In his monograph on Cirripedia from 1851, Darwin pointed to a highly unusual, plateless, and most likely parasitic barnacle of uncertain phylogenetic affinity. Darwin's barnacle was Anelasma squalicola, found on deep-water sharks of the family Etmopteridae, or lantern sharks. The barnacle is uncommon and is therefore rarely studied. Recent observations by us have shown that they occur at an unusually high prevalence on the velvet belly lantern shark, Etmopterus spinax, in restricted fjord areas of western Norway. A phylogenetic analysis based on ribosomal DNA data (16S, 18S, and 28S) from 99 selected barnacle species, including all available pedunculate barnacle sequences from GenBank, shows that A. squalicola is most closely related (sister taxon) to the pedunculate barnacle Capitulum mitella. Both C. mitella and species of Pollicipes, situated one node higher in the tree, are conventional suspension feeders from the rocky intertidal. Our phylogenetic analysis now makes it possible to establish morphological homologies between A. squalicola and its sister taxon and provides the evolutionary framework to explain the unprecedented transition from a filter-feeding barnacle to a parasitic mode of life.

Anelasma squalicola is a barnacle found attached to deep-water lantern sharks of the family Etmopteridae and is the only known cirriped on fish hosts. While A. squalicola is equipped with mouth and thoracic appendages (cirri), which are used for suspension feeding in conventional barnacles, its attachment device (peduncle) appears to have evolved into a feeding device, embedded into the tissue of its host. Here we demonstrate, through comparisons of the feeding apparatuses between A. squalicola and conventional suspension-feeding barnacles, that mouthparts and cirri of A. squalicola are highly reduced, and incapable of suspension-feeding activities. We show that in conventional suspension-feeding barnacles strong symmetries exist within these vital trophic structures. In A. squalicola strong asymmetries are widespread, indicating that those structures have been uncoupled from natural selection. The digestive tract is consistently empty, suggesting that feeding via cirri does not occur in A. squalicola. In addition, comparisons of stable isotope ratios (δ13C and δ15N) between A. squalicola, its shark host, and a conventional suspending feeding barnacle indicate that A. squalicola is taking nutrition directly from its host shark and not from the surrounding water. Our results strongly indicate that this barnacle has abandoned suspension feeding and now solely relies on obtaining nutrition from its host by a de novo evolved feeding mechanism.