Fan worms (Annelida: Sabellidae) from Indonesia collected by the Snellius II Expedition (1984) with descriptions of three new species and tube microstructure

The Indonesian archipelago is one of the most diverse regions in the marine World. Many contributions on polychaete worms have been published since the Dutch Siboga Expedition to the Indonesian archipelago at the end of the 19th century. In this study, we examined specimens of Sabellidae Latreille, 1825 collected during the Snellius II Expedition (1984) to Indonesia, carried out by the Dutch Research Vessel (RV) “Tyro” and the Indonesian RV “Samudera”. The results include reports of Acromegalomma acrophthalmos, A. interruptum, A. sp., Bispira manicata, B. porifera, B. secusoluta, Branchiomma boholense, Notaulax pyrrohogaster, N. tenuitorques, N. sp. 3, Parasabella crassichaetae, Perkinsiana anodina, and Sabellastarte spectabilis. In addition, three new species are described: Acromegalomma sumbense sp. nov., Claviramus olivager sp. nov., and Notaulax montiporicola sp. nov., the latter in living coral (Montipora nodosa). Further, Sabella (Potamilla) polyophthalmos Grube is transferred to Pseudopotamilla. Additional histological accounts of B. porifera and tube microstructure of A. acrophthalmos, B. porifera, P. anodina, Pseudopotamilla polyophthalmos and Sabellastarte spectabilis are also included.


INTRODUCTION
The Indonesian archipelago, the South China and the Philippine Seas are among the most diverse regions in the Western Pacific. Many contributions about polychaete worms have been published since the Dutch Siboga Expedition to the Indonesian archipelago at the end of the 19th century. A compilation of these Siboga reports can be found in Aguado, San Martín & ten Hove (2008) and .

Species name
Type locality Synonymies and current name Myxicola ommatophora Grube, 1878 Philippines Original name currently valid Sabella acrophthalmos Grube, 1878 Philippines -Megalomma acrophthalmos fide Hartman, 1959;Knight-Jones, 1997; Tovar-Hernández & Carrera-Parra, 2011 -Acromegalomma acrophthalmos fide  in some species. Shirlastain A was helpful in analysis of the main morphological features in some species. The distribution map for Acromegalomma was produced using SimpleMappr (Shorthouse, 2010). Species properly illustrated, as indicated per species, were not figured again.
The tube microstructure of Acromegalomma acrophthalmos (Grube, 1878), Bispira porifera (Grube, 1878), Perkinsiana anodina , Pseudopotamilla polyophthalmos (Grube, 1878) and Sabellastarte spectabilis (Grube, 1878) was studied using a scanning electron microscope. Ethanol preserved tubes were first air-dried and longitudinally as well as transversely cut with a razor blade. Samples were then coated with gold, and studied under high vacuum conditions with the environmental scanning electron microscope (ESEM) Philips XL30, at the Faculty of Earth Sciences, Sosnowiec, Poland. A single longitudinal and transverse section of each tube was studied. In addition the tube wall structure adjacent to the lumen of a single tube of each species was studied. The studied tubes were deposited at the Faculty of Earth Sciences in Sosnowiec, Poland.

Nomenclatural acts
The electronic version of this article in Portable Document Format (PDF) will represent a published work according to the International Commission on Zoological Nomenclature (ICZN), and hence the new names contained in the electronic version are effectively published under that Code from the electronic edition alone. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix http://zoobank.org/. The LSID for this publication is: [urn:lsid: zoobank.org:pub:382D313F-0138-4194-B3B2-5BA84817374A]. The online version of this work is archived and available from the following digital repositories: PeerJ, PubMed Central and CLOCKSS.
Description. Large specimen, trunk~60 mm long, 0.8 mm wide. Twenty two pairs of radioles. Subdistal eyes in most radioles (spherical and spiral). Anterior peristomial ring exposed between dorsal pockets. Posterior peristomial ring collar: dorsal margins fused to the faecal groove. Caruncle present. Dorsal lappets present. Inferior thoracic chaetae Type B (with progressively tapering distal tip). Interramal eyespots absent. Tube attached to large basal stone and composed of shell fragments, coralline sand and small stones. Maximum outer diameter of tube: 10 mm.
Tube microstructure. Tube's lumen surface nearly smooth with few small bumps with sub-circular outline, without any regular pattern. Tube wall lamellar; lamellae thin, about 2-3 mm thick, straight in cross section. Lamellae composed of moderately developed, straight, long, thin and parallel fibers with relatively constant diameter about Remarks. As usual in the 19th century, Grube did not specifically mark his specimens as types of any kind. On the basis of the fact that Grube (1878: 258) explicitly states that he had only one specimen, Wiktor (loc. cit.) identified it as the holotype. A. miyukiae (Nishi, 1998) First to 5th dorsalmost pairs of radioles Not fused to faecal groove Absent Absent Absent Absent Well exposed Type A -Thailand, Andaman Sea A. multioculatum (Fitzhugh, 2002) On most radioles Fused to faecal groove Absent ? Absent Present Well exposed Among the known species from the Indonesian archipelago, Australia and the South China and Philippine Seas (Table 2), only two have a caruncle: A. jubatum (Capa & Murray, 2015a) and A. acrophthalmos. The caruncle in some species of Acromegalomma was documented by Tovar-Hernández & Salazar-Vallejo (2008: figs 2A, E-F, 3-5). Externally, it resembles the caruncle in other polychaete families such as Amphinomidae and Spionidae. Internally, it is an organ innervated directly from the cerebral ganglion, supported with hyaline cartilage, homologous to the median organ of sabellariids  fig. 10B) described a smooth structure -not homologous to the caruncle-, called keel. The keel is a smooth projection of the peristomium arising between the dorsal lips, forming a ventrally-directed ridge.
Major differences between A. acrophthalmos and A. jubatum are the following: eyes are present in most radioles in A. acrophthalmos (only in first six pairs of radioles in A. jubatum); long dorsal lappets in A. acrophthalmos, extending beyond collar margins (dorsal lappets shorter than collar margins in A. jubatum); and caruncle as long as collar length in A. acrophthalmos (only half as long as collar in A. jubatum).
Although A. acrophthalmos is one of the first sabellid species described from Indonesia, is about 50 mm long, and lives in the intertidal, there are scarce records of its presence in the region. Acromegalomma acrophthalmos is known from its type locality (Philippines); Ambon ; Negros Island, Philippines (Tovar-Hernández & Carrera-Parra, 2011); and Ambon Bay, Maluku, Indonesia (present study).
