How to estimate life history stage durations from stage structured population data

A method is presented for estimating time-dependent stage durations using complete census data from a stage-structured population. The data consists of counts of a stage population at a series of sample times and counts of the number of deaths within, recruitments to and maturations from, the stage between sample times. The data are analysed using spline functions to estimate continuous rates of recruitment, maturation and death. The stage population is modelled using a delay differential equation formulation, which allows the derivation of a stage duration estimator. The ability of this estimator to reconstruct stage durations is tested by simulation, both in the large sample limit and the far more stringent case of small population size. The method is applied to some data from a Notonecta stock tank experiment.