Biodiversity conservation across taxa and landscapes requires many small as well as single large habitat fragments.

Agricultural intensification has been shown to reduce biodiversity through processes such as habitat degradation and fragmentation. We tested whether several small or single large habitat fragments (re-visiting the 'single large or several small' debate) support more species across a wide range of taxonomic groups (plants, leafhoppers, true bugs, snails). Our study comprised 14 small (<1 ha) and 14 large (1.5-8 ha) fragments of calcareous grassland in Central Germany along orthogonal gradients of landscape complexity and habitat connectivity. Each taxon was sampled on six plots per fragment. Across taxa, species richness did not differ between large and small fragments, whereas species-area accumulation curves showed that both overall and specialist species richness was much higher on several small fragments of calcareous grassland than on few large fragments. On average, 85% of the overall species richness was recorded on all small fragments taken together (4.6 ha), whereas the two largest ones (15.1 ha) only accounted for 37% of the species. This could be due to the greater geographic extent covered by many small fragments. However, community composition differed strongly between large and small fragments, and some of the rarest specialist species appeared to be confined to large fragments. The surrounding landscape did not show any consistent effects on species richness and community composition. Our results show that both single large and many small fragments are needed to promote landscape-wide biodiversity across taxa. We therefore question the focus on large fragments only and call for a new diversified habitat fragmentation strategy for biodiversity conservation.