Capa & Murray (2009) reported one specimen from Dampier Archipelago (Western Australia) as Megalomma cf. acrophthalmos, presenting a low, smooth keel. Later, Capa & Murray (2015a) reported another specimen from Lizard Island (Eastern Australia) (also as Megalomma cf. acrophthalmos but no specimen of Megalomma cf. acrophthalmos sensu Capa & Murray (2009)), having a caruncle, but the distal end of radioles were regenerating, eyes could thus not be studied. Type material of Acromegalomma acrophthalmos examined by Knight-Jones (1997: 316), and discussed in relation to their 20 specimens from the Philippines by Tovar-Hernández & Carrera-Parra (2011), as well as our specimen from Indonesia all have a caruncle. Detailed illustrations of morphological features of A. acrophthalmos can be found in Knight-Jones (1997) and Tovar-Hernández & Carrera-Parra (2011).
Remarks. Among the species of Acromegalomma reported in the Indian Ocean, only two have dorsal collar margins not fused to faecal groove (Table 2): A. interruptum and A. miyukiae. Both species can be distinguished by the presence of shallow dorsal pockets in A. interruptum (absent in A. miyukiae) and eyes in dorsalmost and lateral radioles in A. interruptum (eyes in first dorsalmost pair of radioles in A. miyukiae).
Description. Trunk 17.5 mm long, 3.5 mm wide. Radiolar crown 7.3 mm long. Fifteen pairs of radioles. Subdistal eyes present in dorsal and lateral radioles (large, spherical, surrounding the tip of dorsalmost pair of radioles; small, spherical, similar in size in lateral radioles). Anterior peristomial ring exposed partially on dorsal side. Posterior peristomial ring collar with dorsal margins fused to faecal groove; dorso-lateral margins with V-shaped notches. Dorsal pockets present, shallow. Keel present. Thoracic tori not contacting shields on anterior chaetigers. Ventral lappets rounded. Ventral sacs present. Inferior thoracic notochaetae Type B (with progressively tapering tips). Interramal eyespots absent. Tube not preserved. Body colour preserved only in dorsal thorax: brown coloured with residual dark spots located on the ventral margin of the thoracic tori.
Tubes: composed of shell fragments and coralline sand (Fig. 3A). Sex and gametes: Unknown. Remarks. Acromegalomma sumbense sp. nov., resembles the specimens reported as Megalomma sp. 2 by Capa & Murray (2009) from Victoria (Australia) and referred herein to Acromegalomma sp. 2, A. kaikourense (Knight-Jones, 1997), described from New Zealand, and A. sp. cf. kaikourense (Capa & Murray, 2015a), from Lizard Island (Australia). All three taxa have dorsal collar margins not fused to the faecal groove, and eyes only in the dorsalmost radiolar pair. However, A. sumbense sp. nov., specimens differ from those taxa by having a caruncle (absent in the others) and radiolar tips as long as three times the diameter (radiolar tips not extending beyond distal margins of eyes in   (Table 2).
Etymology. The specific name is an adjective derived from Sumba, the type locality.
Tubes of Indonesian specimens are composed from brown dark sand and architomy is present in some specimens. It includes worms undergoing pre-fission or post-fission. In reproducing worms prior to fission, the posterior abdomen was cream coloured, tapering abruptly towards the posterior end (Fig. 5A). Buds or fragments separated from a parental worm (post-fission) showed incomplete regeneration: rudimentary crowns and -as yet-only abdominal segments (Figs. 5A-5C). The architomy in B. manicata is similar to that described for the Caribbean B. brunnea (Dávila-Jiménez, Tovar-Hernández & Simões, 2017).
Bispira secusoluta (Hoagland, 1920)  Description. Trunk 7 mm long (incomplete specimen), 0.8 mm wide. Radiolar crown 6 mm long (twice longer than thorax) with six pairs of radioles. Radiolar eyes absent. Narrow flanges along radiolar length (Fig. 5F). Palmate membrane 1/4 as long as radiolar crown, or as three thoracic segments. Dorsal lips as long as palmate membrane. Rounded lobes medial to dorsal lips. Posterior peristomial ring collar with dorsal margins not fused to faecal groove (Fig. 5E), lateral margins notched. Ventral lappets long, triangular, with small, triangular-like extensions overlapping at midline (Fig. 5D). Ventral shield of collar with two large, lung-shaped ciliated areas (Fig. 5D). Thorax with eight and abdomen with 21 chaetigers. Interramal eyespots present along thorax and abdomen. Thoracic and abdominal shields with a pair of oval to rounded patches of cilia per shield. Tube not preserved.
Remarks. This species was described from Sombrero Islands (Philippines) by Hoagland (1920) as Sabella secusolutus, then transferred to Bispira by Knight-Jones & Perkins (1998). The word solutus is a Latin adjective, Sabella is feminine (as is Bispira) and the species name should have been secusoluta -with feminine ending, ICZN, 1999, Art.34-and is corrected here. The record of Bispira secusoluta in Sanggar Bay is the first record of the species since its description in 1920.
Bispira secusoluta and four of its congeners lack radiolar eyes: B. brunnea, B. porifera, B. wireni (Johansson, 1922) andB. oatesiana (Benham, 1927). Bispira brunnea and B. secusoluta have ventral lappets with triangular extensions overlapping at midline (without such extensions in B. porifera and B. oatesiana). Bispira secusoluta differs from B. brunnea in having paired, large, lung-shaped ciliary pads on the collar segment and small, rounded patches of cilia in the shields of thoracic and abdominal segments (elliptic ciliary pads, and lacking patches of cilia in thoracic and abdominal segments in B. brunnea). The presence of paired patches of cilia in body segments has been only reported in Pseudobranchiomma schizogenica by Tovar-Hernández & Dean (2014), but probably these structures have been overlooked and are present in many species.
