Kiore (Rattus exulans) distribution and relative abundance on a small highly modified island

ABSTRACT

Kiore (Pacific rat; Rattus exulans) is both a target for eradication and a taonga or highly valued species in New Zealand, and its abundance and distribution vary considerably throughout the country. We investigated reports of an abundant kiore population on Slipper Island (Whakahau), off the east coast of New Zealand’s North Island, in March 2017. We trapped kiore to examine their distribution across a range of habitats with varying degrees of human activity. Kiore were captured in all habitats, with particularly high abundance at a campground with a fruiting fig tree (50 kiore per 100 trap nights corrected for sprung traps). We found no evidence of other rat species; Slipper Island appears to remain one of few New Zealand islands with kiore but without ship rats (Rattus rattus) and Norway rats (R. norvegicus), the two other rat species present in New Zealand. Slipper Island potentially provides opportunities to research kiore behaviour and population dynamics in a New Zealand commensal environment, and genetics of an isolated island population.     

The effects of mice on stoats in southern beech forests

Introduced stoats (Mustela erminea) are important invasive predators in southern beech (Nothofagus sp.) forests in New Zealand. In these forests, one of their primary prey species – introduced house mice (Mus musculus), fluctuate dramatically between years, driven by the irregular heavy seed‐fall (masting) of the beech trees. We examined the effects of mice on stoats in this system by comparing the weights, age structure and population densities of stoats caught on two large islands in Fiordland, New Zealand – one that has mice (Resolution Island) and one that does not (Secretary Island). On Resolution Island, the stoat population showed a history of recruitment spikes and troughs linked to beech masting, whereas the Secretary Island population had more constant recruitment, indicating that rodents are probably the primary cause for the ‘boom and bust’ population cycle of stoats in beech forests. Resolutions Island stoats were 10% heavier on average than Secretary Island stoats, supporting the hypothesis that the availability of larger prey (mice verses wētā) leads to larger stoats. Beech masting years on this island were also correlated with a higher weight for stoats born in the year of the masting event. The detailed demographic information on the stoat populations of these two islands supports previously suggested interactions among mice, stoats and beech masting. These interactions may have important consequences for the endemic species that interact with fluctuating populations of mice and stoats.     

Eradication of Norway Rats for Recovery of Seabird Habitat on Langara Island, British Columbia

Introduced Rattus norvegicus (Norway rats) caused the decline of Synthliboramphus antiquus (ancient murrelets) and other seabirds breeding on Langara Island (approximately 3,100 ha), British Columbia. Using funds from the litigation settlement following the Nestucca oil spill, Environment Canada eradicated Norway rats using a technique developed in New Zealand which involved dispensing wax baits containing the anticoagulant brodifacoum at 50 ppm from fixed bait stations. Bait stations were placed every 75 to 100 m on a grid over the entire island (1 station/ha). Rats removed bait for 26 days, after which crews placed baits in protective plastic bags in each bait station. Stations loaded with baits were left on the island and rechecked four times over 2 years, after which bait stations and remaining bait were removed. The eradication succeeded. No signs of rats have been detected on Langara Island and its associated islands since January 1996. No rats were trapped during 1,700 trap‐nights following the poison campaign. Incisor marks of rats were not found on apples or oil‐dipped chew‐sticks. Corvus corax (common ravens) likely suffered greater than 50% mortality from the eradication after apparently gaining access to the poison directly from bait stations and from scavenging rat carcasses. A monitoring and response system is being developed in conjunction with current users of the islands. The success on Langara Island demonstrates how the technique proven on small New Zealand islands of less than 300 ha can be effectively extrapolated to much larger islands.     

