastal waters might fragment cells. Additionally, estuarine copepods have been observed to modify large single cells and chain-forming diatoms, both in laboratory cultures and natural particle assemblages, during feeding, resulting in a higher final concentration of small particles (O'Connors et al. 1976, Deason 1980, Roman & Rublee 1980). Many of the particles so produced are of a size that could easily be ingested by ciliates. Thus particle modification may represent a pathway by which large single-celled or chain-forming algae become available to both planktonic and benthic protozoan predators at a more 'palatable' size. Such a pulse of appropriatelysized food might enhance their growth. The hypothesis that fragmentation of diatom chains enhances ciliate growth rate, by providing them with food items that would otherwise be unavailable to them, was tested using the hypotrich marine ciliate Euplotes vannus Miiller, and the chain-forming diatom Skeletonema costatum. Euplotes vannus was chosen as the test organism for this study primarily because it readily associates with surfaces. This behavior insured that the non-motile Skeletonema costatum cells (which accumulated on the bottom of the spot plate wells used in these experiments) would remain available to the ciliate predator. E. vannus is a behaviorally multidimensional organism found in a plethora of habitats including surface and interstitial marine benthic habitats, salt marshes, estuarine plankton and littoral zone assemblages, as well as in freshwater lake, pond and stream plankton, neuston and benthos (Barnforth 1985), and marine snow (Caron et al. 1982). Mar. Ecol. Prog. Ser. 47: 205-209. 1988 Materials and methods. Cultures of Euplotes vannus (obtained from A. Repak, Quinnipiac College, Hamden, Connecticut, USA), Dunliella tertiolecta and Skeletonema costatum (from Guillards's Bigelow Laboratory collection, Maine, USA) were maintained in DV medium (Provasoli 1963) adjusted to a pH of 7.5 to 7.8, in 125 m1 flasks for the ciliates, and in 30 m1 screwcap test tubes for the algae. Expt 2, however, employed S. costatum cultures isolated (by G.M.C.) from western Long Island Sound, New York, USA, in February 1986. E. vannus cultures were kept at room temperature and the algae a t 18°C under a 14/10 LD cycle. Prior to an experiment E. vannus were maintained on a diet of D. tertiolecta. All experiments were run at room temperature using exponentially growing S. costatum as food. Skeletonema costatum chains were fragmented using a sand vortexing technique. Beach sand (from Shinnecock beach, Long Island) was cleaned of organics by soaking in 10 % HCl for 24 h , followed by a rinsing in organic-free, deionized water (ODW), a subsequent soaking in a weak NaOH bath, and a final ODW water rinse. The cleaned sand was then oven dried. Diatom chains were modified by introducing 3 m1 of sand into a 30 m1 screw-capped test tube, adding 7 m1 of S. costatum culture, and vortexing at a standard 4 setting, for 60 S using a Vortex-Genie. This procedure was repeated and the treatments pooled to provide about 40 m1 of treated culture. For the controls, 7 m1 of S. costatum was added to 3 m1 of sand without subsequent vortexing. Instead, the tube was gently inverted twice to mix sand and culture. This procedure was repeated until about 40 m1 of control culture was achieved. Cell counts were then taken for both control and treatment cultures. Numbers of cells per chain and total numbers of chains were recorded for each culture. Control suspensions were diluted with DV medium, when necessary, to equalize control and experimental algal concentrations. The above method effectively modifies long chains to shorter ones (Figs. lA , B) . After modification, control and treatment Skeletonema costatum cultures were transferred to individual spot plate wells (0.70 m1 per well). Ciliates from stock cultures were then introduced, 1 individual per well, into the spot plates, via micropipetting. Prior to introduction into experimental spot plate wells, each ciliate was taken through 3 washes of sterile DV to remove the Dunaliella tertiolecfa cells. Cover slips were then placed over each plate well onto a thin encircling bead of Dow Chemical stopcock grease, to prevent evaporation while permitting the exchange of respiratory gases. The experiments utilized 2 plates (9 wells per plate, i.e. 18 replicates) for each control (unmodified diatoms) and treatment (modified diatoms) condition (Expt 1 treatment condition employed 15 replicates). All data represent averages per well based on the 15 or 18 replicates. Ciliate numbers per well were followed over time for both controls and treatments, using a Wild M-5 dissecting microscope set on dark field. Their numbers at times exceeded 40 cells per well and occasionally had to be fixed at the termination of an experiment to ease counting. After completion of an experiment, numbers and distributions of Skeletonema costatum chain lengths were measured. A minimum of 50 cells was counted for the initial samples. This was not possible for post-experimentation treatment samples, since most cells had been eaten during the experiment (Figs. 2A, B). In these instances 2 full hemacytometer grids (4 fields, volume = 9 X I O ~ ml) were counted. Growth rates were calculated using the instantaneous growth equation N, = Noer' t (Capriulo et al. 1983) repetitively for each time interval, as an initial value differential equation. By solving for r one interval at a time, it is necessary to assume that growth is exponential with a fixed value of r only for that interval. For the next interval r may take on any new value. In this way changes in rover time can be studied. Results. The vortexing of the Skeletonema costatum cultures produced a distribution of cells dominated by single cell and 2 cell-long chains, with some 3, 4 or 5 cell-long chains also present (Figs. lA, B). This represented a major redistribution of chain lengths and provided Euplotes vannus with an altered food size spectrum, relative to the control cultures. In both experiments, ciliates provided with this modified food sustained higher numbers (Figs. l C , D) and exhibited higher growth rates (Figs. l E , F) than those ciliates fed unmodified food. The control and treatment density response data were different from each other at the 0.09 and 0.005 levels of significance for Expts 1 and 2, respectively, based on the Wilcoxon 2-sample test for differences in ranked observations (Sokal & Rohlf 1969). The average cumulative numbers of cihates in Expt 1 reached a peak of 14.93 per well (f SE = 2.76) for the treatment and 10.22 (f SE = 1.82) for the controls, with corresponding average instantaneous growth rates (r) over the 230 h experiment of 0.015 and 0.010 h-', respectively. Initially both control and treatment responses were similar. A devlation in growth rates began on the 3rd day. Differences in growth between treatment and control wells were greatest at this time, with r = 0.0346 and 0.0190 h-', respectively. By the 5th day of observation no differences in r values could be detected, probably due to food limitations. The analogous values for Expt 2 ( t = 65 h) were 29 (k SE = 2.29) and 19 (+ SE 1.07) chates per well for treatment and control, with respective average i s of 0.045 and 0.040 h-'. The largest difference in r values Capriulo et al.: Diatom chain fragmentation 207