Estimating genetic change for selection programs designed to improve p
erformance of fish populations requires the application of statistical
and biological approaches specific to the conditions used to maintain
the populations. The unique set of limitations is determined by such
factors as number of individuals reared, availability of control popul
ations, variability of environmental conditions, rearing strategy, and
completeness of data collection. All approaches available to the fish
breeder are designed to remove to the extent possible, both fixed and
random environmental effects from estimates of change. Some methods a
lso attempt to account for genetic trends due to previous selection. G
enerally, influences attributable to genetic drift and inbreeding are
confounded with effects of selection. Four approaches to estimating ge
netic change have been developed for use in animal breeding. The most
common is the use of control populations to estimate environmental cha
nge and possibly account for some inbreeding effects related to popula
tion size. The second is to use divergent selection as a means of inte
rnally correcting for environmental changes and adjusting for asymmetr
y of response to selection. Contemporary comparisons is a common metho
d with species for which gametes can be stored. The method is based on
producing progeny from parents of early and late generation progeny i
n a single season to provide a direct comparison of performance of ind
ividuals from two or more generations. The most recently developed app
roach involves the use of animal models, based on best linear unbiased
prediction procedures, to estimate directly the average breeding valu
es for sequential generations. Information for all individuals in a pe
digreed population is utilized to provide simultaneous adjustments for
environmental effects and genetic trends.