‘relative’ and ‘absolute’ rules, and performed experiments that suggested
female behaviour conformed to an absolute rule (no learning). However,
simulations by Mitchell (1990) indicated that absolute rules could not
account for females that are especially adept at distinguishing different
egg densities on seeds.
The role of experience inCallosobruchushost discrimination was
examined in detail by Horng (1997), who combined foraging models
with empirical observations of female behaviour and larval competition.
Females from populations with very competitive larvae did appear to use
a relative rule in accepting or rejecting egg-laden seeds, but an absolute
rule was adequate to explain the weaker discrimination by females
from populations with less competitive larvae (Horng, 1997). The beetle–
legume interaction may therefore serve as an example of genetic variation
in the expression of parasite learning. A statistical analysis of behaviour
sequences indicated that female behaviour is best described by an
‘adjusted-threshold’ model, in which thresholds for accepting hosts with
different egg densities are modified by the female’s current egg load and
the amount of time since her last oviposition (Hornget al., 1999).Host Size and Parasite Evolution: a Selection Experiment
Genetic variation in host discrimination by female seed beetles appears
to reflect differences in the intensity and type of larval competition
within seeds. Yet this explanation begs the question as to why larvae
from completely interfertile populations should compete in such different
ways (Messina, 1991a,b). Several empirical and theoretical investigations
of this question have converged on the idea that larval competitiveness
is itself primarily an adaptation to seed size (Smith and Lessells, 1985;
Toquenagaet al., 1994; Tuda, 1998). The Asian strain discussed above
was obtained from a region where most hosts are small and other
local beetle populations exhibit contest-type competition within seeds
(Mitchell, 1991).
Because optimal larval behaviour depends on the frequencies of
possible phenotypes, Smith and Lessells (1985) used game theory to
predict evolutionarily stable levels of aggression among seed-parasite
larvae. They concluded that the cost of scramble or exploitative competi-
tion in a large seed is low enough for ‘tolerant’ behaviour (such as a ten-
dency to form peripheral, non-intersecting burrows) to be advantageous,
whereas an aggressive, ‘contest’ phenotype is superior in a small host (but
see Colegrave, 1995). Both Toquenaga et al. (1994) and Tuda (1998)
obtained data suggesting that transferring a population to a different-sized
host alters larval competitiveness in the expected direction.
I performed a mass-selection experiment (F.J. Messina, unpublished
data) to test directly the effects of host size on both larval competitiveness
and host discrimination inC. maculatus. A response to selection may
be expected in this system because earlier experiments confirmed76 F.J. Messina