Among the currently known species of Bispira in the Indian Ocean, B. secusoluta and B. porifera have no radiolar eyes (Table 3). Bispira porifera can be distinguished by the presence of dorsal spongy cushions, absent in B. secusoluta. Description. Trunk 46 mm long, 8 mm wide. Radioles without eyes. Anterior peristomial ring purple, forming two rounded, prominent lobes (Figs. 6A-6C). Posterior peristomial ring collar with dorsal margins not fused to faecal groove. Two pairs of cushion-like masses dorsally, separated by mid-dorsal faecal groove: anterior pair purple, granular, extending around peristome (Figs. 6A and 6C); posterior pair pale, spongy, with many oval to circular-shaped pores, unequal in diameter (Figs. 6A and 6C), extended to third thoracic chaetiger. Ventral lappets of collar prominent, triangular, separated. Ventral sacs present outside radiolar crown, purple (Fig. 6B). Thorax with eight chaetigers. Narrow mat of yellow glandular tissue visible on dorsal side of posterior thoracic segments, and four anterior abdominal segments. Tori of chaetigers 1-5 in contact with ventral shields (tori occupy the entire distance between notopodia and ventral shield margin), tori in chaetigers 6-8 not contacting shields. Abdomen with 173 chaetigers. Interramal eyespots Histology. The anterior pair of dorsal cushion-like masses, located on the anterior peristomial ring (Figs. 6A-6C), is a strongly innervated area (Figs. 6D-6G). Posterior pair of dorsal cushion-like masses is composed of a sinuous, wide glandular epithelium, 500 mm thick. Neural packages and glandular epithelium are surrounded by a dense cartilaginous matrix and muscular fibers, which run along thorax (Figs. 6D-6H).
Tube microstructure. The tube lumen relatively smooth, but in places with wrinkles, elongate pits and somewhat circular bumps. Two sets of fibers in alternate orientations at about 35-40 to each other. Fibers in single set moderately-developed, but long (>15 mm), usually straight and parallel to each other, but some fibers slightly undulating or curved. Some fibers with widened drop-like nodes at irregular intervals. Fibers 0.15-0.30 mm wide, interspaces usually wider than diameter of fibers. Interspaces of adjacent fibers filled with homogeneous organics with smooth surface. Tube wall dense, non-porous, except for rare elongate pits (1-4 mm long) oriented parallel to the fibers (Figs. 7A and 7B).
Spongy dorsal cushions have been reported only in two species of Bispira: B. porifera (Grube, 1878) and B. paraporifera Tovar-Hernández & Salazar Vallejo, 2006. Both species have been found associated with dead coral blocks, but there is no information about structure and function of this peculiar tissue. Willey (1905) correctly suggested that spongy cushions of B. porifera are glandular, but this only applies to the second pair of cushions, consisting of glandular columnar epithelium. The anterior pair of cushions, located above peristomium, is not glandular, but strongly innervated, not unexpectedly so since 19/72 the brain ganglion is located in that area. Types of glandular cells were not determined, but their function might be associated with mucous secretion for tube construction or re-construction of damaged parts. Knight-Jones & Perkins (1998) suggested that this tissue might have a function in embryo incubation, but there is no evidence to support this hypothesis. Information of reproduction in B. porifera is null, but its congener B. brunnea (Treadwell, 1917), is a Caribbean broadcast species with sperm morphology adapted to external fertilization in the water column (Dávila-Jiménez, Tovar-Hernández & Simões, 2017).
Branchiomma is unique by the presence of paired stylodes (epithelial flaps) along the outer surface of the radiolar axes, an autapomorphy . Proper identification is particularly challenging for this genus because there is morphological variation in taxonomically informative characters at species level (Keppel, Tovar-Hernández & Ruiz, 2018). As a result, the nomenclature of the genus is in a state of flux and it is currently under review using molecular identification techniques (del Pasqua et al., 2018;Belato et al., 2018). Branchiomma is among the most visible polychaetes of the hard substrate fouling communities, and several species have been reported outside their naturally expected distribution ranges (Tovar-Hernández, Méndez & Salgado- Barragán, 2009;Keppel, Tovar-Hernández & Ruiz, 2015). The most recent account of alien Branchiomma includes eight species (Keppel, Tovar-Hernández & Ruiz, 2015;Keppel, Tovar-Hernández & Ruiz, 2018). Cases of high phenotypic plasticity in taxa from Australia and Mediterranean Sea, and probably all around the world, high infraspecific genetic variability, cryptic species and unexpected translocations of species beyond original distributions were documented by Capa, Pons & Hutchings (2013) and del Pasqua et al. (2018).
Remarks. Branchiomma boholense was originally described from Bohol islands (Philippines). It has been reported from the Eastern Mediterranean (Malta, Atlit, Tel-Aviv and Alexandria) by Knight-Jones, Knight-Jones & Ergen (1991); Cyprus by Çinar (2005) but then corrected to B. bairdi (Çinar, 2009) (2017), except for the presence of paired basal stylodes and short, blunt radiolar tips, occupying the space of 3-5 pinnules. Specimens examined here, all juvenile and adult stages, present unpaired basal stylodes and long, filiform radiolar tips (occupying the space of approximately 10-13 pinnules in adults, 5-6 in juveniles). Capa, Pons & Hutchings (2013) already emphasized the presence of basal stylodes being paired in some few taxa (B. lucullanum, B. bombyx and B. galei) and showing plasticity in the majority of the studied species: specimens within the same species were found with single or paired basal stylodes. However, a fixed relation of presence-absence of this feature with juvenile or adult stages, or regenerating worms has not been observed.
Two species of Branchiomma were described by Grube from the Central Indo-Pacific: B. cingulatum (Grube, 1871) from Fiji and B. boholense (Grube, 1878) from Bohol Islands, Philippines. Both can be distinguished by the absence of macrostylodes in B. cingulatum and the presence of tongue-like macrostylodes in B. boholense. Branchiomma cingulatum was also reported from Ambon by Augener (1933, as Dasychone).
Number of species: 5, including one new species described below. Radioles in semi-circular radiolar lobes, each radiole with two rows of vacuolated cells. Palmate membrane absent; radiolar tips with foliaceous flanges, expanded or curled; basal flanges absent; radiolar eyes absent. Dorsal lips with radiolar appendages, pinnular appendages apparently absent; ventral radiolar appendages present, few to several pairs. Ventral lips present, ventral sacs and parallel lamellae absent. Anterior peristomial ring with broad, triangular, ventral lobe. Posterior peristomial ring collar with wide mid-dorsal gap, mid-ventral incision and ventral lappets. Peristomial vascular loops present in some species. Peristomial eyespots may be present. Glandular ridge on chaetiger 2 present or not. Ventral shields present on thorax; present or absent on abdomen. Interramal eyespots absent. Collar chaetae similar to superior notochaetae of following chaetigers, elongate, narrowly hooded; inferior thoracic notochaetae broadly hooded, or narrowly and broadly hooded. Thoracic uncini acicular, with short teeth above main fang arranged on transverse rows, hood present, handle long, main fang may be bifid. Neuropodial companion chaetae absent. Abdominal uncini avicular, with distinctly short handle, developed squared to rectangular breast and several transverse rows of short teeth above main fang. Abdominal neurochaetae in single group of narrowly hooded chaetae. Pygidium with eyespots present in at least some species. Anal cirrus absent.
