Slow sand filters (SSFs) for drinking water production are habitats for diverse microbes from multiple domains of life, which are integral to the ability of SSFs to purify water. While cultivation-independent analyses of the prokaryotic communities of SSFs have provided valuable insights, little attention has been paid to fungi inhabiting SSFs. This study characterized the fungal communities in the sand biofilm of one established, one inoculated and one non-inoculated SSF. The removal of the top-layer of sand ("scraping") allowed fungal communities in the top and subsurface layers of sand to be analyzed using amplicon sequencing of the ITS2 region of fungal rRNA genes. The top layers of SSF sand contained fungal communities dominated by phylum Ascomycota (43.5-75.6 %). After scraping, high abundances (>
70 %) of phylum Rozellomycota were revealed in the established filter. These fungi were also detected in an inoculated filter, but not in a non-inoculated filter, suggesting potential dispersal to new filters by inoculation. The diverse Rozellomycota sequences potentially represented 6 different order-level clades, with most being related to previously observed Branch03 Rozellomycota. Their roles in SSF function are unknown but may be related to the removal of indicator bacteria as this phylum includes potential parasites of grazing eukaryotes. Fungi known to constitute microbial risk or contribute to micropollutant biodegradation were in low abundance and only sporadically detected. Lifestyle traits could be predicted for 61.8 % of fungi in the SSF biofilm
most of these were saprotrophic microfungi or yeasts. This study presents an overview of the composition of fungal communities in full-scale SSF, and their potential interactions with water quality. It also highlights the need for more knowledge regarding the ecology of "dark matter"-fungi, such as Rozellomycota, and presents an accessible and societally relevant environment for future research of these microbes.