Flagellia , a new subgenus of Haliclona ( Porifera , Haplosclerida )

Haplosclerid sponges possessing a unique asymmetric fl agelliform type of sigmoid microsclere have been reported from all global oceans. This peculiar spicule, characterized by a circular or elliptical shape, with a longer and sharper curved ending at one side and a shorter and more gradually curved ending at the opposing side, is proposed to be termed ‘fl agellosigma’. These sponges invariably also possess smaller normal sigmas while their skeletal structure of oxea megascleres is markedly confused. They are assigned to the large genus Haliclona Grant, 1841 (family Chalinidae) in a new subgenus, Haliclona (Flagellia) subgen. nov. The species belonging to the new subgenus are reviewed and four species new to science are described, Haliclona (Flagellia) indonesiae subgen. et sp. nov., H. (F.) amirantensis subgen. et sp. nov., H. (F.) hiberniae subgen. et sp. nov. and H. (F.) hajdui subgen. et sp. nov. One species, H. (F.) hentscheli nom. nov., is given a new name on account of secondary homonymy caused by its transfer to the genus Haliclona. One species remains unnamed because of paucity of material. Already known species, reassigned to the new subgenus are H. (F.) hamata subgen. et comb. nov., H. (F.) fl agellifera subgen. et comb. nov., H. (F.) porosa subgen. et comb. nov., H. (F.) edaphus subgen. et comb. nov. and H. (F.) anataria subgen. et comb. nov. Additional species are likely hiding among many erroneous records of ‘Gellius fl agellifer’ from wide ranging parts of the global oceans.


Introduction
The present study addresses a group of haplosclerid sponge specimens united by a remarkable type of microsclere, an asymmetrical flagellated or flagelliform sigma, which is here proposed to be named 'flagellosigma'.Specimens possessing this spicule type, which invariably also possess a smaller 'normal' sigma type, occur in many parts of the global oceans and have been routinely assigned to Gellius flagellifer Ridley & Dendy, 1886, currently named Haliclona (Gellius) flagellifera.Additionally, several specimens with the same general spicule complement (flagellosigmas and normal sigmas) have been described as separate species ('Desmacella' porosa Fristedt, 1887, Gelliodes hamata Thiele, 1903, Gellius incrustans Hentschel, 1912, Gellius edaphus De Laubenfels, 1930, Gellius rhaphidiophorus Brøndsted, 1933 and Gellius anatarius Lévi & Lévi, 1983).These additional species with flagellosigmas are currently mostly assigned to the subgenus Haliclona (Gellius), with the exception of Hemigellius porosus (Fristedt, 1887) and Gelliodes hamata.The collections of the Naturalis Biodiversity Center at Leiden incorporate a representative set of flagellosigma-bearing specimens from various parts of the world, allowing a review of the status of Haliclona (Gellius) flagellifera and its morphological siblings.Despite the superficial resemblance of all records and reports of this seemingly 'cosmopolitan species' this study demonstrates local differences in shape and spicule types and sizes, leading to the hypothesis that the flagellosigma can be considered apomorphic to a monophyletic group of sponge species, and merits recognition at the (sub)genus level.Taxonomic decisions of species status and erection of new species is done below only using collected specimens.The status of taxa discussed on the basis of reports in literature is restricted to tentative suggestions.

Material and methods
Specimens and slides of the following institutions have been studied: -Naturalis Biodiversity Center, Leiden, The Netherlands, specimens from the collections of the former Rijksmuseum van Natuurlijke Historie (registration numbers preceded by the acronym RMNH) and the former Zoologisch Museum van de Universiteit van Amsterdam (registration numbers preceded by the acronym ZMA) -The Natural History Museum, London (registration numbers preceded by the acronym BMNH) -Senckenberg Museum, Frankfurt (registration numbers preceded by the acronym SMF) -Smithsonian Institution, National Museum of Natural History (registration number preceded by acronym USNM) -Western Australian Museum, Perth (registration number preceded by acronym WAM) Specimen processing included making thick sections and dissociated spicule slides for light microscopic and SEM examination closely following previously described procedures (for details see Van Soest et al. 2014: 62).Length and thickness measurements are based on 25 randomly chosen spicules of each distinguished type.Measurements and shape characterization of flagellosigmas were carried out following the meristic parameters depicted in Fig. 1 and described below.

