The elevational gradient of species richness: a uniform pattern?