A database on host plant records from 437 ingroup taxa has been used t
o test a number of hypotheses on the interaction between butterflies a
nd their host plants using phylogenetic methods (simple character opti
mization, concentrated changes test, and independent contrasts test).
The butterfly phylogeny was assembled from various sources and host pl
ant clades were identified according to Chase et al.'s rbcl-based phyl
ogeny. The ancestral host plant appears to be associated within a high
ly derived rosid clade, including the family Fabaceae. As fossil data
suggest that this clade is older than the butterflies, they must have
colonized already diversified plants. Previous studies also suggest th
at the patterns of association in most insect-plant interactions are m
ore shaped by host shifts, through colonization and specialization, th
an by cospeciation. Consequently, we have focused explicitly on the me
chanisms behind host shifts. Our results confirm, in the light of new
phylogenetic evidence, the pattern reported by Ehrlich and Raven that
related butterflies feed on related plants. We show that host shifts h
ave generally been more common between closely related plants than bet
ween mon distantly related plants. This finding, together with the pos
sibility of a higher tendency of recolonizing ancestral hosts, helps t
o explain the apparent large-scale conservation in the patterns of ass
ociation between insects and their host plants, patterns which at the
same time are more flexible on a more detailed level. Plant growth for
m was an even mon conservative aspect of the interaction between butte
rflies and their host plants than plant phylogeny. However, this is la
rgely explained by a higher probability of colonizations and host shif
ts while feeding on trees than on other growth forms.