Definition of a flagellosigma
A flagellosigma (Fig. 1) differs in two major aspects from sigmas in general: (1) asymmetry in the length of the opposite endings, and (2) asymmetry in the curvature of these endings.Furthermore, compared to a normal sigma the flagellosigma has the spicular axes either more or less equal or it has the length-and width axes reverted, resulting in a circular, ovoid or elliptic outline.The flagellosigma occurs only in the order Haplosclerida.Sigmas in other groups often have a slight asymmetry in the opposite endings, one being more sharply curved than the other, but their lengths are almost always equal and the shape is mostly a shallow C-shape with normally longer length axes and shorter width axes.Only in a very few cases, e.g., in Mycale (Naviculina) diversisigmata Van Soest, 1984, the shape is circular or elliptical and the two endings may have different lengths, but in that case these are more or less symmetrical in their curvature, not overtly asymmetrical (Fig. 2).The term flagellosigma is proposed to be strictly limited to sigma shapes as depicted in Fig. 1.In order to investigate the meristic and shape differences between flagellosigmas of different individuals and species, a series of measurements and qualifiers are proposed.

Measurements
1. length of the longest ending, the long axis 2. length of the shortest ending 3. width of the spicule at the widest part, usually the short axis 4. thickness of the spicule at its most outward curve European Journal of Taxonomy 351: 1-48 (2017)  VAN SOEST R.W.M., Haliclona (Flagellia) subgen.nov.Qualifiers 1. shape of the spicule, either circular, ovoid, or elliptical 2. curvature of the longest ending, either rounded, short-straight, long-straight, long-upturned 3. curvature of the shortest ending, either shallow or wide