Remarks. Claviramus Fitzhugh, 2002 currently includes five species worldwide. Claviramus candelus (Grube, 1863), the type species of the genus, was originally described as Sabella candela Grube, 1863, from the Northern Adriatic Sea. Langerhans (1884) transferred the species to Jasmineira Langerhans, 1880, and described a new species J. oculata Langerhans, 1884, from Madeira. Cochrane (2000) redescribed those species within Jasmineira in detail, based on type and additional specimens. Fitzhugh established the genus Claviramus, based on the presence of radiolar tips with foliaceous flanges (Fitzhugh, 2002: fig. 43) and transferred J. candela and J. oculata to Claviramus. The third species, C. grubei Fitzhugh, 2002, was described from Thailand, Andaman Sea. Claviramus kyushuensis Nishi, Tanaka & Tovar-Hernández, 2019 was described from Japan and in the present study, the fifth species is described below from Indonesia.
The generic diagnosis above primarily follows Fitzhugh (2002). The emendation provided here is to include the presence of (1) peristomial vascular loops present in Claviramus kyushuensis from Japan (Nishi, Tanaka & Tovar-Hernández, 2019) and the new species described below from Bay of Sanggar, (2) the presence of bifid main fangs of thoracic uncini reported in C. kyushuensis, and (3) presence of abdominal shields in C. kyushuensis as well as the new species described below.
Remarks. Claviramus olivager sp. nov., is a new species based on a set of unique, distinctive features: absence of glandular ridge on chaetiger 2, distal radiolar flanges without filament, and a mid-ventral incision extending for half flange length, abdominal shields well developed, and ventral lobe of anterior peristomial ring not extending beyond collar margins (Table 4).
Claviramus candelus and C. olivager sp. nov., have glandular abdominal shields and both lack a glandular ridge on chaetiger 2, but ventral margin of collar is entire, of a constant height in C. candelus, whereas it is incised, forming two slightly elevated lappets in C. olivager sp. nov. Claviramus grubei and C. oculatus have a distal filament (cirrus) on their radiolar tips (absent in C. olivager sp. nov.), and the ventral lobe of anterior peristomial ring is broadly triangular, extending slightly beyond collar margin (not exposed in C. olivager sp. nov.). Both taxa have a short mid-ventral incision of distal radiolar flanges, less than 1/4 as long as flange length while it is 1/2 of flange length in C. olivager sp. nov. (Table 4).
Nishi, Tanaka & Tovar-Hernández (2019) described a new species of Claviramus from Ariake sound (Japan): C. kyushuensis. In our study, we had opportunity to examine three of their paratypes. These specimens are hermaphrodites with oocytes and sperm in thoracic and abdominal segments. The holotype of Claviramus olivager sp. nov., is a female with oocytes in thorax and abdomen, but an exhaustive search of spermatozoa was not assessed, so we cannot exclude simultaneous hermaphroditism. Claviramus olivager sp. nov., differs from C. kyushuensis from Japan in lacking a glandular ridge on chaetiger 2 (present in C. kyushuensis) and having tips of main fang of thoracic uncini entire (bifid in C. kyushuensis).
Distal flanges are very fragile and easily broken off from radioles during manipulation. Cochrane (2000) also showed broken radioles in some specimens belonging to C. candelus (as Jasmineira). Under this scenario, it is feasible that more specimens still erroneously can be found under Jasmineira. However, Jasmineira and Claviramus might also be distinguished on the basis of the presence of inferior thoracic bayonet notochaetae (absent in Claviramus), uncinial morphology and presence of a breaking plane sensu Cochrane (2003) or abscission zone sensu Tovar-Hernández (2008).
Etymology. The specific name refers to the shape of radiolar tips, that resembles the elongate oval-shaped shells of the mollusk genus Oliva Bruguière, 1789.
Remarks. Perkins (1984) revised the genus, described new species, provided several synonyms, and proposed new combinations of species within Notaulax. Later, a revisory contribution of members of the genus from Australia has been conducted by Capa & Murray (2015a); a second, from Mexico is in process by Tovar-Hernández. Diagnoses to genus level are available in Perkins (1984),  and . The generic diagnosis was emended in order to include: (1) the presence of abdominal interramal eyespots (as reported in Tovar-Hernández, de León-González & Bybee (2017)); (2) the presence of a high, peristomial ring as seen in N. pyrrohogaster and N. tenuitorques (see below) and also in N. bahamensis Perkins, 1984 andN. nudicollis (Krøyer, 1856) (Perkins, 1984: figs 25E-F, 35E-F). (3) The difference between chaetae from anterior and posterior abdominal segments.
Seven species of Notaulax are known to occur embedded in coral masses , one was found associated with a fossil reef; other species have been found fouling in marinas and ports (Tovar-Hernández, de León-González & Bybee, 2017), for others there is no information on the substrates from which they were collected .
Two species of Notaulax from Australia went unnoticed as such in previous papers: N. velata (Haswell, 1885, as Sabella) from Port Jackson (Sydney) and N. longithoracalis (Hartmann-Schröder, 1980, as Hypsicomus) from Port Samson (Western Australia). Based on their original descriptions and drawings, both species have chaetae, collar shape and radiolar ocelli typical of Notaulax. Thus, N. longithoracalis was included in Table 5, whereas N. velata was left out because its original description is incomplete in critical features for species comparisons.
Capa & Murray (2015a) give the length of radiolar tips as short (Notaulax sp. 1) or medium (as long as a pinnule: Notaulax sp. 2, sp. 3). However, it is unknown if the pinnules considered by Capa & Murray (2015a) were distal, median or basal. As the length of pinnules in Notaulax increases from the radiolar base to the median radiolar zone, and then decreases gradually towards the tip, in our study the length of bare radiolar tips is expressed as the corresponding space of a given number of pinnules.
