Signatures of invasion: using an integrative approach to infer the spread of melon fly, Zeugodacus cucurbitae (Diptera: Tephritidae), across Southeast Asia and the West Pacific

Invasion into new areas by already widespread pest organisms often occurs through non-obvious routes, with the origins of such invasions difficult to determine. Understanding population structure using multiple datatypes can help untangle past dispersal events and reveal putative contemporary invasion pathways. The tephritid fruit fly, Zeugodacus cucurbitae (Coquillett), is a serious pest of cucurbits and other commercial crops and is considered native to the Indo-Oriental region, but is invasive in both Africa and the Pacific. Here, we combine molecular (microsatellites and COI) and morphological (male genetalia length and wing shape geometric morphometrics) data within an integrative taxonomic framework to test hypotheses concerning levels of Z. cucurbitae population variation observed in Southeast Asia (native range, 10 sites, ~200 individuals) versus the West Pacific (invasive range, 4 sites, ~80 individuals), and whether single or multiple introductions of Z. cucurbitae have occurred into the West Pacific. We also use this case to explicitly test if using an integrative approach provides more information about hypothesized invasion pathways than either genetic or morphological approaches would do alone. All datasets support Z. cucurbitae as being more variable in Southeast Asia than the West Pacific, and within these regions populations appear to be structured geographically. In particular, mainland and Sundaic Southeast Asian locations formed separate clusters, and New Guinea and Solomon Islands were not closely related to Guam and Hawaii. Evidence supports a separate single origin for New Guinea from the Melanesian arc, the Solomon Islands from Malaysia/Singapore, and Guam from mainland Asia, but multiple introductions into Hawaii from mainland Asia. Taken together, we argue that there is great value in integrating evidence from multiple sources as it can provide finer resolution of population relationships than any single data source alone