The present study compares the effect of three storage media (silica, RNAlater®, ethanol) and time to extraction (1 week, 1 month and 3 months) on mitochondrial and nuclear marker amplification success in faecal DNA extracts from a sympatric community of small to medium-sized Central African forest ungulates (genera Cephalophus, Tragelaphus, Hyemoschus). The effect of storage type and time on nuclear DNA concentrations, genotyping errors and percentage recovery of consensus genotypes was also examined. Regardless of storage method, mitochondrial and nuclear amplification success was high in DNA extracted within the first week after collection. Over longer storage periods, RNAlater yielded better amplification success rates in the mitochondrial assay. However, samples stored on silica showed (i) highest nuclear DNA concentrations, (ii) best microsatellite genotyping success, (iii) lowest genotyping errors, and (iv) greatest percentage recovery of the consensus genotype. The quantity of nuclear DNA was generally a good predictor of microsatellite performance with 83% amplification success or greater achieved with sample DNA concentrations of ≥ 50 pg/µL. If faecal DNA samples are to be used for nuclear microsatellite analyses, we recommend silica as the best storage method. However, for maximum mitochondrial amplification success, RNAlater appears to be the best storage medium. In contrast, ethanol appeared inferior to the other two methods examined here and should not be used to store tropical ungulate faeces. Regardless of storage method, samples should be extracted as soon as possible after collection to ensure optimal recovery of DNA.