The use of non-pollen palynomorphs, and among them spores of coprophilous fungi, has become greatly important in palaeoecological studies. Particularly, the genus Sporormiella has been demonstrated to be the most valuable proxy for the presence of wild and domestic herbivores. This genus could also be used to determine livestock density and reconstruct pastoral pressure during the Holocene. Non-standard counting methods have been established to determine coprophilous fungal spore abundance in sediments. Moreover, these analyses are faced with the recurrent problem of setting the minimum counting sum as small as possible to save time. We researched the reliability of Sporormiella concentration estimates based on different counting sums, using low to high count samples. Box-plots indicate that the variability of inferred Sporormiella concentrations decreases progressively with increasing sums. Statistical comparisons show that the means of box-plots became stabilised after the counts have reached 300-350 exotic marker grains. Moreover, a count of 300-350 exotic marker grains is sufficient to produce a Sporormiella concentration estimate, whatever the amount. Finally, we propose that this counting limit is valid for other fungal spores as well.