Remarks. de Saint-Joseph (1894) included Sabella pyrrhogaster in his new genus Eurato, and Bush (1905) subsequently designated it as the type species of the genus. According to Hartman (1959: 546) Eurato is a subjective synonym of Hypsicomus Grube, 1870. Hypsicomus has two pairs of accessory lamellae in the posterior peristomial ring, between the dorsal collar margins (absent in Notaulax), and collar chaetae in a typical short bundle (collar fascicles longitudinal to oblique in Notaulax). Based on these main differences, Perkins (1984) attributed S. pyrrhogaster to the genus Notaulax.
Grube (1878) stated specifically that ocelli were absent in Notaulax pyrrhogaster (as Sabella). His description indicates the largest specimen had damaged radiolar lobes. According to Perkins (1984) it is likely that radiolar ocelli were sloughed off with epidermal tissue, or faded. Our specimens have 28-30 radiolar ocelli per row, first appearing above mid-radiole length and continuing to distal pinnules. Basal radiolar ocelli are widely spaced out, whereas distal ocelli are very close to each other. The ventral and dorsal margins of the collar are as illustrated by Grube (1878: pl. XV, figs 1, 1a).
Notaulax pyrrhogaster and N. tenuitorques were both described from Bohol, Philippines. These species differ by the shape of the ventral margin of the collar, the arrangement of collar chaetae, and distribution of radiolar ocelli. The ventral margin of the collar is long, triangular in N. pyrrhogaster (low, rounded in N. tenuitorques). Collar chaetae are arranged in slightly curved rows in N. pyrrhogaster (in straight oblique rows in N. tenuitorques). Notaulax pyrrhogaster has radiolar ocelli in single rows from mid-radiole length to the distal pinnules (ocelli in groups of 15-17 ocelli at three quarters of radiole length, then one row of ocelli distally in N. tenuitorques).
Notaulax pyrrhogaster differs from other species from Australia and Japan in having a single row of 28-30 ocelli (from the mid-radiole length to distal pinnules), and a long, entire, triangular ventral collar margin (Table 5).
Remarks. As explained above, Grube did not specifically mark his specimens as types of any kind. Grube (1878: 258) explicitly states that he had only one specimen. Wiktor (loc. cit.) marked the specimen known to her as holotype. However, 13 years later Hartwich (1993: 142) found a second specimen in the private collection of Grube, bought in 1881 by the Zoological Museum Berlin, and consequently labelled by Hartwich as ? syntype. A full evaluation of the material, syntype, holotype, or whatever, only can be given on the basis of the real material, in the context of a full taxonomic revision, not the intention of the present paper.
The original description by Grube (1878) emphasizes the remarkable long radiolar lobes, the low collar and the presence of radiolar ocelli in S. tenuitorques, as compared with two other species of Potamilla. Hartman (1959) placed S. (Potamilla) tenuitorques in Potamilla. However, Grube illustrated (Pl. 14: fig. 2) the typical crown, collar and chaetae of what nowadays is regarded as present in Notaulax. Our study corroborates the transfer of Sabella (Potamilla) tenuitorques to Notaulax, by the presence of long radiolar lobes, with dorsal and ventral flanges; and collar chaetae fascicles as single, elongate, oblique row of spine-like chaetae. Our description permits to distinguish N. tenuitorques from Australian, Philippine and Japanese congeners (Table 5). Among these, only N. pyrrhogaster and N. sp. 3 Capa & Murray (2015a) have radiolar ocelli distributed in groups, but in the first species the distal ocelli are distributed in a single row. In N. pyrrhogaster, N. longithoracalis and N. sp. 3 the collar fascicle is straight, oblique. However, the second species has only 4-6 radiolar ocelli in a single row, N. pyrrhogaster have 28-30 ocelli in groups basally and then distributed in a single row distally, whereas N. sp. 3 have 20 ocelli in groups at the middle of radiole only (Table 5). Notaulax tenuitorques and N. pyrrhogaster were found within the same locality. A detailed comparison between the latter species is given in the remarks for N. pyrrhogaster.
Abdominal shields divided into two areas by faecal grove, which does not take stain (Fig. 13F). Entire dorsum and lateral sides of body pale, unstained (Figs. 13D-13F).
Remarks. This peculiar new species found in the living coral Montipora nodosa, at 30 m depth on the Western coast of Binongko Island, in the Banda Sea, is the first record of a sabellid polychaete inhabiting a living Montipora de Blainville, 1830. Nishi et al. (2017) provide a good synthesis of sabellids inhabiting corals. The specimens here reviewed were without the coral from which they were extracted, but labeled as "found in living coral". The relation between corals and sabellids remains unknown, but this kind of interaction corresponds to bioclaustration. Species of Anamobaea, Hypsicomus, Notaulax, Perkinsiana and Pseudopotamilla have been reported in coral masses either, but if these were alive or dead, or if there was a real symbiosis is unclear (Perkins, 1984;Tovar-Hernández & Salazar Vallejo, 2006;Nishi et al., 2017). A case of a sabellid modifying a coral surface is reported below for Perkinsiana anodina (see below). Notaulax montiporicola sp. nov., is similar to most species in the genus in that the base of the radiolar crown is long, equipped with dorsal and ventral flanges, and chaetal fascicles have a straight, longitudinal row of spine-like chaetae. However, N. montiporicola sp. nov., does not have radiolar ocelli (present in Anamobaea Krøyer, 1856 and Notaulax); it is unknown if such ocelli may fade off after years of preservation, although the peristomial eyes (not reported in Notaulax neither in Anamobaea) have not faded over the years. In addition, Notaulax montiporicola sp. nov., is unique by the presence of two remarkable types of mucros in the abdominal paleate chaetae: with candle flame-shaped mucros in anterior abdominal segments, sail-shaped mucros posteriorly. Unfortunately, only three specimens were collected and the use of scanning electron microscopy for a better documentation of these changing chaetal forms was not possible. In addition, the new taxon show characters that do not match entirely with Anamobaea or Notaulax such as the absence of radiolar eyes, palmate membrane, radiolar flanges, and ventral sacs. A full phylogenetic analysis, not possible in the context of the present paper, is needed to decide whether or not these characters merit a genetic distinction.