From small Maria to massive Campbell: Forty years of rat eradications from New Zealand islands

Abstract

Over the last four decades the eradication of rats from islands around New Zealand has moved from accidental eradication following the exploratory use of baits for rat control to carefully planned complex eradications of rats and cats (Felis catus) on large islands. Introduced rodents have now been eradicated from more than 90 islands. Of these successful campaigns, those on Breaksea Island, the Mercury Islands, Kapiti Island, and Tuhua Island are used here as case studies because they represent milestones for techniques used or results achieved. Successful methods used on islands range from bait stations and silos serviced on foot to aerial spread by helicopters using satellite navigation systems. The development of these methods has benefited from adaptive management. By applying lessons learned from previous operations the size, complexity, and cost effectiveness of the campaigns has gradually increased. The islands now permanently cleared of introduced rodents are being used for restoration of island‐seabird systems and recovery of threatened species such as large flightless invertebrates, lizards, tuatara, forest birds, and some species of plants. The most ambitious campaigns have been on remote subantarctic Campbell Island (11 300 ha) and warm temperate Raoul Island (2938 ha), aimed to provide long‐term benefits for endemic plant and animal species including land and seabirds. Other islands that could benefit from rat removal are close inshore and within the natural dispersal range of rats and stoats (Mustela erminea). Priorities for future development therefore include more effective methods for detecting rodent invasions, especially ship rats (Rattus rattus) and mice (Mus musculus), broader community involvement in invasion prevention, and improved understanding of reinvasion risk management.     

Factors affecting <Emphasis Type="Italic">Sarcocystis</Emphasis> infection of rats on small tropical islands

The purpose of our research was to explore the limits of rat-python-Sarcocystis distribution among rats on the offshore tropical islands of Singapore, and to examine the effect of island size, insular isolation, landscape peculiarities and anthropogenic disturbance. Commensal rats (Rattus rattus) inhabited all of these islands, regardless of the islands size, proximity to the mainland, biogeographic features, and/or degree of anthropogenic interference. Rats caught on Sakijang Pelepah Island had well deliminated white bellies that are similar to those of sylvatic or feral rats. The prevalence of Sarcocystis spp. on individual islands ranged from 57 to 100%. This is consistent with the range found in forested habitats on Singapore Island. It also exhibited a similar diversity to Sarcocystis spp. and the predominance of Sarcocystis singaporensis. On Sakijang Pelepah Island, one rat (white bellied) was exceptionally heavily infected with both Sarcocystis villivilosi and Sarcocystis sulawesiensis. The muscles of the rats from nearly all of the islands contained immature sarcocysts, which implies that active transmission is taking place on these islands. This suggests that reticulated pythons (Python reticulatus), which are the definitive hosts of rat Sarcocystis, might have been established or frequented in all the islands of Singapore. Our study shows that the Sarcocystis infection load of the rats was negatively correlated with human disturbance, hinting that human disturbance restricts the pythons mobility and thus, reduces infection of Sarcocytis in the islands rats.     

Modelling the distribution and interaction of introduced rodents on New Zealand offshore islands

Aim To establish the factors that correlate with the distribution of the four most commonly introduced rodent species on New Zealand offshore islands — ship rat (Rattus rattus), Norway rat (R. norvegicus), Pacific rat or kiore (R. exulans) and house mouse (Mus musculus) — and examine if these distributions are interactive at the archipelago scale. Location The 297 offshore islands of the New Zealand archipelago (latitude: 34° S to 47° S; longitude: 166° E to 179° E). Methods Data on the distribution of all four introduced rodent species and the characteristics of New Zealand offshore islands were collated from published surveys and maps. The distribution of individual rodent species was regressed on island characteristics using a logistic generalized linear model. Interactions were examined by including the distributions of other rodent species as predictors in models. Results All four rodent species appear to be limited by a variety of factors, which differ between species in both number and type. The distribution of ship rats is limited by the most factors, reflecting the extent of its distribution across the archipelago. The distribution of mice is the least explicative. Only the three rat species interacted in their distribution. The distribution of kiore on offshore islands is significantly negatively related to that of ship rats and to a lesser extent Norway rats. The distribution of mice did not appear affected in any way by the number of other rodent species on an island. Main conclusions Differences in competitive ability and dispersal allow all four species to inhabit the New Zealand archipelago. Kiore distribution appears to be most limited by ship rat (and to a lesser extent Norway rat) distribution. The distribution of kiore was not found to interact with the distribution of mice on offshore islands, as has been suggested by others. The distribution of mice on offshore islands was difficult to model, which highlights the difficulties in managing this species. Overall the results offer valuable insights for management methods to assist preventing the invasion of offshore islands.     

