1. Monocultural rubber plantations have replaced tropical forest, causing biodiversity loss. While protecting intact or semi-intact biodiverse forest is paramount, improving biodiversity value within the 11.4 million hectares of existing rubber plantations could offer important conservation benefits, if yields are also maintained. Some farmers practice agroforestry with high-yielding clonal rubber varieties to increase and diversify incomes. Here, we ask whether such rubber agroforestry improves biodiversity value or affects rubber yields relative to monoculture.2. We surveyed birds, fruit-feeding butterflies and reptiles in 25 monocultural and 39 agroforest smallholder rubber plots in Thailand, the world's biggest rubber producer.Management and vegetation structure data were collected from each plot, and landscape composition around plots was quantified. Rubber yield data were collected for a separate set of 34 monocultural and 47 agroforest rubber plots in the same region.3. Reported rubber yields did not differ between agroforests and monocultures, meaning adoption of agroforestry in this context should not increase land demand for natural rubber. Butterfly richness was greater in agroforests, where richness increased with greater natural forest extent in the landscape. Bird and reptile richness were similar between agroforests and monocultures, but bird richness increased with the height of herbaceous vegetation inside rubber plots.
Species composition of butterflies differed between agroforests and monocul-tures, and in response to natural forest extent, while bird composition was influenced by herbaceous vegetation height within plots, the density of non-rubber trees within plots (representing agroforestry complexity) and natural forest extent in the landscape.
Reptile composition was influenced by canopy cover and openThis is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.