RESUMO
Sabellid and sabellariid polychaetes are regarded as sister groups in a number of recent phylogenetic analyses. This is based mainly on a shared specific arrangement of chaetae referred to as chaetal inversion. Remarkably, the uncini have a notopodial position in the abdomen, whereas capillary chaetae occur in the neuropodia in both taxa in contrast to the situation in putative relatives. However, in sabellids uncini and capillary chaetae change their position completely at the border between thorax and abdomen, whereas uncini are missing in the parathorax of Sabellariidae. Due to this difference the significance of the chaetal inversion for systematics has been subject to discussion for years. Serial semithin sections of parapodia of the Sabellidae Sabella pavonina, Branchiomma bombyx, Fabricia stellaris, and of the Sabellariidae Sabellaria alveolata were studied in order to obtain detailed information on their chaetal arrangement and sites of chaetal origin. SEM investigations and computer-aided 3D-reconstructions provide deep insight into the spatial organization of the rami. Though differing externally, the principal chaetal arrangement and the location of the formative sites turned out to be almost identical within the species of Sabellidae. Most chaetae are aligned in straight transverse rows with a dorsal site of origin within neuropodia and a ventral one in notopodia as is common in sedentary polychaetes. Semicircular and spiral arrangements are revealed to be modified transverse rows. Only in thoracic notopodia does an additional dorsocaudal formative site form distinct rows. The chaetal inversion in Sabellidae is additionally characterized by an abrupt change of capillary chaetae and uncini along with a sudden change of the parapodial morphology at the border between thorax and abdomen. All chaetae of S. alveolata are aligned in transverse rows with the same location of the formative sites as in sabellids and other sedentary polychaetes. However, in contrast to sabellids the chaetae are not inverted across a parathoracic abdominal border. Moreover, there is no inversion of the parapodial structure from parathorax to abdomen and the neuropodial chaetal composition changes gradually from parathorax to abdomen. The chaetal arrangement in Sabellariidae thus cannot be described as inverted along the body-axis as in Sabellidae. Evolutionary steps implied by the assumption of an inverted chaetal pattern in a supposed common ancestor are discussed. It is concluded that the specific chaetal arrangement of Sabellidae and Sabellariidae arose independently and therefore provides no support for a sistergroup relationship of sabellids and sabellariids.