This material is published by Oikos (2001, vol 93, pp 497-498).
What determines territory configurations of badgers?
C. Patrick Doncaster
Blackwell and Macdonald (2000) have re-analysed models by Doncaster and Woodroffe (1993) for badger territory configurations, and drawn opposite conclusions in suggesting that sett locations are not the crucial factor in determining territory shapes and sizes. Doncaster and Woodroffe had constructed Dirichlet tessellations around badger main setts and compared their borders with the locations of actual border latrines. A close correspondence at some sites provided direct evidence that territory size and configuration could be determined solely from the locations of main setts. In contrast, Blackwell and Macdonald use the latrine locations to define Dirichlet tessellations, in order to compare the centres of these modelled territories to actual sett locations. Their method employs statistical techniques that avoid some complications in the way the original model was compared to alternatives. However, because their analysis pre-determines the size and shape of territories from latrine sites, it cannot test hypotheses about factors regulating size and shape. In this respect their conclusions over interpret the analysis, which can only test the hypothesis that Dirichlet centres associate with sett locations. Here I describe two consequences of this limitation. Firstly, any test of an exact coincidence between Dirichlet centres and setts is necessarily sensitive to underlying structure in the mosaic of habitats across which border latrines are situated, and to missing data on latrine sites. These influences can explain the observed displacement of centres from setts reported by Blackwell and Macdonald. Secondly, alternative tests of non-random association between centres and setts can provide indirect evidence for a role in territory configuration of some single point feature at least in the vicinity of the sett, if not the sett itself. A simple test of this sort indeed confirms a strong pairing of centres with setts at two out of three study sites.
Return to home page