During April, June/July and September 1986, phytoplankton biomass was followed in marine enclosures in the Roskilde Fjord, Denmark, manipulated by the addition of benthic suspension feeders, planktivorous fish, nutrients and contact to the sediment. During periods with high insolation and temperatures from 11 to 22 "C, the manipulations caused marked changes in the development of phytoplankton biomass. Additions of Mytilus edulis reduced phytoplankton biomass to between 10 and 59 % of controls, whereas addition of nutrients raised phytoplankton biomass to an average of 256 % of controls. Generally, low growth rates of mussels were found in enclosures containing mussels alone. Addition of planktivorous fish and nutrients increased growth rates of mussels. During June/July, when inorganic nitrogen limited phytoplankton growth, autotrophic picoplankton (1 to 2 pm cell diameter) constituted 70 to 93 % of phytoplankton biovolume in enclosures containing mussels compared to 4 to 20 % in controls. The mussels reduced phytoplankton biomass by only about 50 % during this period, presumably due to low retention efficiency of the small cells. In April and September, however, when nitrogen did not control phytoplankton growth, picoplankton comprised < 0.001 % of phytoplankton biovolume. During September (temperatures 11 to 13OC), M. edulis reduced chlorophyll levels to 10 O/ O of controls and the effects of nutrient additions were significantly reduced in enclosures containing mussels. The effect of fish additions revealed that zooplankton grazing removed (average for June and September) about 20 % of phytoplankton biomass. The sediment acted prin~arily as a nutrient source during summer However, in April at temperatures > 6 "C and in September, benthic suspension feeders maintained chlorophyll levels below those in enclosures with no sediment contact. During April, only minor changes were recorded in the chlorophyll levels in the various enclosures, presumably due to the low temperature (0 to 4.5OC) and a phytoplankton population of decaying diatoms. Thus, several factors or combinations of factors controlled the phytoplankton biomass in the enclosures. The most important of these factors was the balance between nutrient input, phytoplankton size structure and the physiological state of the mussels.

Enclosure experiments were conducted during April. June/July and September in the eutrophc estuary Roskilde Fjord, Denmark, to reveal the effects of inorganic nutrients, suspensionfeedng bivalves Mytilus edulis and planktivorous fish (three-spined sticklebacks Gasterosteus aculeatus) on the zooplankton community > 45 pm. The additlon of inorganic nutrients did not increase zooplankton biomass although it did increase the chlorophyll level, indicating that zooplankton production was not food hmited. Filtration by M. edulis reduced the number of tintinnid ciliates and rotifers during all 3 experiments, but not the abundance of the larger zooplankton species. Additions of planktivorous fish reduced the densities of larger zooplankton species Acartia tonsa and Pleopis polyphernoides but not of smaller species. An immense increase in numbers of A. tonsa and P. polyphernoides was observed in enclosures without fish, indicating that the larger crustacean zooplankton is strongly predator controlled. Thus, the qualitative and quantitative development of the zooplankton community in the enclosures was controlled in 2 ways; from the top of the size spectrum by G. aculeatus and from the bottom of the size spectrum by M. edulis.

During three periods of 16 to 25 days, bacterioplankton production, bacterial cell volume, chlorophyll a, CO2 assimilation, and particulate organic carbon were measured in enclosures situated in the eutrophic estuary Roskilde Fjord, Denmark. The enclosures were manipulated with respect to sediment contact and contents of inorganic nutrients, planktivorous fish, and suspension-feeding bivalves. Nutrient enrichment, the presence of suspension feeders, and sediment contact induced pronounced changes in bacterial production, as well as minor changes in bacterial cell volume; however, these effects seemed to be indirect, transmitted via phytoplankton. Bacterial production, measured as [3H]thymidine incorporation, closely followed changes in phytoplankton biomass and production, with time lags of 5 to 10 days. Good correlations of mean bacterioplankton production to chlorophyll a concentration and CO2 assimilation suggested phytoplankton to be the dominating source of bacterial substrate, apparently independent of nutrient stress. Zooplankton >140 jum, bivalves, and sediment seemed to provide insignificant, if any, substrate for bacterioplankton, and benthic suspension feeders seemed not to act as direct competitors for dissolved organic carbon. The bacterioplankton mean cell volume, measured by image analysis, changed seasonally, with the smallest cells during the summer. Within each period, the bacterial cell volume correlated positively to growth rate and negatively to temperature.