Comparison of methods to estimate the secondary production of a populationof the polychaete Ditrupa arietina (O.F. Muller).

We compared the ability of three direct methods (increment summation, loss summation and instantaneous growth rate coefficient), the size-frequency me thod, and eight indirect methods (Banse and Mosher, Benke, Brey, Edgar, Mor in and Bourassa, Plante and Downing, Robertson and Schwinghamer et al.) to measure or to predict the secondary production of a Mediterranean populatio n of the polychaete Ditrupa arietina. This comparison was carried out both on the 1994 cohort during its two-year lifespan and on the entire populatio n during the two years of the study (1994-1996). Our results showed that al l three direct methods are more or less equivalent. Results referring to th e year 1994-1995 showed a strong overestimation of the actual production by the size frequency method. This is attributed to the existence of an avera ge individual maximal size below which the actual production equals zero. D uring the same year, the results obtained using indirect methods were highl y variable, partly, but not solely, due to the type of environment for whic h the regression models were built. The incorporation of temperature as an independent variable within the most recent regression models did not contr ibute to improving the quality of the predictions. Results referring to the year 1995-1996 confirm the incapacity of both the size frequency and the i ndirect methods to predict changes in the P/B ratio in relation with age st ructure and recruitment irregularities. These results are discussed in the context of inferring productivities both at the population and the communit y level. (C) Elsevier, Paris / Ifremer / Cnrs / Ird.