THE PROBLEMS WITH MULTISPECIES CONSERVATION - DO HOTSPOTS, IDEAL RESERVES AND EXISTING RESERVES COINCIDE

South Africa urgently requires a national strategic plan for the conse rvation of the country's biodiversity. The formulation of such a plan would be relatively easy if centres (hotspots) of richness, endemism a nd rarity were congruent, both within and among many different taxa, i f these hotspots captured a large proportion of the total species, and if hotspots fell within existing reserves. The investigation of six v ertebrate taxa (viz freshwater fish, frogs, tortoises and terrapins, s nakes, birds, and various mammal orders) at a national scale reveals t hat hotspots are not coincident within taxa. Centres of richness are c oncentrated in the north-eastern areas of the country, whereas endemis m is concentrated in the south-west, and centres of rare and threatene d (Red Data Book) species can be in either area. Paired comparisons am ong taxa reveal greatly varying proportional overlaps of species richn ess hotspots (0-72%). The proportion of total species falling in hotsp ots ranges from low (66% for fish) to high (92% for birds). Hotspots a re thus not an efficient method of siting representative (ideal) reser ves. In order to design a more representative reserve system to protec t all vertebrate species, a complementarity algorithm was applied to a ll taxa separately, and then to all taxa combined. The combined analys is yielded more efficient results (66 reserves are required to represe nt all 1074 species at least once) than the separate analyses (97 rese rves). Many of these representative reserves coincide with both hotspo ts and existing reserves, and over 85% of the hotspots of most taxa co incide with existing reserves, thus South Africa's vertebrate fauna co uld be more effectively protected with only moderate acquisition of ne w, well-sited reserves. A biome analysis reveals that these reserves w ill also have to incorporate areas of those biomes that are currently inadequately protected, viz grassland, Succulent Karoo and Nama-Karoo. The methods outlined in this paper should be applied to as many other taxa as possible, in order to aid the formulation of a national strat egic plan for biodiversity conservation.