size of the population (Ne) is large.
Unfortunately, ‘large’ is difficult to define. In
the context of conservation of threatened
species, Franklin (1980) suggested Ne= 500
as an acceptable minimum, whereas Lande
(1995) pointed out that a target of Ne= 5000
is more appropriate. Furthermore, in any
given generation, relationship between effec-
tive population size (Ne) and the number of
adults (N) is a complex function; however, in
the absence of exceptional circumstances, it
will tend to be in the range 0.25 Ne/N
0.75 (Nunney, 2000). From these figures it is
clear that a reasonable goal is to found and
maintain the population with N 1000
(Pimentel, 1990). Practical constraints may
preclude such a large founding population,
but there is no question that several samples
of N100 unrelated adults are necessary to
reflect the genetic variation of the source
population (Mackauer, 1976; Bartlett, 1994).
The size of a founding population is not
the only parameter relevant to initiating a
captive-rearing programme. It must also be
decided which natural populations will be
sampled. To maximize field adaptation, a
source population should (if possible) be
from a region that is climatically similar to
the release sites (McDonald, 1976). But how
many source populations should be used?
Using more than one source population has
a large potential advantage of increasing
genetic variability. However, mixing individ-
uals from different geographical locations
can lead to the breakdown of geographically
distinct coadapted gene complexes. Such
breakdown results in a general loss of fitness
(e.g. Burton et al., 1999) and can cause an
unpredictable change in some traits (see
Carson and Templeton, 1984). Male mating
traits have been shown to exhibit geographi-
cal genetic coadaptation (e.g. Aspi, 2000) and
there could be a major problem if, in an SIT
programme, there is a change in mating
behaviour. We have no way of predicting
when coadaptation is likely to be a problem.
The best indicator of a potential problem is a
large genetic distance between the popula-
tions. Since genetic distance is not necessar-
ily well correlated with geographical
distance (see Burton et al., 1999), it is prudent
to rear samples from different populations
independently until genetic testing or other
evaluation can be carried out.
Once the founding population has been
introduced into the rearing facility, it is very
important to avoid the ‘crash’ of the
‘crash–recovery’ cycle often seen in the ini-
tial stages of captive rearing (Leppla et al.,
1983). The crash occurs because the found-
ing population is generally maladapted to
the rearing environment. As a result, most
genotypes fail to reproduce, but a few are
successful. For example, Leppla et al.(1983)
found that, during the establishment of a
new medfly colony, fewer than half of the
females were reproducing over the first ten
generations.
The exclusive success of a few genotypes
dramatically reduces Neand creates a real
danger of extremely rapid inbreeding.
Temporarily dividing the founding sample
into a large number of very small breeding
units can minimize the problem. It may be
necessary to expend significant effort to
ensure the reproductive success of as many
of these units as possible. The survival of
many independent subpopulations ensures
the reproductive success of a large number
of genotypes and avoids large-scale genetic
losses.Colony maintenanceBartlett (1994) lists some of the important
variables that generally differ between the
environment of natural and captive-reared
populations. The most likely cause of declin-
ing quality is the inability of individuals
adapted to the captive-rearing conditions to
function under field conditions. Notably, in
the natural environment, the physical para-
meters (e.g. temperature) are variable and
the availability of resources (e.g. hosts
and/or mates) is generally low. There are
two qualitatively different ways of dealing
with this problem.
First, stocks can be maintained as a large
number of inbred (isofemale) lines (Roush
and Hopper, 1995). This is the best solution
for preventing adaptation to rearing condi-
tions, since inbred lines cannot adapt, and is
the optimal strategy when quality is rapidly80 L. Nunney