Evidence of Host Discrimination
Attempts to detect host discrimination among herbivorous insects
have focused on species that are endoparasitic (feed within plant tissues)
and infest small, discrete hosts, such as seeds, buds, flowers and fruits
(Prokopy, 1972). Although host discrimination may be weak among
external, leaf-feeding insects (Groeterset al., 1992; Mappes and Mäkelä,
1993), these species have also received less attention (Schoonhoven,
1990; Poirier and Borden, 1991). Perhaps more relevant than the insect’s
feeding mode is the relative mobility of the insect larva (whether it can
move to new hosts when resources become scarce) and the number of
larvae that can be supported by one host individual (Vasconcellos-Neto
and Monteiro, 1993). Benefits to host discrimination should also depend
on whether a female can accurately ‘survey’ conspecific density on a
potential host before she deposits her own eggs.
Three complementary lines of evidence have been used to determine
whether seed parasites respond to conspecific eggs or larvae on potential
hosts. The most common assay has been to provide females with occupied
vs. unoccupied hosts in a laboratory choice test. A related technique is to
offer a female a single host in a no-choice situation and to record her
responses as a function of whether or not the host is already infested.
Relevant variables include the probability of host acceptance, the time
elapsed until a seed is accepted and the number of eggs laid. At the
population level, host discrimination can be detected by exposing hosts to
a single female or a group of females, and then measuring the dispersion
of eggs among hosts. Whereas most natural insect populations show
clumped or aggregated distributions, strong host discrimination should
produce a uniform or regular distribution of eggs among host individuals.
Reliance on only one of these techniques may produce misleading
results in respect of the presence or strength of host discrimination.
Subtle, statistical preferences obtained in laboratory choice tests may
not translate into effective host discrimination under natural conditions,
where differences in intrinsic host quality or other factors may over-
whelm any effects of previous infestation on host acceptance (Klijnstra
and Schoonhoven, 1987). An even dispersion of juvenile stages in the
field might indicate cannibalism or strong contest competition within
seeds rather than host discrimination by ovipositing females. Conversely,
a failure to observe an even dispersion of juvenile stages does not
preclude host discrimination; periods of limited host availability may
produce an aggregated or random dispersion of eggs in species for which
host discrimination has been well documented (Averill and Prokopy,
1989; Fox and Mousseau, 1995).
Host discrimination has been detected in all major groups of
seed parasites, including weevils (Fergusonet al., 1999a,b), seed flies
(Zimmerman, 1982; Pittara and Katsoyannos, 1990; Lalonde and Roitberg,
1992), cone flies (McClureet al., 1998; Quiringet al., 1998), seed-eatingHost Discrimination by Seed Parasites 67