Due to the current species extinction crisis, there is an urgent need to identify the most threatened areas of ex-ceptionally high biodiversity and rates of endemism (i.e., "hotspots"; Mittermeier et al. 1998; Myers 1988; Reid 1998). Conservation strategies represent a crucial issue in the mediterranean biome because this area, which represents only 2% of the world's surface, houses 20% of the world's total floristic richness (Médail & Quézel 1997). Myers initially (1988, 1990) defined 14 hotspots in the tropical biome and four in mediterranean biocli-mates (southwestern Australia, Cape Region of South Af-rica, California, and part of Chile). Like the four other mediterranean areas, the Mediterranean Basin is one of the world's major centers for plant diversity, where 10% of the world's higher plants can be found in an area rep-resenting only 1.6% of the Earth's surface (Médail & Quézel 1997). The prominent role played by these areas as reservoirs for plant biodiversity has been emphasized by Myers (1990). He hesitated, however, to group the whole Mediterranean Basin into one single hotspot be-cause it covers such a large surface area, and insufficient data were available for certain regions. In this context, Médail and Quézel (1997) performed a global survey of plant richness and endemism to more precisely define hotspots in the Mediterranean Basin; they identified 10 hotspots. Three main approaches, however, have been taken in recent studies performed by international con-servation organizations to define priority conservation areas in the Mediterranean Basin. The first approach, the "megadiversity countries" project (Mittermeier et al. 1997), defined 17 megadiverse countries based primarily on plant endemism, species richness, and political boundaries. The top countries listed were Indonesia, Colombia, Brazil, Australia, Mex-ico, and Madagascar, but the Mediterranean Basin was completely left out, although, as some authors have re-cently pointed out (Heywood 1995; Médail & Quézel 1997), several tropical or subtropical countries have lower plant diversity than the Mediterranean Basin. For example, all of tropical Africa has the same plant rich-ness (30,000 taxa) as the circum-Mediterranean region in a surface area four times larger. Furthermore, the Mediterranean Basin possesses 10.8 species/1000 km 2 , which is higher than the 3.1 species/1000 km 2 in China, 4.7 in Zaire and in India, and 6.5 in Brazil, but lower than the 40 in Colombia and 90 in Panama (Médail & Quézel 1997). In the second approach, on the opposite extreme, some recent conservation strategies treat the Mediterra-nean Basin as a unique hotspot. The Global 200 project, performed by Olson and Dinerstein (1997, 1998), included the whole area in one large ecoregion defined by one rare major habitat type (i.e., Mediterranean shrublands and woodlands). These authors emphasized that "more-detailed, fine-scale analyses are essential to identify im-portant targets within ecoregions." In a recent work, Mittermeier et al. (1998) defined 24 hotspots in which, based on plant endemism, the whole Mediterranean Ba-sin was listed second after tropical Andes. Nevertheless, including all the different parts of the Mediterranean Ba-sin in one unique hotspot seems to oversimplify the situ-ation, even on a global scale. In fact, this area has many highly differentiated biogeographical patterns and land-use practices in space and time within its complex and heterogenous landscapes. The Mediterranean region constitutes both a refuge area and one that encourages floral exchange and active plant speciation (Quézel 1978, 1985). In the western ba-sin, high-endemism areas are related to the age of the geological platform and relict endemics prevail, whereas in the east, vicariant endemism is high due to the moder-ate role of glaciations and the presence of ultramafic rocks (Verlaque et al. 1997). Between the northern and southern coasts, human effects have created two differ-ent situations (Barbero et al. 1990). The collapse of the * email firstname.lastname@example.org Paper submitted September 22, 1998; revised manuscript accepted June 9, 1999.