Revision of Ilyphagus Chamberlin, 1919 (Polychaeta, Flabelligeridae)

Abstract Ilyphagus Chamberlin, 1919 includes abyssal, fragile benthic species. Most species have large cephalic cages but chaetae are brittle and easily lost which may explain why the original definition included species with a cephalic cage or without it. The type species, Ilyphagus bythincola Chamberlin, 1919, together with another species (Ilyphagus pluto Chamberlin, 1919) were described as lacking a cephalic cage whereas a third species (Ilyphagus ascendens Chamberlin, 1919) was described with one. To clarify this situation, all available type and non-type materials were studied. Ilyphagus is redefined to include species with digitiform bodies, abundant filiform papillae and a thin body wall; their neurochaetae are thick, anchylosed aristate spines, and all species have a cephalic cage (in the type species the presence of a cage is inferred from the remaining chaetal scars). Ilyphagus pluto, which also lacks a a cephalic cage is determined here to be a holothurian. The redefined genus contains Ilyphagus bythincola (incl. Ilyphagus ascendens), Ilyphagus coronatus Monro, 1939, Ilyphagus hirsutus Monro, 1937, and Ilyphagus wyvillei (McIntosh, 1885).


Introduction
Deep-sea animals are often bizarre. By the end of the XIX century, or during the early XX century, the dredged animals collected during deep-sea expeditions were surprising marine zoologists. Trying to cope with the unexpected body patterns, some general features might have been regarded as diagnostic for grouping species that could be more easily studied. This might explain why Chamberlin (1919) proposed a new name for some polychaetes that were regarded as feeding on mud. By combining the Greek words for mud (Ilys) and glutton (phagos), he established Ilyphagus. For its definition, he relied upon three different species: one with a well-developed cephalic cage, and two others apparently deprived of it. Thus, the generic diagnosis was wide enough to include species in either condition, and that concept prevailed in the current definition (Fauchald 1977). The cephalic cage is a distinctive feature for most flabelligerids. It is made up of long chaetae, usually pointing forward and stemming from at least one of the first few chaetigers. Although it has been a useful diagnostic feature to separate flabelligerid genera, it has been enigmatic how a single genus might contain two distinct morphological patterns; one with a cephalic cage and the other without it. Further, the body shape for Ilyphagus has been regarded as depressed or even disc-shaped (Fauchald 1977:117). This, in turn, is explained by the collapse of the body due to compression because the body is subcylindrical in life. Dredging or sieving sediments might distort the body shape, because the thin body wall is easily broken, such that the body becomes depressed, or flat. Monro (1937:305) made some interesting comments on the genus. He essentially indicated two issues: first, that Ilyphagus is an abyssal genus, and second, that there was a single species that included the three described by Chamberlin (I. bythincola, I. pluto, and I. ascendens), and his I. hirsutus. Monro was correct on the first statement, because all species do come from deep water habitats; for the second assertion, however, he failed to acknowledge the differences in papillae development and chaetae. Later, he indicated that all species in the genus were described as lacking a cephalic cage (Monro 1939:131), which was not the case, and that the cephalic cage chaetae may be present, but may be broken during collection. The latter is correct and will be discussed later.
Ilyphagus species are poorly known because of their life in deep-water and because they have not been found in hydrothermal vents, cold seeps or whale remains, which have recently received a lot of attention. Further, the species of Ilyphagus have very low abundances, most species being known from a single or few specimens. They might live in the sediment-water interface, with the body barely covered by sediment and the very long cephalic cage chaetae are probably used as an anti-predation device.
Because there are problems in the definition for the genus and for the species, the purpose of this revision was to study all material available and currently regarded as belonging in Ilyphagus. It was expected that this study would result in a redefinition of the genus, a clarification of the diagnostic features, and probably a modification of the included species. The presence of a cephalic cage is confirmed for the genus and after redefining it, it contains four species.

