The photograph, below, shows the transparent internal base (left arrow) of the filament in a damaged trawl-captured vampire.
Figure. The heavily pigmented external base (middle arrow) of the filament of V. infernalis extends from the pocket between adjacent arms, and the beginning of the slender, white, distal portion of the filament (right arrow) follows. Most of the filament was broken off during capture. Much of our understanding of the phylogenetic position of the vampire depends on determining the homology of the retractile filaments. Photograph by R. Young.
Figure. The insitu photograph above of V. infernalis was taken by an ROV of the Monterey Bay Aquarium Research Institute. The filaments are withdrawn but the filament pocket is visible between the bases of arms I and II. © MBARI 2001, http://www.mbari.org/rd/midwater.
The primitive (plesiomorphic) arrangement of arms in coleoid cephalopods is thought to be 10 equal arms. This arrangement is known in some fossil coleoids (belemnoids) and the presence of ten unequal arms in modern decapods is easily derived from such a condition. Therefore, octopods, with eight arms, have apparently lost one pair. If the vampire filaments represent modified arms II and if this is the pair that is lacking in octopods, then strong support would exist for a vampire-octopod affinity. Embryological evidence suggests that the missing arm pair in octopods is either arms II or III (Boletzky, 1978-79).
Pickford (e.g., 1940) concluded that the filaments were modified arms II based on their position in series with the arms, venous connections and innervation. Young (1967) clarified the innervation and argued that the filaments were not arms but homologues of the preocular tentacles of Nautilus. The axial nerves of arms send most of their fibers to the anterior subesophageal mass (brachial lobe) of the brain. Young found that the filament nerves bypass this lobe with efferent fibers arising from the middle subesophageal mass and the afferent fibers reaching an enlarged ventral magnocellular lobe. Dilly, et al. (1977), however, demonstated that large nerve bundles from the tentacles of the deep-sea squid Mastigoteuthis also pass to an enlarged ventral magnocellular lobe. Presumably the combined sensory function of thousands of suckers carried by the tentacles of Mastigoteuthis requires the large nerve bundle and the enlarged ventral magnocellular lobe. Similarly the sensory function of the filaments of Vampyroteuthis probably emphasized the connection to and size of the ventral magnocellular lobe while a loss of coordination with the principal arms resulted in a great reduction (at least) in the connection to the anterior subesophageal mass. As a result, the innervation of the filaments became a weak argument against filament-arm homology.
The homology appeared to be resolved when J. Z. Young (1977) reported a connection between the axial nerve of the vampire filament and ganglia of the circumoral commissure that connects the axial nerves of all arms. This convincing evidence, however, could not be confirmed by Young and Vecchione (1996).
Recent confirmation of Pickford's earlier evidence (Pickford, 1946) that the vampire hatchling has relatively thick (i.e., more arm-like) filaments supports the possible homology between arms and filaments (Young and Vecchione, in press). The weight of evidence indicates that the filaments represent arms II. Proof will depend on examining vampire embryonic development which is not feasible at present.