kuehniella eggs (see Barrett and Schmidt,
1991), T. brassicaeproduced on this host were
smaller and less vigorous. Thus, the smaller
wasps were likely to have had more diffi-
culty parasitizing O. nubilalisegg masses.
Secondly, the increasing rejection rate of O.
nubilalis eggs by T. brassicae also had a
genetic basis. They pointed out that the rejec-
tion rate of O. nubilaliseggs increased with
the number of generations that T. brassicae
had been reared on A. kuehniella eggs.
Presumably, the size of the wasps in the first
generation that emerged from A. kuehniella
eggs was similar to that of wasps emerging
after 12 generations on this host. If this was
indeed the case, then wasp size per secould
be eliminated as the only explanation for the
decrease in acceptance of O. nubilaliseggs.
This, coupled with the failure of some para-
sitized hosts to issue wasp offspring, sug-
gested that both phenotypic (i.e.
deterioration) and genetic changes (i.e. a fail-
ure to recognize and develop in O. nubilalis
eggs) were responsible for the change in
acceptance and suitability of O. nubilaliseggs
to T. brassicaereared on A. kuehniellaeggs.
Summary
While trial and error, serendipity, experience
and economics play major roles in develop-
ing an augmentative biological control pro-
gramme, laboratory and field experiments
must still form the foundation for such pro-
grammes (e.g. Taylor and Stern, 1971; Parker
and Pinnell, 1974; van Dijken et al.,1986).
They provide the framework in which to
evaluate the likely prospects for a pro-
gramme’s success or for modifying elements
in a programme that would enhance its suc-
cess. Traditionally, augmentative biological
control has been viewed as simply a matter
of releasing sufficient numbers of a biologi-
cal control agent, such as Trichogramma, to
ensure that a high proportion of the potential
host population is encountered. This
assumes that the only factor determining the
success of an augmentative biological control
programme is the rate at which hosts are
encountered. Growing evidence suggests
otherwise. The degree of biological control
that can be expected is much more likely to
involve the particular attributes manifested
by these hosts in terms of the quality of the
parasitoid offspring arising from them. The
particular host attributes assessed by a para-
sitoid as part of her ‘decision’ to allocate off-
spring are specific to and vary with the host
species, parasitoid species, parasitoid state,
host condition and circumstance. Thus, host
attributes such as egg size, stage of a devel-
oping embryo within the host egg, chorion
hardness or multilayered egg masses, i.e.
those attributes associated with host accep-
tance by Trichogrammaspecies, are but exam-
ples. Each parasitoid–host interaction is
likely to differ and each interaction needs to
be assessed.
The experimental approach advocated
here involves behavioural observations, but
observations that are tied to the size and sex
of the resulting offspring and the degree to
which the parental female is willing to
exploit a particular type of host resource.
This latter aspect measures the reticence
manifested by a parental female in exploiting
a host resource in terms of the proportion of
her egg load she is willing to lay in that host
type (e.g. species, size, age, stage of develop-
ment, etc.). A parasitoid will often exploit a
few marginal hosts if she has had little suc-
cess in encountering hosts, but such hosts
will seldom be heavily exploited. Thus, the
reticence in host use allows host preferences
to be assessed and ranked. This then deter-
mines the attributes of those hosts that are
most likely to be exploited in the field, i.e.
whether augmentative biological control is
likely to succeed. This ranking of host charac-
teristics is also the basis for assessing the sta-
tus of the host (pest) population in the field.
The behavioural approach we advocate is
rooted in a parental female’s reproductive
success – that is, in terms of her offspring’s
reproductive prospects (= a parental female’s
reproductive success). Given a parental
female with a particular set of behavioural
attributes (e.g. how readily she accepts hosts
of a given size or how willing she is to invest
time in exploiting a specific host stage), her
reproductive success is defined in terms of
the success that her offspring manifest in
reproducing when compared with the242 R.F. Luck and L.D. Forster