Taxonomic history of flagellosigma-bearing sponges
The first description and depiction of a flagellosigma spicule is from Schmidt (1870: 53) who refers to an encrustation on Pachastrella abyssi Schmidt, 1870, from a depth of 410 m, with robust oxeas ('stämmige Spindelnadeln') and very peculiar sigmas ("sehr eigentümliche Spangen", illustrated in pl.V, fig.15), as an apparent Desmacella sp.These very short remarks, made in a section under the caption "Desmacella vagabunda nova species", do not actually name the encrustation, but since they follow properly named descriptions of Desmacella vagabunda Schmidt, 1870, Desmacella annexa Schmidt, 1870 andDesmacella vicina Schmidt, 1870 in the same section, it is reasonable to assume Schmidt meant the unnamed sponge to belong to his genus Desmacella Schmidt, 1870.
Schmidt's description of Desmacella vagabunda leaves no doubt that this is a proper Desmacella with tylostyles ("Stecknadeln") and sigmas ("Spangen") (which were of widely different sizes between 14 and 100 µm).The same applies to the description of D. annexa, which he said was a variety with the same spicules as D. vagabunda, but with thin oxea-like spicules added to the complement.I was able to study a slide of the type (BMNH 1870.5.3.29, Florida, 195 fthms) and can confirm the properties of this Desmacella.D. vicina is likewise described with tylostyles and sigmas (but these are only limited to two size variations), confirmed by the study of two slides of Schmidt's collection (BMNH 1870.5.3.40, BMNH 1870.5.3.113).The locality for Desmacella vagabunda is given as Florida, D. annexa likewise ("von ebendaher", and Schmidt's BMNH slide reads "Florida"), but no localities are given for D. vicina (nevertheless Schmidt's BMNH slide also says "Florida") and the flagellosigma-bearing D. sp.However the latter was encrusting the Caribbean species Pachastrella abyssi, which allows the conclusion that this also concerns a species from the Central West Atlantic region, probably Florida.
Vosmaer (1880: 108) assigned Schmidt's Desmacella vagabunda to the genus Desmacodes Schmidt, 1870, for which he devised a new concept containing species from many different genera and families.He correctly defined D. vagabundus as having tylostyles ("tr o .ac")and sigmas ("∞").Several years later, Vosmaer (1885: 28) reassigned D. vagabunda to the genus Gellius Gray, 1867, as 'XXX.Gellius vagabundus (O.S.) Vosm.', stating that Gellius has priority over Desmacodes.He also redescribed the species vagabundus, in a very loose and frustrating way, based on extensive material collected by the Willem Barents Expeditions (1878-1882) in the Barents Sea.He maintained that the species is quite variable and assigned specimens with tylostyles, styles and oxeas, and with or without sigmas, to it.The species would also have a very large distribution.In the Barents Sea material, he distinguished three varieties α, ß and γ, the latter of which was described as having the spiculation of oxeas, flagellosigmas and small normal sigmas (cf.illustrations in Vosmaer 1885: pl.V, figs 36-38) (see further below).Ridley & Dendy (1886: 333;1887: 42, pl. XIII, figs 5, 10) described Gellius flagellifer from the Southern Indian Ocean (off Marion Island) from a depth of 90-135 m.In their remarks they referred to Vosmaer's Gellius vagabundus var.γ, and tentatively assigned this to their new species (pointing out that Gellius vagabundus sensu Schmidt is a Desmacella).They apparently were unaware of Schmidt's (1870: pl. V, fig. 15) drawing of a flagellosigma very similar to that found in their new species.Fristedt (1887: 440, pl. 24 figs 36-37, pl. 28 fig. 15), unaware of Ridley & Dendy's G. flagellifer, described Desmacella porosa from Arctic waters, Davis Strait, 126 m depth.He referred to Schmidt (1870: pl. V fig. 15) to demonstrate the similarity of the flagellosigmas in Schmidt's and his own species.Fristedt did not mention the presence of small normal sigmas in his specimen.
European Journal of Taxonomy 351: 1-48 ( 2017) Lambe (1896: 185, pl. I fig. 4) reported Gellius flagellifer from the Gulf of St. Lawrence, Eastern Canada, at a depth of 68-136 m, in the process synonymizing Desmacella porosa with it.This record was mentioned by De Laubenfels (1949: 38) and Gosner (1971).Topsent (1896: 281, pl. VIII fig. 4) reported Gellius flagellifer from the nearby Gulf of Biscay off the coast of SW France, depth 400 m.He did not mention the presence of small normal sigmas (but see below).Lundbeck (1902: 71, pl. II fig. 9, pl.XIV figs 1 and 73, pl.XIV fig.2) reported both Gellius flagellifer (Lundbeck 1902: 71, from a locality between SE Iceland and the Faroe Islands, 756 m depth) and Desmacella (as Gellius) porosa (Lundbeck 1902: 73), one specimen from a locality off the N coast of Iceland at a depth of 214 m, the other from an unknown depth between Iceland and the Faroe Islands.Lundbeck maintained that there were differences between the two species, and reassigned specimens reported by Topsent and Lambe to G. porosus.Burton (1959a: 19) listed Lundbeck's record of Gellius flagellifer as Haliclona flagellifer in his catalogue of the sponges of Iceland, without referring to or mentioning G. porosus.Thiele (1903: 942, fig. 7) described the finger-shaped Gelliodes hamata from the shallow-water off Ternate, Indonesia, which combined flagellosigmas and smaller normal sigmas with a spongin enforced reticulation of oxeas.He referred to the similarity of the flagellosigmas to those of Gellius flagellifer, but the skeletal structure appeared different enough to him to distinguish them at the genus and species level.Topsent (1904: 231) reported Gellius flagellifer from the Azores, at a depth of 845-1360 m, and this time properly mentioned the presence of normal sigmas, admitting that they were also present in the sample from the Gulf of Biscay, reported in 1896 (see above).Lundbeck (1909: 434) reported Gellius porosus from several localities off the coast of East Greenland, at depths of 90-252 m.No descriptive information was provided.Hentschel (1912: 390, pl. XV fig. 3, pl. XXI fig. 45) described a small encrusting species Gellius incrustans, possessing flagellosigmas, from the Aru Islands, Indonesia, depth 12 m.He pointed out the similarity with Gellius flagellifer, but judged that the spicule size differences were sufficient for specific distinctness.Hentschel (1916: 11) reported Gellius porosus from Spitsbergen (Svalbard), depth 141-147 m.No taxonomic information was provided.Hentschel (1929: 978) summarized the occurrence of Gellius flagellifer and Gellius porosus in the Arctic region, relying on Lundbeck (1902), without providing additional data.Stephens (1916: 233;1917: 5;1921: 5) extensively described specimens of Gellius flagellifer from off the W coast of Ireland, at depths of 90-1328 m.She supported the synonymization of Gellius porosus with it.