Competition, coexistence, and adaptation amongst rodent invaders to Pacific and New Zealand islands

Variation in skull size was investigated for three species of rats (kiore –Rattus exulans Peale; ship rat –R. rattus L.; Norway rat –R. norvegicus Berkenhout) which were introduced by humans to various islands in New Zealand and other Pacific islands. Data from seventy-one islands and 882 specimens are examined for evidence of the effects of latitude, island size and interspecific competition among rats and the house mouse (Mus musculus L.) on skull size, using multiple regressions. For R. exulans, skull size increases with latitude as predicted by Bergmann's rule, but no such effect occurs for the other two rats. There was a positive relationship between island size and the number of species inhabiting it, and some species combinations were more likely to occur than others. For example, R. exulans and R. norvegicus were more likely to occur together, while R. rattus and R. exulans were rarely sympatric. R. exulans and R. rattus skull size was negatively correlated with the number of other rodents on the same island. R. exulans skull size increased on smaller islands in some island groups, perhaps because increased density and consequent increased intraspecific competition on smaller islands favours increased body size. This effect is more pronounced in tropical islands (Solomon islands), than in subtropical ones (Hawaiian islands) and less so in temperate New Zealand. Collectively the data demonstrate that rapid evolution of body size in predictable directions can follow within 150 years of the introduction of species to new receiving communities.     

A pleasing consequence of Norway rat eradication: two shrew species recover

Four to 10 years after the successful eradication of the Norway rat (Rattus norvegicus) from three islands of the Sept–Îles Archipelago and one in the Molène Archipelago (Brittany, France), the abundance index of the lesser white‐toothed shrew (Crocidura suaveolens) increased by factors of 7–25, depending on the island and the year. Moreover, in the same region, the abundance index of the greater white‐toothed shrew (Crocidura russula) on Tomé Island increased by factors of 9 and 17, one and two years after the Norway rat eradication, respectively. The maximum variation of the abundance index for the lesser white‐toothed shrew during seven years on the rat‐free island of Béniguet in the same region was a factor of only 2.5. Moreover, the distribution of the lesser white‐toothed shrew on Bono island, restricted before the eradication to two steep areas with few rats, increased and encompassed virtually the entire island four years after rats disappeared. These results suggest strong detrimental interactions between the introduced Norway rat and the two Crocidura shrew species on temperate oceanic islands. However, our data do not indicate the ecological mechanisms at work in these interactions. The main reason this shrew recovery was detected after rat eradication was the inclusion in the eradication protocol of the evaluation of impacts on the local biota of eliminating alien species. The rigor of the sampling procedure was also crucial to this discovery. This example demonstrates that an eradication operation can be extremely useful for both scientists and managers if it is planned as a research project.     

Heavy rimu (Dacrydium cupressinum) mast seeding and rat (Rattus spp.) population eruptions on Stewart Island/Rakiura

Abstract

This study aimed to quantify changes in rat abundance and population structure before, during, and after a rimu (Dacrydium cupressinum) mast seed event in lowland forest on Stewart Island, New Zealand. Rats, primarily ship rats (Rattus rattus), were trapped in low numbers throughout the study period (March 2000‐March 2003), except when they erupted to very high abundance in spring 2002, shortly after heavy rimu seed fall. In the immediate post‐peak phase, scavenging of trapped rats increased substantially; rats were seen and trapped in daylight; and weights of adult female rats were low in relation to their size, which suggests that food shortage was the cause of the subsequent steep decline in abundance. Rat eruptions have been observed on Stewart Island after heavy rimu seed fall several times over the past 40 years. Eruptions of rats caused by heavy rimu seed‐fall may have triggered the invasion of nearby islands by rats, and caused the extinction of several native species on Stewart Island.     