Remarks. Diagnoses to genus level are available in
Notwithstanding the fact that Capa & Murray (2015b) examined many specimens, including some types, they were unable to attribute all to a previously described or an evident new species. The taxonomic study of Parasabella species is not simple because many of the original descriptions are brief, incomplete, or not illustrated; further, records are doubtful because unique morphological diagnostic features are lacking. Moreover, there are cases of translocation of species out of their natural distribution range (Capa & Murray, 2015b).
A full revision of the genus, based on redescriptions of types, designation of neotypes, and supported by genetic data is needed. In the Indo Pacific Region, the identity of lineages as P. cf. aulaconota, P. japonica or P. cf. rugosa should be elucidated. In Japan, three species of Parasabella were described: P. japonica, P. fullo (Grube, 1878) and P. albicans. Parasabella fullo was redescribed by Keppel, Ruiz & Tovar-Hernández (2020), whereas the the status of the other two species demands further study. Two more nominal species should be taken into account: P. rufovittata (Grube, 1881), described as Sabella from Singapore (type material in the Zoological Museum of Berlin, No. 870) and P. oculea (Pillai, 1965) from the Philippines (type material in the University of Sri Lanka, RTS 25) since these taxa might have a larger distribution in the region. Material examined. Indonesian-Dutch Snellius II Expedition, Sta. 4068, Indonesia, NE coast of Sumba, 09 57′S, 120 48′E, Agassiz trawl, 50 m depth, sandy bottom with sponges and gorgonians, on a whitish ascidian mat, legit R.W.M. van Soest, September 16, 1984, 13 specimens [RMNH.VER. 19945].
Description. Trunk 4.5 mm long, 0.7-0.8 mm wide. Radiolar crown 2.5 times longer than thorax, with seven pairs of radioles. Radiolar eyes absent. Thorax with 6-10 chaetigers. Thoracic ventral shields contact neuropodial tori. Inferior thoracic notochaetae Type A, with hoods three times wider than shaft, and up to three times as long as its maximum width. Thoracic uncini with medium length handles, and neck half of breast length. Abdomen with 15 chaetigers.
Remarks. Parasabella crassichaetae was originally described from Shellharbour (New South Wales, Australia) on an orange sponge from 22.4 m depth (Capa & Murray, 2015b). Its apparent translocation to Queensland, Western Australia and Hawaii, where the species was found in dead coral rubble, sponges, algae and artificial surfaces in ports and harbours, is remarkable. Our specimens have inferior thoracic chaeta type A: broadly hooded with short, abruptly tapering tips, and as such match this complex.
Tubes from the NE coast of Sumba were found below the surface of ascidian mat, with anterior tubes openings directed toward the surface. One specimen was found with three copepods attached to its radiolar tips.
Capa & Murray (2015b) reported regeneration of the posterior end in the holotype and presence of eggs in additional specimens. Specimens reviewed in our study include many buds produced by asexual reproduction. These buds were found within tubes of parents, with transverse fission and regeneration. Some buds have vestigial radiolar crowns (0.2 mm long), with developing radioles without pinnules, or presenting only an anterior blastema. Other buds consist only of trunks formed by 18-23 abdominal segments (no thoracic segments). Transverse fission and regeneration were also reported in Parasabella columbi (Kinberg, 1867) from Argentina (Tovar-Hernández, de León-González & Bybee, 2017).
Detailed illustrations of Parasabella crassichaetae complex were given in the original description by Capa & Murray (2015b).
Genus Perkinsiana Knight-Jones, 1983 (p. 273-274) Perkinsiana. -Fitzhugh, 1989: 78;Capa, 2007: 549;Tovar-Hernández et al., 2012: 57;Capa et al., 2019: 199. Type species: Sabella rubra Langerhans, 1880, by original designation.  Knight-Jones (1983) to accommodate species previously included in Demonax Kinberg, 1867 (= Parasabella fide Tovar-Hernández & Harris, 2010), Potamilla Malmgren, 1866and Potamethus Chamberlin, 1919 modified the generic diagnosis by Knight-Jones (1983), and pointed out that Perkinsiana could not be defined by any synapomorphy. Rouse (1996) emended the generic diagnosis to incorporate features found in other species in his opinion belonging to the taxon.  emended the genus again in order to include changes needed after the exclusion of P. riwo Rouse, 1996(included in Kirkia Nogueira, López & Rossi, 2004 and transferred to Aracia Nogueira, Fitzhugh & Rossi, 2010, when Kirkia was recognized as a homonym) and the addition of two new species (Pekinsiana (sic) longa P. anodina Capa, 2007). Tovar-Hernández et al. (2012) provided a new emendation to Perkinsiana and defined three types (A, B, C) of abdominal chaetae in the genus. Type A for chaetae with a broad hood and progressively tapered. Type B for chaetae having a basal broad knee and tips abruptly narrowed. Type C for elongate abdominal chaetae with narrow hoods. Description. Radiolar flanges and palmate membrane absent. Radioles with rectangular outer margin and digitiform tips. Anterior peristomial ring large, as long as two thoracic segments, exposed dorso-laterally. Posterior peristomial ring collar with dorsal margins separated by a wide gap. Ventral shield of collar quadrangular, broader than the following shields, with straight anterior margin. Ventral shield of chaetiger 2 rectangular with triangular anterior margin. Other shields rectangular with straight anterior margins and of similar length along thorax, separated from uncinial tori by wide gap. Abdominal chaetigers with chaetae Type B. Tubes embedded in live coral, anterior parts emerging as small cones above coral surface (Figs. 17A and 17B). Tube microstructure. Lumen relatively smooth, showing faint micro-relief caused by sparsely spaced fibers. Three different sets of fibers oriented at about 45 or 90 to each other. Fibers in single sets are moderately-developed, long (>13 mm), usually straight and parallel to each other. Interspaces of fibers larger than fiber diameter. Interspaces of adjacent fibers filled with homogeneous organics with smooth surface. Fibers with constant width (0.10 to 0.20 mm). Tube wall dense, non-porous (Fig. 17C).

Remarks. Perkinsiana was proposed by
Remarks. Perkinsiana anodina was described from Western Australia with specimens living in dead coral, large granite boulders with small colonies of live and dead corals, on sand with scattered bommies, and on mussel clumps from pilings of a jetty. It was also reported from the Northern Territory  and on coral ruble in Queensland (Capa & Murray, 2015a).
