Automated determination of critical oxygen concentration for routinely active fish

Abstract

We describe an automated respirometer and control system that determines critical (metabolism-limiting) oxygen concentration for routinely active fish. A microcomputer monitors fish metabolic rate as oxygen concentration in the closed respirometer declines; the critical oxygen concentration is signaled by metabolic-rate change, which is resolved via statistically based rules that consider both magnitude and consistency of rate deviations. After the critical oxygen concentration is found, data are written to disk, the respirometer is reoxygenated and another trial is initiated. This sequence can be repeated indefinitely without human intervention, allowing replicate estimates from a single fish. Ideally, metabolic rate should be calculated frequently during a trial, to minimize exposure of the fish to sub-critical oxygen concentrations. However, precision of measurement is limited by ‘noise’ related to length of time interval over which changes in oxygen concentration are determined, respirometer chamber volume, and fish respiration rate. Short time intervals lead to excessive noise, whereas long time intervals result in insufficient numbers of rate measurements. In a respirometer chamber of calculable optimum volume, measurements made by averaging oxygen readings taken very rapidly over two to three minute intervals provide a good compromise. We present data from experiments with bluegill,Lepomis macrochirus, to illustrate the method and show that critical oxygen concentrations identified by the system are consistent with estimates made by humans viewing graphs of the same experiments