Home ranges of introduced mammalian carnivores at Trounson Kauri Park,Northland, New Zealand

Abstract

At Trounson Kauri Park, we monitored the movements of 21 feral cats (Felis catus), 11 stoats (Mustela erminea) and one male ferret (Mustelafuro). In feral cats, the average minimum home range was 446 ha (±82 SE) for 14 males, significantly larger than the average minimum of 117 ha (±40 SE) for seven females. In stoats the average minimum home range was 107 ha (±20 SE), for nine males compared with 81 ha (±31 SE) for two females. The single male ferret had a minimum home range of 197 ha. Adult male feral cats lived on apparently separate, non‐overlapping home ranges; females occupied exclusive home ranges which were overlapped by adult males; home ranges of sub‐adult male feral cats overlapped those of other sub‐adult male, adult male and female feral cats. The home ranges of two neighbouring male stoats overlapped, although their core ranges did not; both these and the ferret's home ranges overlapped those of the neighbouring feral cats. The feral cats were often located in cover in pastureland or near the edge of larger tracts of forest within their home ranges; stoats near waterways in the larger tracts of forest; and the ferret near the forest/pasture margins of Trounson Kauri Park. Our results suggest that control devices targeting all these species should be set at a minimum spacing of 800 m in order to put the majority of the resident and immigrant predators at risk.     

Role of fruits in the diet of small mustelids (<Emphasis Type="Italic">Mustela</Emphasis> sp.) from the western Italian Alps

The food habits of both weasels and stoats (Mustela nivalis and M. erminea) and the occurrence of fruit in their diet were studied by scat analysis in the Gran Paradiso National Park (western Italian Alps). Being impossible to distinguish between stoats and weasels scats, they were grouped together as the genus Mustela. Despite the high consumption of small rodents, which represented the main food item in the diet of small mustelids (41.5% in mean volume), wild fruits (mainly Rosaceae and Ericaceae) were also relatively common (23.3% in mean volume). Fruits were exploited throughout the year, with an intensive use in seasons of high availability. Their consumption did not significantly differ between habitat types, despite a significant decrease of fruit availability from deciduous mixed forest to alpine prairie. This evidence suggests that this trophic behaviour could represent a consolidated feeding strategy of both weasels and stoats, which seem to adopt an opportunistic behaviour as shown by larger size mustelids. We argue that fruits, owing to their good availability and high profitability, could play an important role in the diet of small mustelids in alpine habitats, slackening their well-documented dependence on rodents.     

Comparison of foraging strategies of incubating king penguins Aptenodytes patagonicus from Macquarie and Heard islands

The foraging strategies of king penguins from Heard and Macquarie islands were compared using satellite telemetry, time-depth recorders and diet samples. Trip durations were 16.8±3.6 days and 14.8±4.1 days at Macquarie and Heard islands, respectively. At Macquarie Island, total distances travelled were 1281±203 km compared to 1425±516 km at Heard Island. The total time the penguins spent at sea was 393±66 h at Macquarie Island and 369±108 h at Heard Island. The penguins from Macquarie Island performed more deep dives than those from Heard Island. King penguins from Macquarie Island travelled 1.5±0.2 km h⁻¹ day⁻¹ compared to 1.3±0.1 km h⁻¹ day⁻¹. At Macquarie Island, 19% of dives were upto 70–90 m depth compared to 35% at Heard Island. The main dietary prey species were the fish Krefftychthis anderssoni and the squid Moroteuthis ingens in both groups. The differences in the at-sea distribution and the foraging behaviour of the two groups of penguins were possibly related to differences in oceanography and bathymetric conditions around the two islands. Dietary differences may be due to interannual variability in prey availability since the two colonies were studied during incubation but in different years.     