The tube wall microstructure of P. anodina is not in any way special as compared to the other studied species. This indicates that symbiosis with coral has had no effect on the tube microstructure of the species, possibly because Perkinsiana anodina has no interaction with the host coral through its tube wall.
Regarding morphology, the presence of a palmate membrane in P. anodina is doubtful.   fig. 6B) reported and illustrated the presence of a low membrane between two lateral radioles, but they resemble pinnular tissue instead of a proper membrane.
In her figure 6C, dorsal radioles are not fused basally by a membrane. Our specimens do not have a palmate membrane joining radioles.

Genus
Number of species: 19-23, under revision, see remarks below. Knight-Jones (1983,  and Knight-Jones et al. (2017). The number of validly described species in Pseudopotamilla is not clear at all since there is no worldwide revision of the genus. WoRMS lists 19 species (Read & Fauchald, 2020b) but 23 species are included in Table 6. Note that we regard the status of some nominal species to be questionable. Discrepancies in numbers between WoRMS and Table 6 are due to recombinations, synonymizations, homonyms, revalidations of some species, and transfer to other sabellid genera such as Acromegalomma.

Remarks. Diagnoses to genus level are available in
Regarding Pseudopotamilla from the Indonesian archipelago, Philippine Seas, Australia and New Zealand, five nominal species have been described. Grube (1878) described Sabella (Potamilla) oligophthalmos from Singapore, and Sabella (Potamilla) polyophthalmos from the Philippines. In our study both species are transferred to Pseudopotamilla (Table 6). Pseudopotamilla alba (Knox, 1951) and P. laciniosa (Ehlers, 1904) were described from New Zealand and appear to be currently valid (Table 6). In addition, Capa (2007) described P. monoculata from the Tasman Sea and recorded two additional, unnamed species as Pseudopotamilla sp. A (from Sydney) and Pseudopotamilla sp. B (from New South Wales, Queensland and Western Australia). Later, Capa & Murray (2015a), placed these two taxa in the preliminary synonymy of Pseudopotamilla sp. cf. P. reniformis from Lizard Island.
Tubes: The tube is embedded in calcareous material secreted by the coral (Fig. 18H), its organic wall is translucent (Fig. 20A). Sex and gametes. Unknown.
Tube microstructure. Tube lumen relatively smooth with some faint longitudinal relief at some places. Fibers of at least four different orientations visible on the lumen surface; however, fibers are poorly developed and difficult to see. Interspaces of fibers much wider than fiber diameter. Fibers straight, relatively long (>5 mm), thin (0.10-0.15 mm wide). Interspaces of adjacent fibers filled with homogeneous organics with smooth surface. Tube wall dense, non-porous. There is not much similarity in tube microstructure with the coral symbiont Perkinsiana anodina, nevertheless both species show multiple orientations of fibers (Figs. 20B and 20C).
Remarks. The original description for Sabella (Potamilla) oligophthalmos Grube, 1878, stated the presence of a few eyes along the basal half of dorsal radioles, except on the dorsal most one (a diagnostic feature for Pseudopotamilla), and the presence of long and narrow dorsal lappets. In his description, Grube (1878) also provided a comparison of his new species with Pseudopotamilla reniformis (as Potamilla).  reported Potamilla oligophthalmos from Western Australia and later, as Potamilla (Pseudopotamilla) oligophthalmos from New Zealand (Augener, 1926). In our study, Pseudopotamilla oligophthalmos is redescribed using specimens from the North East coast of Sumba, that were found from 45 to 125 m depth, in dead coral blocks which agrees with the original description.
Pseudopotamilla oligophthalmos differs from P. monoculata in features as number and distribution of radiolar eyes and collar morphology ( Table 7). The presence of glandular areas like those visible in P. oligophthalmos have only been reported for Pseudopotamilla platensis (Hartman, 1953), a species from Argentina, reported up to a depth of 200 m (Tovar-Hernández, de León-González & Bybee, 2017).  proposed two changes in the status of other taxa: he referred Potamilla laciniosa Ehlers, 1904 to Potamilla oligophthalmos (Grube, 1878); and synonymized Sabella (Potamilla) polyophthalmos Grube, 1878 (from the Philippines) with Sabella (Potamilla) oligophthalmos Grube, 1878 (from Singapore), giving the latter priority. It is remarkable that Capa (2007: 559) mentioned some papers to have considered P. laciniosa as valid species, but she did not provide sources for this opinion, these were probably Hartmann-Schröder (1989) and Glasby et al. (2009), both listed in WoRMS (Read & Fauchald, 2020c). Although types of P. polyophthalmos and P. oligophthalmos were not examined by  to support this synonymy, Augener (1926, p. 111) explained that apart from differences in the number of radiolar eyes, Potamilla polyophthalmos and P. oligophthalmos belong to a single species. He did not explain why oligophthalmos should have priority over polyophthalmos, perhaps it was merely based on alphabetic order: in the original descriptions P. polyophthalmos was described and figured first (Grube, 1878: 247−248, pl. 15, fig. 2), P. oligophthalmos came later in the text and was not figured (Grube, 1878: 248−249), one might expect Augener's priority proposal to have been the other way around. Pseudopotamilla polyophthalmos was described with 5−6 eyes per radiole, with a maximum of 10 (many = poly-), whereas P. oligophthalmos only with 1−2 eyes on most radioles (exceptionally up to 3; few = oligo-), except for dorsalmost pair. In our study, we prefer to keep both species separate, until a detailed revision of the supposed synonymy based on the examination of Grube's type material and comparison with Augener's specimens will prove otherwise. The prominent eyespots in (some) species of the genus Pseudopotamilla resulted in three homonyms for two specific epithets: "oligophthalmos" and "polyophthalmos". Pseudopotamilla oligophthalmos originally described as Sabella (Potamilla) oligophthalmos was referred to Pseudopotamilla by Hartman (1959: 557), making Pseudopotamilla oligophthalma (Iroso, 1921) homonymous. The latter was described from Naples as Potamilla oligophthalma and considered to be a synonym of Pseudopotamilla reniformis by Fauvel (1927) and Hartman (1959 Remarks. Grube (1878) described Sabella (Potamilla) polyophthalmos from the Philippines. His description, emphasizing and illustrating the presence of compound eyes in most radioles, except the dorsalmost, matches Pseudopotamilla. Its species have 5−6 eyes per radiole, with a maximum of 10. In our opinion it is a valid species. As indicated with the diagnosis of the genus Pseudopotamilla, above, a worldwide revision is desirable, based on examination of types and topotypical material.