Methods
All specimens were studied under the stereomicroscope. They were often stained by a few seconds immersion in an oversaturated solution of methyl-green in 70% ethanol which is temporary. Individual chaetae or parapodial rami were observed in compound microscopes. The plates were made by selecting one or by editing several digital pictures of the same objects. The anterior end was dissected to study the head structure and the associated appendages. The materials belong to the following collections.

Museum and collections acronyms
LACM-AHF Museum of Natural History, Los Angeles, Allan Hancock Foundation Polychaete Collection.

NHML
The Natural History Museum, London.

SIORAS
Shirshov Institute of Oceanology, Russian Academy of Sciences, Moscow.

USNM
National Museum of Natural History, Smithsonian Institution, Washington.

Morphology
Body. The body is sausage-shaped; it may be short and oval, or cigar-shaped, and may even be swollen anteriorly. In some species, juveniles have thinner bodies and may be confused with some species of Bradabyssa Hartman, 1967; however, in Ilyphagus, the body wall is thin and covered by delicate, long papillae, which are not embedded by the tunic, resulting in a pilose surface. The papillae are filiform, barely swollen distally, if any at all, and are often covered by fine sediment particles. Once the excessive sediment is removed, individual papillae might have a thin or a thick layer of adherent sediment particles. Parapodia are poorly developed and the brittle chaetae emerge from the body wall. Gonopodial lobes are difficult to detect due to the abundant papillae; when they are well developed, they can be visible in chaetigers 5-6. Cephalic cage. The species belonging in Ilyphagus carry very long chaetae in the first 1-2 chaetigers. The first parapodia are markedly displaced dorsally, with both parapodial rami very close to each other and approaching the middorsal line. The second chaetiger has notopodia more dorsally displaced than the first but they are quite separated from the corresponding neuropodia. The second chaetiger's notochaetae are arranged in an oblique line or in a bundle, whereas the first notochaetae are more frontally located. Further, these notochaetae are often the longest, being as long as the whole body, or even longer; this chaetal length is remarkable among the family and among all Polychaeta. Sometimes, these long notochaetae can appear spirally twisted, as in the maldanid polychaete Nicomache maculata Arwidsson (Kennedy and Kryvi 1978). However, the chaetae in Ilyphagus, instead of being formed by spirally twisted fibers, have parallel microvilli and the external surface has a series of constrictions along the shaft which slowly rotate towards the tip. Nevertheless, these spiral chaetae are brittle and seldom available for observation, such that it is unknown how widespread they are, or if the finding detailed below is based upon accidental growth. All cephalic cage chaetae can be broken off, but their presence can be determined by chaetal scars or holes through the body wall, or by making a longitudinal dissection along the anterior end, if there are several specimens available and papillae cover chaetal scars. A large fan of companion chaetae is visible where the notochaetae of the first chaetiger are present.