Distribution and abundance of small mammals in relation to habitat in Pureora Forest Park

Populations of ship rats (Rattus rattus), Norway rats (R. norvegicus), feral house mice (Mus musculus), stoats (Mustela erminea), weasels (M. nivalis), and ferrets (M. furo) were sampled with killtraps every three months from November 1982 to November 1987 in logged and unlogged native forest and in exotic plantations of various ages at Pureora Forest Park, central North Island. Mice (n=522 collected) were fewest in unlogged native forest, more abundant in road edge cutover forest, and most abundant in a young (5-10 year old) plantation. Traps catching most mice were set in dense ground cover under a low, sparse canopy. Ship rats (n=1793) were absent from the young plantation, present but not abundant in older exotic forest, and abundant in all native forest regardless of logging history. Traps set on warmer, steeper sites caught most ship rats, and those set in early successional habitats caught fewest. There was a marked reciprocal relationship between the distributions of ship rats and of mice: the proportion of mice in the total catch of rodents decreased significantly at the least, disturbed forest sites (P <0.001). Most (81%) Norway rats (n=43) were caught in a single trap in unlogged native forest on the bank of a stream. Stoats (n=57) were most abundant in the older exotic plantations; weasels (n=16) in the young plantation and along road edges in native forest; and ferrets (n=11) in unlogged native forest. Hedgehogs (n=290) were common in unlogged native forest far from any roads and also in older exotic forest. Our data suggest that selective logging and conversion to exotics have different effects on each of the six species we monitored. We hypothesise that (1) selective logging is likely to stimulate temporary increases in the numbers of mice and weasels, but not rats or stoats, and (2) after conversion to exotic forest, mice and occasionally weasels will be abundant at first but will gradually be replaced by ship rats and stoats as the forest matures.     

The morphology, distribution and conservation implications of introduced rats, Rattus spp. in the granitic Seychelles

There are no native land mammals in the Seychelles archipelago other than bats. Introduced rats have reduced the conservation value of most islands. This paper compares the results of rat‐trapping carried out on eight islands in the granitic Seychelles, between July 1999 and April 2000. Trapping was carried out in both the dry and wet seasons. Three introduced rodent species were caught, including two species of rat (ship rat Rattus rattus Linnaeus and Norway rat R. norvegicus Berkenhout), but only one Rattus species occurred on each island. Both rat species were smaller than European or Asian conspecifics, and there were variations in the size and appearance of rats on different islands. Inter‐island differences in size and pelage colour are discussed.     

Using genetic techniques to quantify reinvasion,survival and in situ breeding rates during control operations

Determining the origin of individuals caught during a control/eradication programme enables conservation managers to assess the reinvasion rates of their target species and evaluate the level of success of their control methods. We examine how genetic techniques can focus management by distinguishing between hypotheses of ‘reinvasion’ and ‘survivor’, and defining kin groups for invasive stoats (Mustela erminea) on Secretary Island, New Zealand. 205 stoats caught on the island were genotyped at 16 microsatellite loci, along with 40 stoats from the opposing mainland coast, and the age and sex were determined for each individual. Using these data, we compare and combine a variety of genetic techniques including genetic clustering, population assignment and kinship‐based techniques to assess the origin of each stoat. The population history and individual movement could be described in fine detail, with results indicating that both in‐situ survival and breeding, and reinvasion are occurring. Immigration to the island was found to be generally low, apart from in 1 year where around 8 stoats emigrated from the mainland. This increased immigration was probably linked to a stoat population spike on the mainland in that year, caused by a masting event of southern beech forest (Nothofagus sp.) and the subsequent rodent irruption. Our study provides an example of some of the ways genetic analyses can feed directly into informing management practices for invasive species.     

