Traditional calculation of catch age composition estimate requires that the age-length data and random length frecuency be sampled from the same population, in other case an age-length key is capable of providing biased estimates of age distribution. By using the iteration algorithm is possible to overcome this limitation.
The catch age composition for Dissostichus eleginoides was obteined with the aplication of a single age-length key to all monthly length distribution using the nonlinear model (Hoenig and Heisey, 1987). We found a catch age distribution pattern wich is not significatly different from the one obteined by use of traditional method