Recent attention has focused on the harmful effects of introduced biological control agents on nontarget species. The parasitoid Compsilura concinnata is a notable example of such biological control gone wrong. Introduced in 1906 primarily for control of gypsy moth, Lymantria dispar, this tachinid fly now attacks more than 180 species of native Lepidoptera in North America. While it did not prevent outbreaks or spread of gypsy moth, we present reanalyzed historical data and experimental findings suggesting that parasitism by C. concinnata is the cause of the enigmatic near‐extirpation of another of North America's most successful invaders, the browntail moth (Euproctis chrysorrhoea). From a range of 160 000 km² a century ago, browntail moth (BTM) populations currently exist only in two spatially restricted coastal enclaves, where they have persisted for decades. We experimentally established BTM populations within this area and found that they were largely free of mortality caused by C. concinnata. Experimental populations of BTM at inland sites outside of the currently occupied coastal enclaves were decimated by C. concinnata, a result consistent with our reanalysis of historical data on C. concinnata parasitism of the browntail moth. The role of C. concinnata in the disappearance of browntail moth outside these enclaves has not been reported before. Despite the beneficial role played by C. concinnata in reversing the browntail moth invasion, we do not advocate introduction of generalist biological control agents. Our findings illustrate that the impact of such organisms can be both unpredictable and far‐reaching.

We used pheromone-baited traps to survey the distribution of winter moth, Operophtera brumata (L.) (Lepidoptera: Geometridae), a new invasive defoliator from Europe in eastern New England. The traps also attracted Bruce spanworm, Operophtera bruceata (Hulst) (Lepidoptera: Geometridae), native to North America. We distinguished between the two species by examining male genitalia and sequencing the mitochondrial cytochrome oxidase subunit 1 (COI) gene, the DNA barcoding region. In 2005, we recovered winter moths at sites stretching from eastern Long Island, southeastern Connecticut, all of Rhode Island, eastern Massachusetts, coastal New Hampshire, and southern coastal Maine. At sites further west and north we captured only Bruce spanworm. In 2006, we confirmed that both winter moth and Bruce spanworm are present in Nova Scotia and in coastal Maine, but only Bruce spanworm was recovered in coastal New Brunswick, Canada; Pennsylvania; Vermont; or Quebec City, Canada. In 2007, we collected Bruce spanworm, but no winter moths, in New Brunswick and the interior areas of Maine, New Hampshire, and New York. Winter moth and Brace spanworm differed in the COI sequence by 7.45% of their nucleotides. The prevalence of intermediate genitalia in the zone of overlap suggested that hybridization between the two species may be occurring. To confirm the presence of hybrids, we sequenced the nuclear gene, glucose-6phosphate dehydrogenase (G6PD). We identified six nucleotides that routinely distinguished winter moth and Bruce spanworm, of which three were always diagnostic. We showed that eggs produced by hybridizing the two species in the laboratory contained copies of both species at these six sites. We found that most of the moths collected in the field with intermediate genitalia had winter moth CO1 and G6PD sequences and thus were not hybrids (or at least F1 hybrids). We found three hybrids out of 158 moths with intermediate genitalia in the region where both species were caught. We conclude that hybrids occur in nature, but are not as common as previously reported. Introgression of genes between the two species may still be significant.

The European winter moth, Operophtera brumata (L.), an invasive forest defoliator, is undergoing a rapid range expansion in northeastern North America. The source of this invasion, and phylogeographic diversity throughout its native range, has not been explored. To do this, we used samples from a pheromone-baited trap survey of O. brumata collected across its native range in Europe, and invasive range in North America. Traps in North America also attract a congeneric species, the Bruce spanworm O. bruceata (Hulst), and the western Bruce spanworm O. b. occidentalis (Hulst). From this sampling, we sequenced two regions of the cytochrome c oxidase subunit I mitochondrial gene; one region corresponds to the DNA ‘barcode’ region, the other is a nonoverlapping section. We used these sequences, in combination with sequence data from a recent survey of the Geometridae in western North America, for phylogenetic and phylogeographic analyses to characterize genetic divergence and variation for O. brumata in North America and Europe, and O. bruceata and O. b. occidentalis in North America. We found O. brumata mtDNA diversity to be dominated by a single widespread, and common haplotype. In contrast, O. bruceata shows high haplotype diversity that is evenly distributed throughout North America. Phylogeographic patterns indicate an introduction of O. brumata in British Columbia likely originated from Germany, and suggest the invasive population in northeastern North America may have its origins in the United Kingdom, and/or Germany. We found uncorrected pairwise sequence divergence between Operophtera species to be ≈7%. O. b. occidentalis is ≈ 5% divergent from O. bruceata, has a restricted range in the Pacific Northwest, and has unique morphological characters. Together these lines of evidence suggest O. b. occidentalis may be deserving of species status. Additionally, a single morphologically unique Operophtera specimen, similar to O. bruceata, was collected in southern Arizona, far outside the known range of O. bruceata. This suggests that North America may contain further, unsampled, Operophtera diversity.