THE DURATION OF COPEPOD LIFE STAGES ESTIMATED FROM STAGE-FREQUENCY DATA

Current methods to determine the stage duration of copepods from stage-frequency data are often based on inappropriate distribution functions of the stage development time; they also lack a clear probability model that describes the random noise in the data. This study aims to estimate the duration of copepod life stages, using an asymmetrical probability distribution function of development time. Data on stage frequency were collected from cultured cohorts of Acartia clausi. Temora longicornis and Pseudocalanus elongatus. In addition, data from cultures of Calanoides carinatus and Calanus australis by Peterson and Painting (J. Plankton Res.. 12, 283-293, 1990) were used. Following Landry (Limnol. Oceanogr., 20. 854-858, 1975). for each stage the number of animals that did not yet pass that stage was plotted as a function of time. A gamma distribution function appeared to describe the expected proportion of these animals versus time. The number of animals is assumed to follow a binomial distribution and the model parameters are estimated by maximizing the likelihood. This method enables individuals with a retarded development, which are usually deleted from the data material, to be dealt with. The median development time was estimated from the gamma distribution functions. from which the duration of the stages was derived. Isochronal development was confirmed for most stages. However, the first feeding naupliar stage. as well as the last copepodite stage (CV), took longer in most species, whereas the pre-feeding naupliar stages had a shorter duration, particularly in P. elongatus. Differences between sexes were not apparent. Contradicting results among cultures and with published results by other workers suggest that part of the observed deviation from near-isochronal development may be due to the culture conditions.