The spectacular diversity of sclerophyll plants in the Cape Floristic Region in South Africa and Australia's Southwest Floristic Region has been attributed to either explosive radiation on infertile soils under fire-prone, summer-dry climates or sustained accretion of species under inferred stable climate regimes. However, the very poor fossil record of these regions has made these ideas difficult to test. Here, we reconstruct ecological-scale plant species richness from an exceptionally well-preserved fossil flora. We show that a hyperdiverse sclerophyll flora existed under high-rainfall, summerwet climates in the Early Pleistocene in southeastern Australia. The sclerophyll flora of this region must, therefore, have suffered subsequent extinctions to result in its current relatively low diversity. This regional loss of sclerophyll diversity occurred at the same time as a loss of rainforest diversity and cannot be explained by ecological substitution of species of one ecological type by another type. We show that sclerophyll hyperdiversity has developed in distinctly non-Mediterranean climates, and this diversity is, therefore, more likely a response to long-term climate stability. Climate stability may have both reduced the intensity of extinctions associated with the Pleistocene climate cycles and promoted the accumulation of species richness by encouraging genetic divergence between populations and discouraging plant dispersal.macrofossil flora | species diversity | species-area curve | glacial-interglacial cycles M editerranean climate regions support ∼20% of vascular plant species on 5% of Earth's surface (1) and show positive anomalies in the latitudinal plant species diversity gradient. Researchers have used two approaches to explain this gradient. The first approach emphasizes the importance of contemporary environmental processes. For example, some models use energy and water availability to explain global patterns of diversity (2, 3). Alternatively, historical models focus on geological-scale processes and often invoke the contrasting antiquity of tropical vs. temperate climates. For example, the tropical conservatism hypothesis (4) argues that the high diversity of tropical floras accumulated over tens of millions of years during warm, ever-wet, early-middle Cenozoic climates. This model infers that temperate floras are depauperate because they were derived relatively recently from a small number of thermophilic lineages that were able to colonize novel temperate climates during the late Cenozoic.