Forest duikers (Cephalophinae) are important targets of the bushmeat trade but a reliable and efficient assessment of these shy animals is often regarded as problematic. Using a power analysis approach, we assessed the performance of three established methods (diurnal, nocturnal and indirect dung surveys) in producing density estimates for the blue and the Ogilby's duiker (Cephalophus monticola, Cephalophus ogilbyi ogilbyi) in a 16 km² section of primary forest, systematically sampled using line transects. Direct survey methods (diurnal and nocturnal) yielded density estimates between 8.3 and 6.8 (blue duiker) and 6.5 and 4.3 animals/ km² (Ogilby's). The coefficients of variation of density estimates implied a resolution of 44% density change detectable in diurnal surveys but only of 140% in nocturnal surveys. Density estimates derived from the dung survey were lower than those of direct surveys (1.5 animals/ km² and 1.4 animals/ km², respectively). Associated coefficients of variation imply a resolution of change in density estimates detectable for changes far larger than 100%. However, dung pile density was susceptible to variation in key parameters (decay and production rates) and difficulties in data acquisition lead us to reject the dung survey as a suitable method in our particular setting. Our results suggest that diurnal direct surveying from line transects can be reliable and effective even for shy forest ungulates under (some) hunting pressure.