Historical biogeography plays an important role in understanding evolu
tionary processes and the history of life, with fossil data, plate tec
tonics, and palaeoclimatology offering major data bases for biogeograp
hic analyses. Here we suggest that specialized interspecific interacti
ons, in combination with molecular data, can play an important role in
such analyses. We use the interaction between Bowlesia incana (Apiace
ae) and the host-specific herbivore Greya powelli (Lepidoptera: Prodox
idae) in California to demonstrate the utility. Distributed in disjunc
t temperate parts of North and South America, B. incana has been propo
sed to have been introduced in historical time (<250 yrs BP) into Nort
h America. Three lines of evidence together suggest that the plant is
of a far older age in North America. First, G. powelli is not known fr
om the South American range and the genus is very unlikely to exist th
ere, making introduction with the plant in North America unlikely. Sec
ond, divergence of mtDNA among members of the genus Greya suggests tha
t the lineage leading to G. powelli originated 2.3-3.8 Mya, thus preda
ting a proposed introduction by several orders of magnitude. Third, ho
st shifts are consistently linked with species divergence within the g
enus Greya, suggesting that the G. powelli lineage has utilized Bowles
ia since a time near its origin. We conclude that B. incana has been p
resent in western North America for a long period of time, and that it
did not arrive by human transport. The use of specialized herbivores
and molecular data adds a powerful tool to historical plant biogeograp
hy. (C) 1998 The Linnean Society of London.