By transferring Sabella (Potamilla) polyophthalmos Grube to the genus Pseudopotamilla as we propose, the new combination has two secondary homonyms: P. polyophthalmos (Langerhans, 1884) (from Madeira) and P. polyophthalma Hartmann-Schröder, 1965 (from Punta Lavapié, central Chile). Examination of types is certainly needed, but the Madeiran species seems to be a junior synonym of P. saxicava (see Knight-Jones et al., 2017), in which case no further action is needed for that species-group name.
Tube microstructure. Tube lumen relatively smooth; poorly developed fibers with the same general orientation. Fibers short (usually shorter than 5 mm); single fibers 0.10 to 0.20 mm wide. Fibers can be slightly curved. Tube wall dense, non-porous (Fig. 21H).
Remarks. Our specimens match the description by Knight-Jones & Mackie (2003) for S. spectabilis from Bohol, Philippines. However, the presence of some radioles with hypertrophied pinnules (wide, flat and long at medium length of radioles) has not been recorded in any species of Sabellastarte. It is remarkable that these unusual pinnules were only seen in the largest specimen. It is not clear if this modification is functional, or due to disease. Spermatids, coelomocytes or full developed gametes were not found in our specimens [RMNH.VER. 19951]. However, S. spectabilis from Hawaii is a protandric hermaphrodite (Bybee, Bailey-Brock & Tamaru, 2006), while the same species is reported to be gonochoric in Micronesia (Rouse & Fitzhugh, 1994). This indicates that more than one species may be included under the same name.
On the other hand, Augener (1933) reported Sabellastarte magnifica Shaw, 1800 from Biliton (Belitung) Island, Indonesia. He emphasized that S. indica Savigny, 1822 and the species S. magnifica from the West Indies are synonyms. This synonymy was not followed by Knight-Jones & Mackie (2003), they regarded Indo-West Pacific and Caribbean taxa as separate species. In addition, Capa, Bybee & Bybee (2010) found different lineages between specimens attributed to Sabellastarte from the Caribbean and the Indopacific (Malaysia and Saipan), and concluded they are separate species.
Nomenclatural discussion on the use of Sabellastarte spectabilis (Grube, 1878) over S. indica (Savigny, 1822) and their authorship Knight-Jones & Mackie (2003: 2269 stated that Sabella indica Savigny, 1822 (subsequently designated as type-species of Sabellastarte by Bush (1905)) is preoccupied by the pectinariid Sabella indica Abildgaard, 1789, making it necessary to find a different type species for the genus Sabellastarte, for which they selected Sabella spectabilis Grube, 1878. This is incorrect because the type species "remains unchanged even when it is a junior synonym or homonym, or a suppressed name" (ICZN, 1999, Art. 67.1.2). However, while reviewing original data sources in order to track this history, we found some discrepancies and inconsistencies necessitating a separate contribution, which will not complete without examination of the syntypes of Sabella indica deposited at the Museum National d'Histoire Naturelle, Paris (MNHN POLY TYPE 608, 609).
Firstly, in Abildgaard (1789) nowhere the Latin name "indica" can be found. Abildgaard only compares a large pectinariid from the East Indies with Amphitrite auricoma Müller, 1776 and Sabella granulata Linnaeus, 1767, both taxa belonging to the family Pectinariidae according to WoRMS (Read & Fauchald, 2020d). Under this scenario, there is no homonymy for Sabella indica Savigny, 1822, and the new proposal of Sabellastarte spectabilis (Grube) as type would be incorrect.
Second, there is confusion about the authorship of Sabella indica. It can be found in the literature and WoRMS as: Sabella indica Gmelin in Linnaeus, 1788; Sabella indica Savigny in Lamarck, 1818 and Sabella indica Savigny, 1822(Hartman, 1959Knight-Jones & Mackie, 2003;Read & Fauchald, 2020e). In this contribution we follow Knight-Jones & Mackie (2003) but is it clear that Sabellastarte indica and S. spectabilis require a thorough revision.
On the microstructure of organic tube wall in sabellids All studied sabellids have a tube wall with a purely organic composition. This organic tube wall has a lamellar microstructure; lamellae are composed of fine, long fibers. The fibers in the single lamellae of some species seem to have a similar general orientation (i.e., Acromegalomma acrophthalmos, Sabellastarte spectabilis), but in other species fiber orientation varies. Usually lamellae of the tube wall are composed of fibers with alternate orientations. There can be up to four different main orientations of fibers in a single lamella. In case of two alternate orientations, fibers are usually located at angles of 35-40 to each other (i.e., Bispira porifera). In case of three different orientations, two of these are at an angle of 45 , respectively 90 as compared with the first (i.e., Perkinsiana anodina). In Pseudopotamilla oligophthalma the orientation of fibers do not fall into well-defined categories. The diameters of fibers seem to be similar in most species studied, being between 0.10 and 0.20 mm, nevertheless, in B. porifera fibers are slightly thicker.

CONCLUSIONS
The Indonesian archipelago, the South China and the Philippine Seas are an important marine biodiversity hotspot with especially rich marine life. Until now, 23 genera and 78 species of Sabellidae were reported from the whole area Augener, 1933;Mesnil & Fauvel, 1939;Pillai, 1965;Gallardo, 1968;Nishi, 1998;Fitzhugh, 2002;Al-Hakim & Glasby, 2004;, Capa, 2008Capa & Murray, 2009;Salazar-Vallejo et al., 2014;Glasby, Lee & Hsueh, 2016;Nishi et al., 2017;Hadiyanto, 2018;Nishi, Tanaka & Tovar-Hernández, 2019;, and the present study). This represents 15.82% of the 493 species, and 57.5% of the 40 genera of sabellids currently known in the entire world . Surely these numbers will increase, at least 11 taxa mentioned as "spec." are now already waiting for formal description (Fitzhugh, 2002;Al-Hakim & Glasby, 2004; this study). Certainly, through exploration of other habitats and depths with the new sampling techniques, and with modern methods for identifying and analyzing biodiversity, we expect a significant raise in numbers. Special attention should be given to molecular analyses that allow for distinction between cryptic "look-alike" taxa, as well as invasive species.