On the enigmatic symbiotic polychaete 'Parasyllidea' humesiPettibone, 1961 (Hesionidae): TAXONOMY, phylogeny and behaviour

Abstract

The hesionid genus Parasyllidea differs from Oxydromus in lacking median antennae. It was originally described to include a single species, P.humesi, known only from its original description. This was based on specimens from mangrove swamps at Pointe-Noire (Republic of Congo, West Africa), living endosymbiotically with the bivalve Tellina nymphalis. Lately, the genus included P.blacki and P.australiensis. A new population of P.humesi was recently found at the upper intertidal level of Rio San Pedro salt marsh in Cádiz Bay (eastern Atlantic, Iberian Peninsula). It was also living endosymbiotically, but with another bivalve, Scrobicularia plana. Some Iberian and Congolese specimens revealed the presence of a small papilla-like central antenna associated with the prostomial median ridge, which raised some doubts on the validity of the genus Parasyllidea. A phylogenetic analysis based on the mitochondrial COI and 16S and the nuclear 18S and 28S genes confirms Parasyllidea as a junior synonym of Oxydromus. Therefore, in this paper, P.humesi is fully re-described as Oxydromus humesi comb. nov. The worm has never been reported as free-living. Previously, the association appeared to be an obligate symbiosis, closer to parasitism, as infested hosts had lower relative biomasses than non-infested ones and the worm did not occur locally inside any other bivalve co-habiting the intertidal salt marsh. The finding of a highly infested population (>85% in the specimens longer than 20mm) of a new host at the lower subtidal part of Rio San Pedro mouth, the bivalve Psammotreta cumana, led us to discuss the host-specificity of O.humesi. In addition, the observation of living specimens during sampling and laboratory handling enabled detailed observations of the host-entering behaviour of the specimens living with S.plana, which are also described and illustrated. Living, uninfested specimens of P.cumana have not been obtained, preventing us from checking the host-entering behaviour in the new host. The significance of the intraspecific attacks observed in experimental conditions is also discussed.