At the landscape level, the four-horned antelope is confined to tropical dry deciduous forests and within these, their distribution is patchy. Various factors have been proposed as determinants for their patchy distribution within landscapes, but none provided an adequate explanation. We hypothesized that availability of a constant supply of forage influenced the species distribution. We found that the four-horned antelope mainly fed on fruits and flowers, and that a total of 60% of the tree species in Panna Tiger Reserve bore fruits at different times of the year. High tree species richness in habitat patches was considered a surrogate for constant supply of forage for the four-horned antelope. Data from 547 sighting locations between 2002 and 2006 and six spatial layers were analysed using maximum entropy to produce a probability distribution model for the four-horned antelope in Panna Tiger Reserve. Our model predicted that habitat patches summing up to only 9.5% of the 543 km2 of Panna Tiger Reserve had high probability of distribution (>0.5) of four-horned antelope. Although all variables contributed to the distribution model of the four-horned antelope, explanatory power was highest for tree species richness within habitat patches. The distribution of four-horned antelope within tropical dry deciduous forests can be treated as an indicator of high tree diversity and hence habitat quality.

A new technique for estimation of tree species richness in tropical forests was developed and tested in this study. The missing species extrapolation technique is based on the assumption that the rate of accumulation of new species with increasing area is the same within sampled and unsampled areas of similar habitat within a region. This new technique was tested on four 1-ha sites of subtropical moist forest in Rio Bravo, Belize for which complete species counts of trees ≥10cm dbh are available. Estimates given by the technique are 115, 98, 93 and 106% of the actual species counts within the four sites. The accuracy of these estimates was found to be equal to or better than most others obtained by using existing methods of extrapolation. The missing species extrapolation technique was also applied to a 21-ha riparian forest patch in the Mountain Pine Ridge, Belize, where it was used to obtain an estimate of the number of tree species ≥10cm dbh in the entire patch.