sitoid species was that for Itoplectis conquisi-
tor(Yazgan, 1972). Diets whose entire chem-
ical composition is known, even if the
structure of some components is not fully
defined (nucleic acids, proteins), can be con-
sidered as chemically defined. A small num-
ber of diets that fit such a definition were
tested successfully for rearing ento-
mophagous insects (Grenier et al., 1994). In
such diets, many complex or ‘crude’ compo-
nents can be added as host substitutes.
Irrespective of the species reared, whether
parasitoids or predators, the most com-
monly used components are hen’s egg yolk,
chicken embryo extract, calf fetal serum,
bovine serum, cow’s milk, yeast extract or
hydrolysate, crude proteins or protein
hydrolysates, meat or liver extracts and
seed oils. For recent reviews of such diets,
see Thompson (1999) and Thompson and
Hagen (1999).
Success in Development of Some Species
in Artificial Conditions
The main successes in artificial mass rearing
have been obtained with hymenopterous
egg and pupal parasitoids, tachinid larval
parasitoids and some polyphagous preda-
tors. Extensive general reviews of artificial
diets for entomophagous arthropods have
been published by Grenier et al. (1994),
Thompson (1999) and Thompson and
Hagen (1999).
Koinobiontic endoparasitic Hymenoptera
(parasitoids that do not immediately kill their
hosts and where the parasitoid larvae
develop in the still living host) are the most
difficult species to rear in vitrobecause the
parasitoid has a close relationship with its
living host, which probably supplies the par-
asitoid with some specific growth factors nec-
essary for normal development of the
parasitoid larva (Greany et al., 1989).
Moreover, endoparasitoids, for which the
diet is not only their food but also their envi-
ronment for larval development, have special
requirements compared with ectoparasitoids
or predators. Thus, special attention has to be
paid to factors such as osmotic pressure and
pH (Grenier et al., 1994).
Comparison of Artificially and Naturally
Reared Natural Enemies
Many parameters used as quality criteria are
linked, such as adult body weight and
longevity, fecundity, flight activity and
searching ability (Kazmer and Luck, 1995).
Quality control procedures could be simpli-
fied and could thus be made less costly if we
were able to use one parameter that is easily
measured (e.g. size) to predict the value of
another trait that is more complex or time-
consuming to determine (e.g. fecundity or
field performance). In parasitoids, body size
may be related to fecundity, longevity, rate of
search and flight ability (Kazmer and Luck,
1995). Bigler (1994) pointed out that the
female body size of a parasitoid could be
used as an index of fitness or a quality para-
meter, as in Trichogramma. But female size is
not a reliable parameter for predicting field
performance when the parasitoids are reared
on factitious or artificial hosts. In
Trichogramma, large-sized wasps developed
from in vitrorearing showed characteristic
abnormalities called ‘big belly’. Despite their
large body size, such adults usually have a
low viability. The size of a normally shaped
Trichogrammaadult produced in vitrois also
larger than that of a wasp that developed in
the natural host (Nordlund et al., 1997). This
is often found in oophagous parasitoids and
is the result of a low number of parasitoid
eggs developing in the large amount of food
that is available to them (Grenier et al., 1995).
In general, the size of Trichogrammaand
other oophagous parasitoids varies accord-
ing to the number of adults developing in
the same host, which consume all the avail-
able host material. Remains of the host pre-
vent proper pupation of parasitoids, and
parasitoid larvae that are excessively large
cannot pupate. In a natural situation with a
great many Trichogrammalarvae in one host,
adult parasitoid size will be reduced accord-
ingly. Under artificial rearing conditions,
however, the quantity of food in the artificial
host egg is usually very large compared with
a natural host egg, and the number of para-
sitoid eggs laid is often too low for the devel-
opment of normal-sized Trichogramma
(Grenier et al., 2001).118 S. Grenier and P. De Clercq