few eggs and the ‘penalty’ for adding eggs to occupied seeds is much
greater. This is particularly true because older larvae have a competitive
advantage over younger larvae and only one larva will develop to adult
emergence in a small seed (Messina, 1991b). We might therefore predict
parallel differences in the strength of host discrimination among
populations.
Geographical populations ofC. maculatusdo vary in the predicted
direction (Messina and Mitchell, 1989; Mitchell, 1990). Females from an
Asian strain with highly competitive larvae produce much more uniform
egg distributions and show stronger preferences for egg-free seeds than
do females from strains with less competitive larvae (Messina, 1991a;
Messinaet al., 1991). Moreover, Asian females provided few seeds simply
stop laying eggs after each seed has received two or three eggs, so that
their lifetime realized fecundities are sharply reduced. Females from
strains with less competitive larvae add several eggs per seed, and their
realized fecundities do not drop substantially unless seeds are absent
(Messina, 1999). Interpopulation differences in host discrimination
depend both on differences in proximate cues and on differences in
female responses to egg-laden seeds (Messinaet al., 1991). Reciprocal
crosses of populations with divergent egg-laying behaviour suggested
almost complete dominance towards strong host discrimination and very
uniform egg laying (Messina, 1989).
Callosobruchusbeetles have also been used to identify non-genetic
sources of variation in host discrimination. Non-diapausing females
emerge from seeds with about eight mature oocytes and, in the absence
of suitable hosts, they will accumulate oocytes in their abdomens for
about 2 days (Credland and Wright, 1989; Wilson and Hill, 1989). In one
experiment, females from a population with strong host discrimination
were deprived of hosts for 10 h and were compared with females with no
host deprivation. Host deprivation caused the frequency of oviposition
‘mistakes’ to rise from 19% to 50%, where a mistake was defined as
adding a second egg to a seed when at least one uninfested seed was still
available (Messinaet al., 1992). In a strain with weak host discrimination
(and less competitive larvae), a period of host deprivation had little effect
on the frequency of egg-laying mistakes, which were already common
among females with no deprivation.
There has been a long-standing interest in whether experience gained
during a foraging bout modifies host discrimination byC. maculatus
females. Mitchell (1975) proposed that learning was needed for females to
maintain uniform egg distributions even after all seeds bear multiple eggs.
In particular, he hypothesized that oviposition would be more likely if a
currently visited seed has fewer eggs than the previous one. An alterna-
tive hypothesis is that a female simply has a lower probability of host
acceptance with each increase in egg density. In this case, the specific
probabilities of acceptance will depend on the female’s physiological
condition or the amount of time since the last egg was laid (Messina and
Renwick, 1985). Wilson (1988) labelled these respective mechanisms asHost Discrimination by Seed Parasites 75