Preservation and incubation time-induced bias in tracer-aided grazing studies on meiofauna

A recent review suggests that meiofauna are important grazers of microphytobenthic primary production as well as of bacterial secondary production. The potential importance of meiofauna grazers may nevertheless have systematically been underestimated, since label leakage from chemically preserved animals has hitherto not been accounted for. Furthermore, a majority of studies have used relatively long incubation times and assumed, rather than proved, that label recycling over this period is negligible. In the present study we tested the influence of sample preservation on label retention in the marine nematode Pellioditis marina Andrassy, 1983 fed H-3-labelled bacteria. Label loss from formaldehyde-preserved specimens averaged 40% after 1 h preservation and amounted to a maximum of 85% after 24 h in formaldehyde, irrespective of formaldehyde concentration; no further leakage occurred beyond 24 h. Glutaraldehyde and ethanol yielded significantly better and poorer results, respectively, but the former fixative still yielded label losses of up to 70%. A comparison of label uptake as a function of time with observations on ingestion and defecation behaviour suggest that on time scales of hours an indication of assimilation (after correction for label leakage) rather than of ingestion is obtained. When killed with formaldehyde at room temperature, P. marina egested a significant part of its gut contents. The sources of bias identified here may have generally led to significant underestimations of true grazing rates. The cumulative effect of label leakage, prey egestion and long incubation times, each at the highest rates observed in this study, may yield as much as a 15-fold underestimation of true food consumption. Cooling samples on ice and fixation with ice-cold formaldehyde, followed by immediate freeze-preservation, and sorting of the nematodes within 2h after thawing, gives average values for label leakage of 50%, and hence allows the application of a proximate correction factor for label losses of 2.