Modelling species distributions, skewed sampling, spatial autocorrelation, and implausible evolutionary histories

Abedin et al. (2025) modelled the distribution range of suitable habitat for a primate, the Bengal slow loris (Nycticebus bengalensis), in Asia. Only a small part of their current habitat is suitable, and this will reduce substantially with climate change. Low ecological connectivity was attributed to the formation of major river systems and limited dispersal abilities. A new model was presented to explain the disjunct distribution of African and Asian lorises. I find five major issues with the study (1) The study area includes the range of another, allopatric, slow loris species; (2) sample locations are highly unequally distributed excluding the eastern third of the species’ range; (3) there is strong spatial autocorrelation between sample locations; (4) the dispersal scenario of lorises from Africa to Asia at ~ 200Mya is totally implausible as this predates the origin of the Order Primate by ~ 100 million years; (5) an overstatement of the role of large rivers in shaping the distribution of Bengal slow lorises as, remarkably, these slow lorises are proficient swimmers and none of the five major rivers systems in Asia has acted as barriers for this species. Range of the Bengal slow loris (Nycticebus bengalensis) with the provincial mean low temperature in the coldest month and the approximate sampling locations used in species distribution modelling by Abedin et al. (2025). The top photograph shows a Bengal slow loris still wet after crossing a stream in eastern Bangladesh (from Hasan et al. 2024) and the bottom photo shows a Bengal slow loris swimming between mangrove islands in southern Myanmar (from Somura and Manome 2024).