Body chaetae. Notochaetae are multiarticulated, at least distally; there are usually 1-3 notochaetae per bundle and often one has long articles, whereas the others may have short articles only basally or throughout the chaeta. Neurochaetae are thicker than notochaetae; those present in the first chaetiger, or first few chaetigers, may be multiarticulated as well. In the former case, there are some chaetigers with transitional chaetae. They have long articles throughout the chaetae in the second chaetiger, but in following chaetigers the articulated region is progressively reduced, such that it becomes restricted to a short distal region. From chaetigers 2-4, anchylosed aristate spines are present. Neurochaetae have basal transverse marks, but these marks become slightly oblique beyond the median region. The distal region of the neurochaetae is delicate, often hyaline, and aristate; in one species, it is hirsute which is due to the rupture and exposure of the abundant oblique fibers .
Anterior end. The anterior end carries two large thick palps and several thick branchial filaments, arranged in 1-2 concentric rows which often resemble a horse-shoe pattern. The nephridial lobes are difficult to find, because they are located on the inner side of the branchial row at about the centre of the branchial plate. The species are apparently devoid of eyes and caruncle; the lack of a caruncle is unexpected because it has been found in all other flabelligerid genera. Although the underlying prostomial wall may be depressed, the ciliary bands may be present; however, because the preservation fluids penetrate slowly, these bands may be difficult to detect. Histological or SEM observations would be required to determine whether the expected bands of cilia are absent, or if they have been reduced. Remarks. The type species, I. bythincola Chamberlin, 1919, originally described as lacking cephalic cage, in reality has one but the chaetae are broken and only their embedded bases or some chaetal scars are left. The other species described as lacking a cephalic cage, I. pluto Chamberlin, 1919, is not a polychaete but an abyssal holothurian which belongs to the synallactid genus Meseres (identified by the late Dr. Cynthia Ahearn, USNM). The inclusion of I. pluto allowed a rather broad concept for body shape and chaetal patterns, because this species is rather cylindrical and completely lacks a cephalic cage. Further, Chamberlin (1919) emphasized the lack of large, wide papillae that resemble tubercles, as those found in some species of Brada. Thus, he recognized the difference and restricted the inclusion to those species lacking thick papillae (or tubercles).
Ilyphagus and Bradabyssa Hartman, 1967 are closely allied (Salazar-Vallejo et al. 2008) because they have anchylosed neurospines which are basally annulated (anchylosed short articles) and distally hyaline, tapering into an arista (hence aristate spines). However, there are four main differences between these genera. First, the relative position of the cephalic cage chaetae: they are arranged as transverse dorsal rows in Ilyphagus, whereas in Bradabyssa they are lateral, fewer and smaller. Second, the development of the body wall: most Bradabyssa species have a thick muscular body wall, whereas in Ilyphagus species it is reduced with poorly-developed muscle layers. Third, the branchial features: in Ilyphagus there are a few thick branchial filaments arranged in a horse-shoe pattern, whereas in Bradabyssa they are abundant and medially separated by the caruncle into two half-moon shaped groups. Fourth, and derived from the latter: the species of Bradabyssa have a well-developed caruncle whereas in Ilyphagus it is reduced or absent.
Species included. Besides the type species, I. bythincola Chamberlin, 1919 from the Eastern Pacific (including I. ascendens Chamberlin, 1919), the genus contains I. coronatus Monro, 1939 from the Antarctic Ocean, I. hirsutus Monro, 1937 from the Central Indian Ocean, and I. wyvillei (McIntosh, 1885) from the Antarctic Ocean.

