The Pentland Hills sponge fauna (Llandovery, Telychian) consists of an unusual, aberrant assemblage, but of low diversity. A new specimen of a unique sponge, Eoghanospongia carlinslowpensis gen. et sp. nov., is described from the classic locality of R82. The mushroom-shaped, probably stalked body (peduncle attachment to body not exposed) resembles that of some living rossellids, especially Caulophacus. The sponge also shows prominent projecting pentactins and monaxon derivatives – a spicule type diagnostic of the Rossellidae among living taxa, albeit in a robust form not described from extant sponges, including Caulophacus or other pedunculate lyssacinosidans. Certain attribution to the Rossellidae is not possible from the single specimen, largely because of weak preservation of the primary spicule skeleton, but no other fossil or modern sponges show any significant similarity to it. Although similarly early relatives of the Rossellidae have recently been described from elsewhere, the new sponge is even more unexpected in being from a shallow-water environment, making the absence of rossellids through the rest of the Palaeozoic much more problematic.

The monospecific hexactinellid sponge genus Symplectella endemic to New Zealand waters was originally assigned to family Rossellidae within the order Lyssacinosida (subclass Hexasterophora), although affinities to family Euplectellidae were also noted. Seventy-eight years later, the genus was transferred to Euplectellidae (subfamily Corbitellinae) on rather subjective grounds. Here, I test these two competing taxonomic hypotheses with molecular phylogenetic methods and demonstrate that Symplectella rowi is indeed a rossellid, as was originally suggested. The genus is officially transferred back to Rossellidae (subfamily Rossellinae), which represents another small step towards a more natural classification system of glass sponges.