Bacteria–flagellate coupling in microcosm experiments in the Central Atlantic Ocean

The coupling between planktonic bacteria and bacterivorous protozoans was examined in microcosm experiments at several oligotrophic and ultra-oligotrophic sites in the subtropical and 5 tropical Atlantic Ocean. Bacterial concentrations at these stations were in the range 2.2–8.1 3 10 21 21 cells ml , heterotrophic nanoflagellates (HNF) in the range 100–800 cells ml , bacterial doubling times (estimated from leucine incorporation) in the range 1–100 days, and chlorophyll a 21 levels in the range 0.03–0.36 mg l . The experimental uncoupling of the microbial loop by differential filtrations did not result in an increased growth and grazing by nanoflagellates despite a stimulation and increase of bacterial abundance and mean cell volume due to the bottle incubations. A strong response of the grazer population occurred after increasing bacterial numbers about 10-fold by the addition of a complex substrate source (yeast extract). Bacteria responded immediately to the substrate enrichment with an increase in mean cell size and abundance, and reached stationary phase already after about 24 h. In contrast, HNF development showed a pronounced lag phase, and it needed between 3 and 7 days until grazers reduced bacterial numbers to about the initial values. The grazing impact on the bacterial assemblage in the bottles resulted in feed-back effects that resembled those known from other, more productive systems: protozoan size-selective grazing removed preferentially larger sized bacteria and shifted the size-distribution towards the initial, natural situation with a dominance of small cocci. Grazing-resistant morphotypes consisted of bacterial aggregates embedded in a polysaccharide matrix whereas filamentous forms did not develop. These experiments provide evidence that bacterial assemblages have the capacity to respond to enhanced substrate availability (for example in micropatches) and to utilise these substrates without significant grazer control.  2000 Elsevier Science B.V. All rights reserved.