Key to species of
Cephalic cage chaetae broken; scars present in chaetiger 1, short dorsal transverse row with 8-10 chaetal scars per side (Fig. 1C, D); holotype of I. ascendens with chaetae at least twice as long as body width, perhaps as long as body length (Fig. 1E). Chaetiger 1 (and perhaps 2) involved in the cephalic cage, dorsal; most cephalic cage chaetae broken from the base, 8 notochaetae; neurochaetae lateral, 7 per side, arranged in a short row, transverse (oblique in I. ascendens, Fig. 1G, H). Non-type specimens with chaetae as long as body length ( Fig. 2A). Anterior dorsal margin of first chaetiger papillated. Anterior chaetigers without larger papillae. Anterior end observed in a non-type specimen (SIORAS).
Cephalic hood short, not exposed, margin smooth. Prostomium low, as dark as surrounding region; eyes not seen. Caruncle not seen. Palps large, thick, shorter than branchiae; palp lobes reduced. Lateral lips well developed; ventral and dorsal lips reduced.
Branchiae thick, digitate, in different sizes, sessile on branchial plate, in a horseshoe pattern (Fig. 2B), one superior single row with four thick larger filaments, and six pairs of lateral filaments arranged in irregular double rows. Largest branchiae longer than palps. Nephridial lobes in branchial plate not seen. Chaetal transition from cephalic cage to body chaetae abrupt (most chaetae broken); first aristate neurospines in chaetiger 3. Gonopodial lobes not seen in holotype; non-type specimens with dark, digitate, small lobes in chaetiger 5 (Fig. 2C), or in chaetigers 5 and 6.
Remarks. The original body shape was digitate rather than sole-like; this distortion was the result of the sudden change of pressure, especially because of the sediment load in the dredge over its body. The damage resulted in the loss of all cephalic cage chaetae, but chaetal scars are visible in the corresponding position. This damage further compressed the body breaking its wall, and making it appear flat. Other specimens (SIORAS) are long, anteriorly swollen, posteriorly tapered; this is more pronounced among juveniles, and although the number of chaetigers is fixed early in development, counting depends on the presence of chaetae and they are often broken off.
Ilyphagus bythincola Chamberlin, 1919 andI. ascendens Chamberlin, 1919 are herein regarded as synonyms. The latter has cephalic cage chaetae in an oblique row, rather than in a transverse one as in I. bythincola. This displacement results in a larger area between chaetae and the anterior margin of chaetiger 1. However, because the anterior end of I. bythincola is severely damaged, and because other body features are similar, there are insufficientdifferences to keep them separate as distinct species.
Ilyphagus bythincola resembles I. hirsutus Monro, 1937, but they differ in the relative number of neurochaetae in chaetiger 1 and in the relative smoothness of neurochaetal tips. Thus, I. bythincola has about 8 neurochaetae in the first chaetiger, whereas there are 3-4 in I. hirsutus, and in the former, the neurochaetal tips are mostly smooth or barely hirsute, whereas in I. hirsutus neurochaetae are markedly hirsute. At the same time, I. bythincola resembles I. wyvillei (McIntosh, 1885), but they differ in the relative number of neurochaetae in chaetiger 1 and in the number of branchial filaments. Thus, in I. bythincola there are 8 neurochaetae per side and about 40 branchial filaments, whereas in I. wyvillei there are 10-12 neurochaetae and about 16 branchial filaments.
Anterior end not exposed; observed by dissection of anterior end in anterior fragment of syntypes, or other non-type specimens (USNM-56698, USNM-56696). Prostomium low cone, without eyes. No caruncle (Fig. 3D). Palps massive, as long as branchiae; palp lobes low. Lateral and dorsal lips fused; ventral lip reduced. Branchiae digitate, 14 filaments, arranged in three irregular rows: superior one with 4 filaments, two lateral groups medially placed each with 3 filaments, and two lateral basal ones with 2 filaments each; largest branchiae as long as palps.
Cephalic cage chaetae mostly broken; size relationships with body length or width unknown; syntypes with chaetae at least as long as body width and one with chaetae almost as long as body length; one very long chaetae straight, with successive constrictions but anchylosed articles perpendicular to the main shaft. Chaetiger 1 involved in cephalic cage; notochaetae of chaetiger 2 very long, thin. Chaetiger 1 with 8 notochaetae in transverse short dorsal row; neurochaetae in C-pattern, opening towards the posterior end, looking like two series, with 8 neurochaetae.
Posterior end (USNM-56696) rounded, pygidium with anus terminal, without cirri. Monro, 1939 can be separated from other cigar-shaped species because of the relative size of body papillae, which appear pilose, and because its neurochaetae, despite possibly appearing hirsute due to fracture, are mostly smooth.
Posterior end globose, damaged; pygidium with anus terminal, without anal cirri. Monro, 1937 resembles I. bythincola because of their dense coverage with fine, long papillae. They differ because I. hirsutus has a projected lobe in the first neuropodia, which is not present in I. bythincola, and because the neurochaetae of the former are markedly hirsute distally, whereas those in I. bythincola are either distally hyaline or slightly hirsute.

