Local and regional abundance of exotic plant species on Mediterranean islands: species traits or island attributes? F. Lloret Centre for Ecological Research and Forestry Applications, Universitat Autònoma de Barcelona, Spain P.W. Lambdon & P.E. Hulme NERC Centre for Ecology and Hydrology, Banchory, UK I. Camarda & G. Brundu Dipartimento di Botanica ed Ecologia vegetale, Università di Sassari, Italy F. Médail Institut Méditerranéen d'Ecologie et de Paléoécologie (IMEP), Université d’Aix-Marseille III, France Keywords: dispersal attributes, exotic plant species, growth form, invasion, local-regional scale, Mediterranean Basin ABSTRACT: Biological invasions by non-native or “exotic” plant species are widely recognised as an important component of human-caused global environmental change, often resulting in a sig- nificant loss in the economic value, biological diversity and function of invaded ecosystems. A fundamental challenge posed by biological invasions has been the identification of characteristics associated with the invasion success of exotic species. We have examined these questions at differ- ent spatial scales. Firstly, the importance of life-history traits and biogeographic origin on the local (within island) and regional (between island) abundance of 396 exotic species was examined across a set of key, well-studied islands in the Mediterranean Basin. Secondly, a larger number of species were assessed across a wider range of islands, using less detailed published information. Naturali- zation rates (the number of islands on which they were recorded to be naturalized) were the least reliable indicators of true invasiveness, but a combined index from both spatial scales gave the best overall picture of invasion success. We found that life-history traits may play a lesser role than in- vasion history. However, local abundance was highly variable between islands and even between ecologically-similar species. The comparison of these approaches demonstrates that invasion pat- terns should be interpreted at both local and regional scales, but the stochastic nature of biological invasions may limit deterministic interpretations of invasion patterns, especially if islands are stud- ied in isolation. 1 INTRODUCTION Biological invasions by non-native or “exotic” plant species are considered one of the most impor- tant environmental problems resulting from human-caused global environmental change. The con- sequences of such invasions include losses to the economic value, biological diversity and function of the affected ecosystems (Mooney and Hobbs 2000, Mack et al. 2000, Pimentel et al. 2001). There are theoretical reasons to expect dramatic effects of plant invaders on isolated habitats, such as those on geographical islands (Lonsdale 1999). Particularly, their importance in Mediterranean- type ecosystems has been extensively documented (Di Castri et al. 1990, Vilà et al. 2001). In order to develop screening procedures for potentially harmful species, it will be necessary to devise methods for assessing invasion risk: a function of both the consequences of invasion, and the likelihood of its occurrence. At present, the consequences are poorly-understood. In contrast, likelihood, in terms of known invasion success, can be evaluated from copious data on the extent and history of plant invasions. “Invasion success” is, in turn, a function of both the likelihood of 1