lines could have partially reverted towards values in the ancestral line on
mung beans.
All replicate cowpea-selected lines evolved to resemble populations
that have long been associated with large hosts. Larval competitiveness
decreased considerably. Mung beans receiving two, equal-aged larvae
from cowpea-selected lines frequently yielded two emerging adults
(> 60% of all seeds), but lines maintained on mung bean almost invariably
yielded only one adult per seed (F.J. Messina, unpublished data). When
the host was cowpea, approximately 90% of seeds with two cowpea-
selected larvae yielded two adults, but only 13% of seeds with two mung
bean-selected larvae did so. There was also a major effect of selection on
the body size of emerging adults. Even in the absence of competition (only
one larva per seed), adults from cowpea-selected lines were significantly
smaller than those from either mung bean-selected lines or the ancestral
population.
Cowpea-selected and mung bean-selected lines diverged in their
degree of host discrimination. Egg distributions were measured for
each line on each host in 12 treatment combinations. Statistical analysis
indicated an effect of test host, as egg-laying females from all lines were
better able to avoid occupied seeds (and produced higherUscores) on
mung bean than on cowpea (Fig. 4.4). This result matches earlier studies
showing more oviposition ‘mistakes’ on larger hosts (Mitchell, 1990).
However, there was also a significant effect of selection host, as well as
a selection-host×test-host interaction. When females were given mung78 F.J. Messina
Fig. 4.4. Egg dispersions byCallosobruchus maculatusfemales derived from
cowpea-selected (C1–3) or mung bean-selected (M1–3) lines and presented either
cowpeas or mung beans.Uscores vary between 0 (= random egg laying) and 1
(= completely uniform distributions) (see Messina and Mitchell, 1989).