ABSTRACT

The effects of genetic erosion on the viability of small populations following habitat fragmentation are understood in theory but the critical early stages of the process have gone undocumented as the changes are rapid and difficult to monitor. We found it is possible to monitor genetic erosion in recently fragmented populations by genotyping with panels of 6–7 hypervariable nuclear microsatellite loci as markers of variability. We studied changes in variability in populations of three small mammals isolated on forest fragments in Thailand when the creation of Chiew Larn reservoir flooded the forested KhlongSaeng valley in 1987 and left about 100 rainforest fragments as islands in the lake. Mark-recapture surveys in years 5–8 post-fragmentation on island and matched undisturbed mainland sites showed that habitat fragmentation led to the onset of genetic erosion in surviving populations of a forest rat, Maxomys surifer, tree mouse, Chiropodomys gliroides, and tree shrew, Tupaia glis. Demographic and genetic responses to fragmentation were species-specific, reflecting differences in life history and behaviour. Allelic variation was invariably lost faster than heterozygosity and, in C. gliroides, genetic erosion preceded demographic decline. We found that small, recently isolated populations lose variation faster than allowed for in current conservation practice and that genetic erosion may commence before the onset of obvious demographic decline. The project has great generality throughout the increasingly fragmented humid tropics and the methods may be used to monitor genetic erosion in isolated populations of the larger mammals that are typically the focus of conservation efforts. The policy implications of our research are that populations in fragmented forests may require both ecological and genetic management if they are to survive and provide ecological services.