Zusammenfassung: The software “BHATTA”, achieved in Turbo Pascal on IBM PC makes possible histogram analysis for population structure. The base of the study is an analysis of population dynamics from histograms of size frequency. The program “BHATTA”, made from the method of logarithmic differences of Bhattacharya, cuts the histograms in order to display Gaussian components and to estimate the parameters: number (size of sample), mean and standard deviation for each generation or cohort to determine growth curves for each cohort.

Zusammenfassung: Biases in estimating population sizes of the plus group are examined for bluefin tuna (Thunnus ) and swordfish (Xiphias gladius ). It is shown that the VPA estimates used in the 1990 ICCAT assessment are sensitive to the systematic ageing errors. The biases in the estimates of population sizes of the plus group might increase exponentially with tracing back to the earlier year.

Zusammenfassung: Length-based population analyses have had a fairly long but not too successful history. These methods are required over a range of situations from when aging is impossible to when it is difficult or prohibitively expensive. There have been 3 sorts of approaches: 1) modal analysis to convert length frequencies to ages 2) age-length key enhancement and 3) imbedded growth models. The last group is what is generally known as length-based methods and the focus of this review. At the 1991 ICES Methods Working Group, 5 techniques were compared and these results are summarized. Two of these techniques, cohort slicing and length-based ADAPT, are applied to swordfish (Xiphias gladius ) data. The imbedded growth model for the former technique is the mean size at age while the latter uses the distribution of size at age. A new method is also presented which begins with cohort slicing then estimates numbers at age and enhances an inferred age-length key at each step of an iterative process.