Brodifacoum, chlorophacinone and coumatetralyl treated pearl millet (Pennisetum typhoides) grains were tested in a desert scrub grassland to evaluate the comparative field efficacy of these three anticoagulant rodenticides. Baits were placed for 10 consecutive days in bait stations in 18 plots of one hectare each in a randomized block design. Per cent reduction in rodent population was calculated by pre and post-treatment active burrow counts, census baiting and live trapping. Results revealed that brodifacoum is significantly more effective than other two anticoagulants. However, analysis of variance revealed significant difference in the estimation of per cent reduction in the rodent population calculated by three methods, the possible reasons for which are also discussed.