Corsica is the biggest French Mediterranean island. It has an area of 8680 km² with coasts of 1000 km long. The coastal strip is very narrow, Corsica being highly mountainous with an elevation of 2710 m. This island is characterized by a Mediterranean climate with an average annual temperature of 16.1 °C and average annual precipitation of 698 mm. The population density is 30 inhabitants/km² mainly concentrated on the coastal strip. In summer, during the activity peak of tourism, the population increase from about 200 000 inhabitants to more than one million.

A comprehensive sampling of the Corsican coastline has been performed, visiting notably sites inspected by Casewitz-Weulersse and Brun [12] (Fig. 1, sites 2, 6, 13, 15, 17 and 22). We conducted the survey of 79 sites from June to September in 2007 and in 2008. At each site, we inspected the ground, under stones or dead woods during a 20 minutes scan. If no Argentine ants were found, the site was stated as L. humile-free. Inversely, when a site was scored as infested, we calculated an infestation index which was the ratio between the number of Argentine ant observations and the total number of inspections during the 20 minutes scan. One corresponded to a highly infested site and zero to a free site. In both cases, infested or not, we recorded the date, the habitat type, the coordinates and a sample was taken of all species encountered and stored in a vial of 75% ethanol. This fine description with coordinates and infestation index of the whole infested sites will permit a better survey in the future of the Argentine ant expansion dynamic.

Argentine ants are present and expanding in Corsica since the last survey by Casewitz-Weulersse and Brun [12]. Ten years later, we found L. humile along the coast but not continuously (Fig. 1). We recorded this invader in 28% of the 79 prospected sites (22/79) (Fig. 1). We recorded three major infested zones, one on the oriental coast in the South of Bastia (Fig. 1, sites 1 to 4) and two on the occidental coast, in the gulf of Ajaccio (Fig. 1, sites 15 to 17) and in the gulf of Propriano (Fig. 1, sites 11 to 14). We recorded Argentine ants in seven urban sites (31.8% of occurrences), four had already been noticed by Casewitz-Weulersse and Brun [12] Propriano, Porticcio, Ajaccio and Calvi (Fig. 1, sites 22, 17, 15, 12, respectively) and three were recorded for the first time, i.e. Campomoro, Porto-Vecchio and Vignola (Fig. 1, sites 10, 7, 18, respectively). Contrary to our expectations, urban sites were not very highly invaded, with a mean infestation index of 0.36 ± 0.07 S.E. (n=7). This could be explained by the use of different chemical treatments for management and control of this invader. In instance, during this ant survey we found in Calvi only one ant trail in a flowerpot whereas this location is the oldest infested site. Consistently, we did not found Argentine ants in Bastia, the second city of Corsica, whereas they were present few kilometres to the south in Biguglia (Fig. 1).

We also recorded Argentine ants in relative high densities in semi-natural or natural sites (15/22) with a mean infestation index of 0.56 ± 0.06 S.E. (n=12) and 0.85 ± 0.04 S.E. (n=3) respectively in coastal grassland and in forest/scrub sites. The forest/scrub habitat was the most infested among the three habitat types (P<0.05, Mann–Whitney U tests). No significant difference was recorded between urban and coastal grassland habitats (P=0.06, Mann–Whitney U test). Thirteen sites were first documented in this study. In Natura, we frequently found L. humile in open landscape, mostly in coastal grassland (12/22 sites, 54.5% of occurrence). We also mapped Argentine ants in three forest/scrub sites (13.6% of occurrence). When they were present in forest, they were always in high densities in adjacent to open habitat (Table 1, sites 3, 5, 13). Moreover, we found L. humile in protected areas with high biological value. This is the case of Cannelle, Capu Laurosu and Cala Piscona (Table 1, sites 6, 11 and 14) which are three of the four locations for a rare endemic ant plant Anchusa crispa [13].

This study documents changes in the distribution of Argentine ants in Corsica over the past decade. Argentine ants appeared to be well established within their introduced range and spreading along the Corsican coasts. The spread of Argentine ants seemed to be mainly the result of jump-dispersal through Human transport. The examples of likely human-mediated dispersal were for sites 8, 9, 16, 20 and 21 (Fig. 1). The majority of the newly infested sites recorded in this study are rather isolated. In these locations, Argentine ants were absent in the adjacent zones suggesting that they did not diffuse naturally but were introduced accidentally from foci by Human. Indeed, natural diffusion distances are usually believed to be of 150 meters per year [14] which is insufficient to explain the colonization of the new sites. This relative slow spread stems principally from the colony reproduction strategy; i.e. by budding [14].

Where the Argentine ant has invaded native habitats in other parts of the world it has predominantly been in open areas [14,15]. Our results are consistent with these studies as, among the 15 natural or semi-natural sites, 12 were characterised by open areas. We also recorded Argentine ants in forest in three times only. This is also consistent with former studies showing that forests are less susceptible to host the invasion or are not invaded. However, in these three sites, Argentine ants were present in high densities, probably naturally diffused from adjacent open areas to forest.

Despite a marked spread over the last decade, Argentine ants were present discontinuously along the coastal strip. This is particularly true for the oriental coast, where we recorded a 80 km-long line free of Argentine ants going from site 5 to site 6. This is not consistent with the widely recognized vulnerability of the island populations to biological invasions [16]. These simplified ecosystems offer generally less resistance to biological invasions. In other invaded islands, Argentine ants colonized almost the entirety of the islands. In Bermuda, Argentine ant was recorded in the 1950s. Since then, the Argentine ant is currently the most common ant both by its occurrence and its extremely high densities. Wetterer and Wetterer [17] found this ant almost everywhere they collected in Bermuda. Argentine ant was found in Madeira for more than 100 years. However, Wetterer [18] balanced its impacts considering that Argentine ant impacts were exaggerated in former studies. Gomez and Espadaler [19] recorded the presence of Argentine ants in the Balearic Islands where the infestation was considered as very important in the islands of Ibiza, Mallorca, and Menorca. In these three Mediterranean islands, Argentine ants spread rapidly, much more easily in the more humid environment [19].

Several studies demonstrated that the spread of Argentine ants could be limited by both abiotic [20] and biotic factors [15,21]. Abiotic conditions have been implicated as one of the most important factors driving the potential of the Argentine ant invasion. At the ultimate of their distribution, Argentine ants are limited by extremes in temperatures [22] and water limitation [23]. Abiotic factors can shape species distributions mainly in harsh environments while biotic factors may play a role in shaping species distributions in milder conditions. Our ant sampling showed the presence of two species widely distributed along the coasts and notably along this Argentine ant-free zone (Fig. 2). These species belonged to the genus Tapinoma, Tapinoma simrothi and Tapinoma nigerrimum. While they usually shared the same habitat, we noticed only twice co-occurrence between Argentine ants and at least one of the two Tapinoma. In these two cases, only one small nest of Argentine ants was found suggesting recent introductions. Consistently, a systematic exclusion seemed to arise between Argentine ants and Tapinoma. Bioassays between L. humile and Tapinoma testing competition for food and space will permit to test the hypothesis of a biotic resistance. Accordingly, we will be able to state whether Tapinoma can delay or prevent the spread of Argentine ants, which would be relatively unusual on island ecosystems.