Revision of Ambisporaceae, with three new genera and one new species and a morphological identification key for all the species currently attributed to this family

The objective of this study was to re-analyse the molecular phylogeny and/or the morphology of all species, which have been attributed to the so-far mono-generic fungal family Ambisporaceae. The genus Ambispora has been well-known for its spore bi-morphy described even from single spore clusters. Triple-walled spores are differentiated on sporiferous saccules, while mono-walled spores are formed on simple subtending hyphae. New phylogenetic analyses reveal dissimilarities of [≥]10% in partial nrDNA gene of three different stable phylogenetic clades and thus suggest the division of Ambispora into three genera, which simultaneously request for advanced morphological separations. These advances are primarily based on the more diverse spore wall composition of the ambisporoid-acaulosporoid morph rather than on the rather simple-glomoid morph. While all known species of the triple-walled morph have an evanescent to semi-permanent outer spore wall, i) Am. fennica, Am. brasiliensis, Am. gerdemannii and Am. nicolsonii have a smooth, permanent central spore wall (Am. fennica clade, A), ii) the central wall of Am. appendicula, Am. callosa, Am. leptoticha and Am. jimgerdemannii is alveolate (Am. appendicula clade, B), and iii) the central wall of Am. granatensis is smooth, but easily degraded, thus rather short-lived and not permanent but evanescent (Am. granatensis clade, C). In conclusion, species of the Am. fennica clade represent the genus Ambispora, while species of the Am. appendicula clade represent the new genus Appendiculaspora, and the mono-specific Am. granatensis clade represents the new genus Ephemerapareta. Species of an additional morph, with triple-walled spores, but apparently formed on subtending hyphae, and having a diagnostic reticulate, football-like middle wall, are here separated from the revised genus Ambispora based solely on morphological analyses, since molecular identification analyses so far failed and remained merely unknown. This later morph and genus is based on the type species Pelotaspora reticulata comb. nov, and on P. austrolatina sp. nov. Concomitant molecular phylogenetic and morphological analyses are needed to attribute not only Pelotaspora spp., but also those species, for which hitherto only the ambisporoid-glomoid morph has been observed correctly within the family Ambisporaceae. Without molecular analyses, such species with glomoid but unknown ambisporoid-acaulosporoid morph have to be retained within Ambispora.