Sampling programmes for monitoring threatened species require some repeatable and semi-quantitative sampling strategy to provide reliable estimates of population density and long-term demographic trends. However, quantitative sampling in low-density populations or highly heterogeneous habitats represents a considerable challenge. We have attempted to quantify the required sampling effort for monitoring populations of threatened mudfish using Gee minnow traps from 12 years of field data and controlled, replicated laboratory studies of Gee minnow trap efficiency. Conservation managers need to define the required measure of data precision necessary to obtain useful measures of long-term population trends. The application of a prescribed sampling design needs to consider not just population density and catch precision and variation, but also capture efficiency and fish behaviour.