Females of most arthropod species mate with more than one male during their lifetime and can store sperm for extensive periods. Consequently, male adaptations to sperm competition and cryptic female choice are the main factors that determine paternity success. In order to understand the processes that lead to differential fertilization success, the relative number of sperm transferred and the number of sperm stored by the female need to be established. To this end, a reliable sperm counting procedure is required. We assessed the reliability and variability of sperm counts in the nuptial-gift-giving spider Pisaura mirabilis (Clerck, 1757) by comparing counts from the same sample and by comparing different storage treatments before sperm counts. We used male sperm storage organs (pedipalps) that were processed immediately after dissection, pedipalps fixed in 80% ethanol and stored at room temperature, and pedipalps that were frozen at -20°C, -40°C, -80°C. We further tested for an influence of the counting solution (CASY®ton versus saline-triton-x buffer) on reliability and variability. The sperm counting protocol resulted in homogeneous distribution of sperm over the counting chamber and highly correlated counts drawn from one sample. Although all treatments kept the sperm in a good state, the samples previously stored in ethanol contained many broken bits of material, which made the counting procedure considerably more time consuming. The buffer used had no effect on the homogeneity and number of sperm found. Whether and how the material was fixed had no significant effect on the average number of sperm found in the palps.