Possible domestic dispersal of Bactrocera dorsalis during 2015 occurrence on the small islands of south‐western Japan

Although Japan has maintained the state of eradication of the oriental fruit fly, Bactrocera dorsalis Hendel complex (Diptera: Tephritidae) since 1986, B. dorsalis complex are occasionally trapped within a limited area and short time period on Japan's small south‐western islands. Trapping events occurred on Amamioshima Island, Kagoshima Prefecture, in 2015, and male adults were also caught in surveillance traps on neighbouring Tokunoshima (south of Amamioshima) and Yakushima (north of Amamioshima) islands as well as on several islands in Okinawa Prefecture (south of Amamioshima). To investigate possible domestic dispersal of the pest from Amamioshima Island, a trajectory analysis was performed to reveal the timing and flight distance of possible dispersal. Surveillance data indicated that the first trappings on Tokunoshima and Yakushima islands occurred in November, which was after the number of trapped male adults on Amamioshima Island peaked in late October. Backward trajectories from Tokunoshima Island beginning 14 days prior to the first trapping passed over Amamioshima Island, which suggested possible dispersal between the islands. Several backward trajectories from the islands in Okinawa Prefecture also passed Amamioshima Island. Moreover, forward trajectories from Amamioshima Island beginning during peak occurrence, from mid‐October to early November, arrived over Tokunoshima, Iheya and Okinawa islands where the flies were caught. Such circumstantial evidence suggests that domestic dispersal of B. dorsalis complex occurred on Amamioshima Island. A situation similar to that on Tokunoshima Island also occurred on Yakushima Island. However, since the number of traps has been limited until the first capture, it was not clear when and from where the immigrants arrived on the island.     

Have the Harmful Effects of Introduced Rats on Islands been Exaggerated?

Introduced rats are now being eradicated from many islands. Increasingly, these eradications are contested by activists claiming moral, legal, cultural, historic or scientific reasons and poorly documented evidence of effects. We reviewed the global literature on the effects of rats on island flora and fauna. We then used New Zealand as a case study because of its four-decade history of rat eradications and many detailed and innovative studies of how rats affect native species. These include use of exclosures, local manipulations of rat populations, video surveillance, and measurements of responses following eradications. The most intensive studies have been on the Pacific rat (Rattus exulans), a small South-East Asian species spread by Polynesians throughout the Pacific. These and the more recently introduced Norway rat (R. norvegicus) and ship (roof) rat (R. rattus) suppress some forest plants, and are associated with extinctions or declines of flightless invertebrates, ground-dwelling reptiles, land birds, and burrowing seabirds. On islands off France, Norway rats are also implicated in declines of shrews. Globally, ship rats were associated with declines or extinctions of the largest number of indigenous vertebrate species (60), including small mammals such as deer mice and bats. Effects of rats on forest trees and seabird populations are sufficiently pervasive to affect ecosystem structure and function. However, the data are patchy. Deficiencies in our knowledge would be reduced by documenting distribution and abundance of indigenous species before and after eradications. Comprehensive measurements of the responses of indigenous species to rat eradications would enable the development of testable models of rat invasion effects.     

A histochemical study of the swimming musculature of Antarctic fish

The kiore, once common throughout New Zealand, had disappeared from most of the country by the end of the 19th century, and is now found only on certain offshore islands and in areas of Fiordland where at least one of the three introduced European rodent species is absent. It is usually accepted that the kiore was displaced by ship rats (Rattus r. rattus) and Norway rats (R. norvegicus). However, recent investigations on Stewart Island have revealed kiore, ship rats, and Norway rats living in close association, but in the absence of mice (Mus musculus). In the area studied the kiore seemed to inhabit mainly grassland. Re‐examination of possible reasons for the decline of the kiore strongly suggests that competition from mice has been a major contributing factor. It seems that in New Zealand a niche no longer exists for kiore once mice, ship rats, and Norway rats have all become established.     

