Field and laboratory experiments were conducted to compare the rates of ingestion of planktonic protozoa for chroococcoid cyanobacteria and heterotrophic bacteria, and the fate of this ingested biomass. Laboratory experiments tested the ability of cyanobacteria and bacteria to support the growth of 3 species of bacterivorous protozoa. Two species of heterotrophic bacteria supported faster growth rates and higher cell ylelds of the protozoa than 3 strains of cyanobacteria. When mixtures of bacteria and cyanobacteria were offered, however, all protozoa grew as rapidly as when bacteria were offered alone. One protozoan showed a marked feeding selectivity against 1 strain of cyanobacteria when offered mixtures of bacteria and cyanobacteria. Grazing rate measurements performed in Vineyard Sound, Massachusetts, USA, revealed removal rates as hlqh as 54 O/ O of the cvanobacterial assemblage d-' and 75% of the heterotrophic bacterial assemblage d-' ~a n o~l a n k t o n i c protists (orqanisms < 20 urn) were the maior consumers of both cyanobacterial and bacterial biomass in this . -environment on 5 sampling dates.-Based on measuremenis of the ingestion rates of nanoplanktonic consumers and the carbon content of cyanobacteria and heterotrophic bacteria, we conclude that cyanobacterial biornass in this coastal environment at times reaches 30 % of the total prokaryote biomass consumed by the nanoplankton. During times of peak abundance of chroococcoid cyanobacteria, this biomass is an important source of organic carbon for planktonic protozoa feeding on bacteriasized particles.

Carbon cycling in a 3-member food web containing a diatom (Phaeodactylurn tricornutum), bacteria, and a herbivorous/bacterivorous microflagellate (Paraphysornonas jmperforata) was examined. Ingestion of prey by the microflagellate was the primary mechanism for remineralization of particulate organic material. Approximately 65 % of the particulate organic carbon (POC) initially present was lost over the course of the 8 d experiments in cultures containing microflagellates. No significant increase in remineralization was observed when bacteria were present. Bacteria were responsible for the uptake of dissolved organic carbon (DOC), but their overall contribution to carbon cycling was small relative to that of the microflagellate. Microflagellates incorporated diatom and bacterial biomass with equal efficiency (44 %) during exponential growth. Only 10 % of the POC ingested by microflagellates was released as DOC while 10 % was released as egested POC. The relatively high weight-specific respiration rate of the microflagellates (X = 2.67 X nl O2 h-') coupled with their relatively small release of DOC indicates that herbivory by heterotrophic microflagellates may be a major mechanism for the regeneration of nutrients from living phytoplankton which circumvents bacterial decomposition.

The uptake of fluorescent particles by protists and filter-feeding metazoa is being used increasingly by microbial ecologists to study feeding behavior and measure grazing rates. Recent studies of microflagellate uptake of these inert particles have yielded inconsistent results. In particular. grazing rates determined from fluorescent particle uptake are often less than rates measured using other techniques. These low uptake rates have been attributed to osmotrophy, food quality or size selection, rapid egestion of inert particles, and the slower feeding by free-living, as opposed to attached, protists. We have found that a variety of flagellates egest food vacuole contents upon fixation with several commonly used agents including glutaraldehyde and formaldehyde. During time course experiments, the observed microsphere uptake rate for a small chrysomonad flagellate using 1 % glutaraldehyde was only 6 % of the rate obtained by using the fixation method of van der Veer (1982) (2 '10 acrolein, 2 % glutaraldehyde and 1 % tannic acid), modified for epifluorescence microscopy. Uptake rates of several mixed flagellate populations also were 2.4 to 3.1 times higher using the modified van der Veer method than with 1 % glutaraldehyde. The average number of ingested microspheres cell-' using this method was simllar to that observed in live cells immobilized with NiS04. Glutaraldehyde also caused the egestion of Synechococcus sp. cells and fluorescently labelled bacteria from the chrysomonad flagellate. We conclude that previous studies using common aldehyde fixation wlth particle uptake for measunng rates of microflagellate bacterivory have significantly underestimated actual rates of consumpt~on, and that these studes must be re-evaluated, and perhaps repeated, using effective fixation methods.

The phagotrophic microflagellate Paraphysomonas imperforata was capable of grazing 2 marine phytoplankton species, the diatom Phaeodactylum tricornutum and the chlorophyte Dunaliella tertiolecta. The phytoplankton species, which are grossly different in size, shape and morphology, were first grown under different degrees of nitrogen and phosphorus limitation. Patterns of nutrient regeneration appeared to be a function of the physiological state of the prey: lags in NH: regeneration untd the end of exponential growth occurred when the prey were N-limited and phosphorus regeneration was negligible dunng the entire growth cycle of the microflagellate when the prey were P-limited. There was, however, evldence for dark uptake of NH; by N-limited prey that were ungrazed, which could have biased the observed lags in nutrient regeneration. Regeneration efficiencies reached as high as 70 O/O for nitrogen and up to 50 % for phosphorus only after prolonged penods in the stationary phase. Because protozoa are able to convert prey nutrients to their own biomass with great efficiency during exponential growth, particularly when the prey are nutrient-limited, the size and complexity of the microbial food web may be related to the nutritional state of the phytoplankton.

As part of a series of grazing experiments in batch cultures, we found that the phagotrophlc microflagellate Paraphysomonas imperforata, while grazing on the diatom Phaeodactylum tn'cornutum or bacteria, was responsible for the bulk of phosphorus regeneration. Regeneration of soluble reactive phosphorus (SRP) and dissolved organic phosphorus (DOP) was negligible in control cultures of the diatom alone, bacteria alone, or the 2 microbes together. When the m~croflagellate grazed on prey organisms that had been precultured with excess nutrients or under nitrogen limitation there was considerable regeneration of total dissolved phosphorus (TDP = SRP + DOP). Rates of TDP regeneration were greatest during exponential growth of the microflagellate and then decreased through the transition and stationary phases. Overall, up to 70 % of the phosphorus initially incorporated into prey biomass was regenerated through the stationary phase. Total excretion of DOP was about 15 to 20% of TDP, although DOP excretion made up a larger fraction of total phosphorus excretion for short periods during the exponential phase of growth. When the prey were phosphoruslimited virtually no TDP was excreted throughout the entire growth cycle of the microflagellate. Our results indicate that Protozoa have higher weight-specific rates of phosphorus excretion than do Metazoa. Although metabolic activity is not the sole indicator of the role Protozoa play in the nutrient regeneration process, our results, together with those from size-fractionation studies on nutrient regeneration, point toward a major role for Protozoa in pelagic waters where they constitute a large fraction of the zooplankton biomass.