Ferrer Hernandez (1918: 22, fig. 3) reported Gellius porosus from off the coast of Asturias, N Spain, at 200 m depth.He attempted to distinguish it from Gellius flagellifer on the shape of the flagellosigmas.Ferrer Hernandez (1923: 262) listed both species without further data from all coasts of Spain.Dendy (1922: 26) described Gellius flagellifer from Saya de Malha in the W Indian Ocean, at a depth of 99 m.He believed that the species was cosmopolitan and that G. porosus is a synonym.Babič (1922: 228, fig. H) remarkably reverted back to Vosmaer's nomenclature and used the name Gellius vagabundus for a flagellosigma-bearing specimen from the Adriatic Sea, depth 45 m.
VAN SOEST R.W.M., Haliclona (Flagellia) subgen.nov.Rezvoi (1924: 243) recorded Gellius porosus from the Kara Sea, Siberia, at a depth of 106 m.He did not mention small normal sigmas.A few years later, Rezvoi (1928: 91) described, from the nearby Barents Sea, both Gellius porosus (which this time was described to contain small 'symmetric' sigmas as well as flagellosigmas), depth range 92-183 m, and Gellius flagellifer, from 91-769 m.The two species were separated on spicule size data.Dendy (1924: 320) briefly described Gellius flagellifer from Three Kings Island, New Zealand, at a depth of 180 m.Bergquist & Warne (1980: 22) and Kelly et al. (2009: 44) confirmed this record (as Sigmadocia flagellifer).Topsent (1928: 314), as Babič before him, took the erroneous view that his earlier reports on Gellius flagellifer were part of what he considered to be Gellius vagabundus (Schmidt, 1870), referring to Vosmaer as the inspiration for this.In fact, Topsent's (1928) description of 'Gellius vagabundus' from 1378 m near São Miguel, Azores, is likely to be a different species from his other described specimens as the sigmas appear dissimilar to the flagellosigmas discussed here.It is possible that this could be a species of Haliclona (Gellius).Burton (1928: 114) described two specimens as Gellius flagellifer from the Andaman Sea, from depths of 310-896 m.This record was repeated by Pattanayak (1999: 450).The same material was reexamined by Pattanayak (2006: 74, pl.IXA, text-fig.51a-c), but fig.51a-c does not show a proper flagellosigma rendering this record uncertain.Burton (1930) listed Gellius flagellifer from Hammerfest, N Norway, without further data.
De Laubenfels (1930: 28;1932: 111, fig. 66) described Gellius edaphus from an intertidal cave off the coast of Southern California, depth approximately 0-1 m.He did not distinguish flagellosigmas and normal sigmas, but a redescription of the type material by Lee et al. (2007: 110, as Xestospongia edapha) revealed their presence.Brøndsted (1933: 18, fig.7) described Gellius rhaphidiophorus from the Davis Strait, W Greenland, from a depth of 410 m.The fragments were similar to previous reports of Gellius porosus and Gellius flagellifer from the Davis Strait, but according to the author the specimen has raphides (not in dragmata).Burton (1938: 7) reported Adocia flagellifer from a depth of 36 m in the Commonwealth Bay, East Antarctic.He provided some information on the flagellosigmas.Koltun (1964Koltun ( : 103, translated in 1966: 102: 102) repeated this record.Dickinson (1945: pl. 14 fig. 27-28, pl. 15 fig. 29) described Sigmadocia edaphus from Baja California (Mexico), at a depth of 120 m. Green & Bakus (1994: 46, fig. 27) reported this species from Southern California, at a depth of 200 m.Koltun (1959) 170 it is clear that his record of Gellius flagellifer is not correct, as he omitted to mention and depict the presence of flagellosigmas.Some years later, Koltun (1962: 186) reported Haliclona porosus from the Paramushir Islands, NW Pacific (see also Hoshino 1987).Burton (1959b: 218) briefly described specimens named Haliclona flagellifera obtained in the Southern Red Sea (at 26 m depth) and near the Maldives, at a depth of 229 m.
Van Soest (2017) described and figured Haliclona (Gellius) sp.aff.flagellifera from the Guyana shelf, off the coast of Suriname, N South America, depth 130 m.He also mentioned the existence of a sample with flagellosigmas from reefs off Santa Marta, Colombia, depth 20 m.Van Soest (1980: 25) and De Weerdt (2000: 64) reported the existence of a specimen in the collection of the Natural History Museum, London (BMNH 1937.11.26.16a), from Turneffe Island, Belize.Alcolado (2002: 67, based on Kaminskaya 1971) gave a description of a Haliclona sp. from Cuba with flagelliform sigmas.These records indicate a widespread occurrence in the Western Atlantic.
From this literature review it is clear that sponges sharing the spicule complement of oxeas, flagellosigmas and small normal sigmas have been been reported from all over the world (Arctic, Antarctic, North Atlantic, Mediterranean, Central West Atlantic, SE Atlantic, Southern Ocean, Western Indian Ocean, Indo-West Pacific, NW Pacific, NE Pacific, East Pacific and SE Pacific).The Marine Ecoregions (sensu Spalding et al. 2007) that yielded reliable reports of specimens of Haliclona (Flagellia) subgen.nov., are presented in Fig. 3.So far, there have been five species named with flagellosigmas, but from the ubiquitous reports of one of them, H. (G.) flagellifera, it is likely that several more such species await to be recognized.The skeletal structure of most of these species conforms to the subgenus Haliclona (Gellius) as defined by De Weerdt (1986Weerdt ( , 2002)), but at least one species has spongin enveloped polyspicular fibers, unlike those of Haliclona (Gellius).It is not unreasonable to hypothesize that these species and specimens possessing the complement of oxeas, flagellosigmas and normal sigmas, distributed globally, form a monophyletic subset of Haliclona to be recognized here at the subgenus level and named Flagellia subgen.nov.The new subgenus differs from Haliclona (Gellius) as defined by De Weerdt (1986Weerdt ( , 2002) ) in the possession of flagellosigmas.