Trial reintroduction of buff weka to an unfenced mainland site in central South Island,New Zealand

While there have been significant conservation successes through restoration of island biodiversity following the eradication of invasive predators, a major challenge remains to reintroduce native species within larger mainland systems that support suites of introduced mammalian predators. Strategies to enhance establishment and persistence of reintroduced populations include pre‐release management to reduce post‐release dispersal, and habitat restoration such as predator control at release sites. Evaluation of such strategies critically requires strategic and intensive post‐release monitoring to identify drivers of success or the specific causes of failures. The buff weka (Gallirallus australis hectori), a flightless rail, was reintroduced to an unfenced mainland island on New Zealand's South. Past reintroductions on the mainland have all failed, but lack of post‐release monitoring has meant the exact cause and timing of failures is unknown. We investigate the ability of buff weka to establish a mainland population in conjunction with high intensity predator control. Nineteen buff weka (15 males, 4 females) were transferred from predator‐free islands in Lake Wakatipu, South Island, to Motatapu Station and held in a pre‐release enclosure for 6 weeks. Using a combination of very high frequency (VHF) and Global Positioning System (GPS) telemetry, released birds were monitored every 2 days for 4 months post‐release. Following release, no buff weka dispersed off Motatapu Station. Survival, however, was low and by the end of the study, 12 (63%) buff weka had been predated by introduced mustelids, ferrets (Mustela furo) and stoats (Mustela erminea). The lack of dispersal by buff weka suggests the presence of favourable resources on Motatapu Station. However, the low survival rate indicates that the predator‐trapping network was insufficient to suppress predator numbers to a level low enough for buff weka population persistence.     

Reduced flight capability in British Virgin Island populations of a wing-dimorphic insect: the role of habitat isolation, persistence, and structure

1. The effects of habitat isolation, persistence, and host‐plant structure on the incidence of dispersal capability (per cent macroptery) in populations of the delphacid planthopper Toya venilia were examined throughout the British Virgin Islands. The host plant of this delphacid is salt grass Sporobolus virginicus, which grows either in undisturbed habitats (large expanses on intertidal salt flats and around the margins of salt ponds, or small patches of sparse vegetation on sand dunes along the shore), or in less persistent, disturbed habitats (managed lawns). 2. Both sexes of T. venilia were significantly more macropterous in disturbed habitats (77.1% in males, 12.5% in females) than in more persistent, undisturbed habitats (19.2% in males, < 1% in females). 3. Males exhibited significantly higher levels of macroptery (26.9 ± 7.6%) than did females (2.0 ± 1.7%), and per cent macroptery was positively density dependent for both sexes in field populations. 4. There was no evidence that the low incidence of female macroptery in a subset of island populations inhabiting natural habitats (1.7 ± 1.2%) was attributable to the effects of isolation on oceanic islands. The incidence of macroptery in British Virgin Island populations of T. venilia was not different from that observed in mainland delphacid species existing in habitats of similar duration. 5. Rather, the persistence of most salt grass habitats throughout the British Virgin Islands best explains the evolution of flight reduction in females of this island‐inhabiting delphacid. 6. Males were significantly more macropterous in populations occupying dune vegetation (37.6 ± 9.8%) than in populations occupying salt flat–pond margin habitats (7.6 ± 5.6%). By contrast, females exhibited low levels of macroptery in both dune (0%) and salt flat–pond margin (< 1%) habitats. Variation in salt‐grass structure probably underlies this habitat‐related difference in macroptery because flight‐capable males of planthoppers are better able to locate females in the sparse‐structured grass growing on dunes. This habitat‐related difference in male macroptery accounted for the generally higher level of macroptery observed in males than in females throughout the islands. 7. The importance of habitat persistence and structure in explaining the incidence of dispersal capability in T. venilia is probably indicative of the key role these two factors play in shaping the dispersal strategies of many insects.