Why did specificity testing fail to predict the field host-range of the gorse pod moth in New Zealand?

Contrary to predictions based on host-range testing, the gorse pod moth (GPM) infests pods of several exotic Genisteae and Loteae species, as well as the target weed gorse Ulex europaeus, throughout New Zealand. The original host-range tests were conducted on moths collected in southern England; however, the offspring of Portuguese moths were also released in New Zealand. We investigated whether failure to predict non-target attack was because (a) a cryptic species was accidentally introduced; (b) asynchrony between the oviposition period of GPM and gorse flowering results in deprivation, causing less preferred plants to become more acceptable for oviposition and (c) the Portuguese GPM population has a different host-range to the tested English population. Dissections of genitalia and molecular data collected on COI mtDNA indicated that a cryptic species was not introduced. Specificity tests on moths sourced from England concurred with the original tests and indicated that GPM should be unlikely to exploit the non-target species that are attacked in New Zealand. In contrast, GPM sourced from Portugal were able to exploit a broader range of plants, although choice oviposition tests indicated that gorse is, nevertheless, the preferred host of this population. Adult GPM activity was often poorly synchronized with gorse flowering in New Zealand and non-target attack was most prevalent when gorse flowers and pods were absent. We conclude that the release of untested moths sourced from Portugal, coupled with asynchrony between the flight period of GPM and gorse flowering explains the unanticipated non-target attack in New Zealand. (C) 2008 Elsevier Inc. All rights reserved.