Responses to selection in the first gen
eration were attained at relatively high
levels of selection intensity, while re
sponses in the second generation were
achieved at much reduced levels of
selection intensity. Unfortunately, posi
tive gains made in cold tolerance in upselected
populations of O. aureus and
its hybrid during the first generation of
selection were lost during the second
generation of selection. In contrast, rapid
gains in down-selected populations were
achieved in O. niloticus, irrespective of
selection intensity. Such results may in
dicate that cold tolerance is near a se
lection limit in these populations. Lack
of response to up-selection in the sec
ond generation may have resulted from
a combination of factors including: sam
pling error, selection plateau,
misidentification of freeze brands and/
or a relaxing of the intensity of selec
tion. It should be noted that the selec
tion differential in the first generation
was based on the upper and lower 10%
of a base population consisting of 600
individuals. However, selection differ
ential in the second generation of
selection was based on the upper and
lower 50% of 200 individuals/population.
Thus the intensity of selection in the
second generation was significantly less
than that practiced in the first generation.
It should be noted that many of the
causes for lack of response and
asymmetry of response are exacerbated
by small founding populations. In the
US, certain tilapia populations were
founded as early as 1957 with as few
as five broodstock (R.O. Smitherman,
Auburn University, pers. comm.). Also
many of the US populations have
experienced multiple and severe
bottlenecks, further reducing genetic
variance. Despite limited phenotypic
variance among several tested
populations (Behrends etal. 1990a), it