Remarks. Ilyphagus hirsutus
Distribution. Originally described from the Central Indian Ocean, in deep water (3385 m); it has not been recorded since.

Ilyphagus pluto Chamberlin, 1919
http://species-id.net/wiki/Ilyphagus_pluto Ilyphagus pluto Chamberlin, 1919:403;Hartman 1960:132. Remarks. This is a holothurian. The stout cylindrical processes forming a 'complete closed circle'from the original description are actually tentacles surrounding the mouth. Each tentacle is short and branched, but each branch is like a wart, making them apparently crenulated. The long, typical reddish brown chaetae found pen-  Monro, 1937. Holotype (NHML-1937 O'Loughlin (2002), two species have been described from the same region: M. torvus (Théel, 1886) and M. macdonaldi Ludwig, 1894; however, the former species has an uncertain generic placement, whereas the second is retained in its genus (O'Loughlin and Ahearn 2005).
Cephalic cage chaetae as long as half body length, or about as long as body width. Chaetigers 1-2 involved in the cephalic cage, chaetiger 1 with 8-9 notochaetae in a single transverse row and 11-12 neurochaetae arranged in a C-pattern, opening to posterior region; chaetiger 2 with 5-6 noto-and 9-10 neurochaetae.
Parapodia poorly developed, chaetae emerge from the body wall. Parapodia lateral, median neuropodia ventrolateral. Noto-and neuropodia close to each other, without especially longer papillae, some slightly thicker papillae bordering chaetae.
Posterior end rounded; pygidium with anus ventro-terminal, without anal cirri. Remarks. Ilyphagus wyvillei (McIntosh, 1885) resembles I. bythincola because they both have globose bodies with distally smooth aristate neurospines. They differ because I. wyvillei has comparatively shorter cephalic cage chaetae than I. bythincola, and because in I. wyvillei there are only 16 branchiae, whereas in I. bythincola there are about 40. On the other hand, I. wyvillei resembles I. coronatus Monro, 1939, because in their first chaetiger, neurochaetae are arranged in a C-pattern, opening to the posterior region, and by having distally smooth neurospines. However, they differ because I. wyvillei has fewer chaetigers (19-22 vs 23-25) and a more globose body but these differences might be modified after more specimens are studied. Two other differences are probably more relevant and must be emphasized:the relative number of neurochaetae in the first chaetiger (11-12 in I. wyvillei, 8 in I. coronatus) , and the start of the aristate neurospines (chaetiger 2 in I. wyvillei, chaetiger 3 in I. coronatus).
The presence of parasitic copepods in the branchial bases of I. wyvillei cannot be confirmed due to the state of the anterior end; however, one portion of a parasite is visible at one of the palps tip, and there is another deep scar in the same palp. McIntosh might have confused the attachment site, because he dissected the anterior end and branchial scars could be confused with these parasite attachment sites.
Brada gravieri McIntosh, 1922 might belong to the same species. There is no type material available; it is probably lost. However, the original illustrations and description noticed the lack of the cephalic cage chaetae, whereas the neurochaetae (pp 7-8) were described as translucent, smooth, devoid of transversal marks. The distal part of neurochaetae is often smooth, hyaline, but the rest of the chaetae have anchylosed articles or transverse markings throughout it. They were collected from relatively close localities but fresh material needs to be examined to clarify this .
Distribution. Originally described from the Antarctic Ocean, it has been found in abyssal depths off Western South America (Levenstein, 1975). The Bering Sea records by Levenstein (1961aLevenstein ( :160, 1966, cannot be confirmed because the specimens were not found. Ahearn (USNM) kindly confirmed that what has been regarded as a flabelligerid species was actually a holothurian. The fast and kind help by José Santos Gómez, our librarian in Chetumal, was fundamental to find some uncommon publications. The careful reading by Pat Hutchings, Teresa Darbyshire and Chris Glasby helped a lot for improving this final version. Additional funds came from the U.S.A. Department of the Interior, Fulbright Commission, MNHN, and from the Université Catholique de l´Ouest, Angers, France.