Systematic descriptions
The order in which the species are treated is geographic, starting with Indonesia, then Western Indian Ocean, followed by Central West Atlantic, North Atlantic, and ending with the Eastern Pacific.Flagellia subgen.nov.urn:lsid:zoobank.org:act:B6ABC6E8-EF63-4D79-A636-B3348CC80D3A

Etymology
The name is derived from the Latin word 'flagellum', meaning 'whip', which refers to the whip-like flagellosigma.

Remarks
This subgenus shares with mainstream Haliclona species a skeleton in which the ascending spicule tracts are interconnected by single megascleres.There is usually no distinct detachable ectosomal skeleton, although tangential arrangement of the oxeas at the surface is common.The choanosomal skeleton tends to be very loosely organized, verging to confused.In that aspect it conforms most closely to species of the subgenus Gellius, but in that subgenus the sigmas are symmetrical and often angular.Although symmetrical normal sigmas are part of the spicule complement of the new subgenus, these are never angular.The habitus of members of the subgenus varies strongly, from small crusts to elaborate plates or arborescent forms.Association with other sponges or other sessile organisms appears common.
The subgenus is found all over the world's oceans.The depth occurrence is wide, but so far is confined to coastal, continental platform and upper bathyal waters.
Ten species are recognized here, four of which are new to science, one is given a new name due to junior homonymy, and one remains unnamed due to limited available material.From the historical overview presented above it is likely that several more species will be found to be extant.
It would perhaps have been logical to choose Haliclona (Flagellia) flagellifera as the type species for the new subgenus as it is the most closely associated name to species belonging to Flagellia subgen.nov.(cf. the historical overview above).However, the holotype of H. (F.) indonesiae sp.nov.was collected VAN SOEST R.W.M., Haliclona (Flagellia) subgen.nov.
recently and is also quite large in size making subsampling for DNA sequencing a viable option for nearfuture phylogenetic studies of the position of Flagellia subgen.nov. in the order Haplosclerida.

Etymology
Named after the country where the holotype was collected.

Description
The holotype (Fig. 4A, A1, A2) is a large plate-like sponge, tending to form a very shallow cup with folding sides.Size 25 × 20 cm, less than 1 cm thick.Color pinkish cream alive, orange-cream in alcohol.Surface smooth, riddled with rounded holes in life, but these contract in alcohol.A few oscules of about 5 mm are present.Consistency firm.The paratype (Fig. 4B) is broken into three flat fragments, but together these comprise also a large plate-like sponge.The life color was noted as light brown, but in alcohol it is slightly darker brown.Surface is similarly smooth and no oscules are apparent.The additional specimens are smaller flat encrustations.

Remarks
In spicule shapes and sizes the new species is extremely close to Indonesian Haliclona (Flagellia) hamata (Thiele, 1903) (see below).The shapes and sizes of the flagellosigmas are virtually identical, and the length of the small normal sigmas is similar.However, there are three distinct differences: the body shape of H. (F.) hamata is digitate to arborescent, the oxeas are larger and especially thicker (264-425 × 13-24 µm), and the normal sigmas are considerably more robust (thickness 1.5-2.5 µm).The combination of these differences confirms the specific status of the two sympatric species.A third Indonesian species H. (F.) hentscheli nom.nov.(see below) differs clearly in having two size categories of flagellosigmas and normal sigmas, and smaller and thinner oxeas.
The presence of this species in NW Australia is here reported on the basis of a photo of an in situ specimen, a photo of an 'on deck' labeled fragment of that specimen, and a light microscopic photo of the skeleton and spicules made from the fragment.These images were graciously provided by one of the manuscript reviewers.Although I did not study the material myself, the images provided sufficient evidence for a positive identification as Haliclona (Flagellia) indonesiae subgen.et sp.nov.

Description
From a thickly encrusting base, the sponge issues upright branches, which may divide higher up.In the holotype these finger-shaped digitations are 2-2.5 × 0.7 cm in size (Thiele 1903), in ZMA 09285 (Fig. 7) the branches are up to 13 cm long and 1 cm in diameter, in ZMA Por.09050a there are only lumpy fragments.The color is light brown or yellow-brown, both in situ and in alcohol.Surface optically smooth, with a few flush oscules of about 3 mm in diameter.There are encrusting bryozoans and hydroids (the holotype is described as bearing small stones and other foreign particles).Consistency firm.

Remarks
As discussed above, this species is distinguished primarily by a digitate-arborescent habitus from the very similar but plate-like morphology of H. (F.) indonesiae sp.nov.The flagellosigmas of the present species are also subtly larger and thicker than those of the new species; the normal sigmas are distinctly more robust.

Etymology
The specific epithet refers to E. Hentschel, author of Gellius incrustans.

Remarks
The specimens described here are judged to be conspecific with Gellius incrustans Hentschel, 1912.However, data provided by Hentschel do not entirely match the present specimens: normal sigmas are described as very common, but no size categories were mentioned; only the largest size is quoted as 43-56 µm, smaller than the present 57-81 µm.Flagellosigmas are quoted as having a largest 'Durchmesser' of 47-51 µm, likewise smaller than in the present specimens.Oxeas were 156-180 × 5-6 µm, according to Hentschel.It remains to be established whether the differences observed here are the result of a less VAN SOEST R.W.M., Haliclona (Flagellia) subgen.nov.than optimal description by Hentschel, or a genuine difference, in which case the present specimens belong to an undescribed species.
Apart from these differences, transferring Gellius incrustans to the combination Haliclona (Flagellia) incrustans, created a junior secondary homonym of Haliclona foraminosa incrustans (Czerniavsky, 1880) (originally Protoschmidtia foraminosa forma incrustans) and of Haliclona simulans incrustans (Carter, 1887) (Carter 1887: 70, originally Isodictya simulans var.incrustans).Burton (1959b: 220) already solved the latter case of homonymy by giving Carter's subspecies the new name Haliclona carteri Burton, 1959.Here the new combination Haliclona (Flagellia) hentscheli nom.nov. is proposed to solve the homonymy with Czerniavsky's (1880) species, which, in spite of its unrecognizable description remains a senior secondary homonym until such time as its status is resolved.Future reallocation of these species to other valid genera will require reinstatement of Hentschel's and Carter's names.Burton's (1928) deep-water record of Gellius flagellifer from the nearby Andaman Sea was possibly the present species, as the upper size of the normal sigmas falls within the variation of the above measurements.However, the oxeas of the Andaman specimens are 280-360 × 12-14 µm, well in excess of those measured above.Combined with the deepwater occurrence (300-900 m) the conspecificity appears doubtful.

Etymology
The specific epithet refers to the type locality.

Description
The sponge (Fig. 10A, circle) forms a central encrustation of approximately 2 × 2 × 0.5 cm on a large Topsentia knoll of 9 cm high and wide.It has an irregular outline around a 5 mm diameter oscule.The color of both sponges was noted as beige and the specimen of Haliclona (Flagellia) was only detected by its softer consistency and a less coarse surface.
Skeleton.Confused anisotropic organization with large open spaces, with spicule tracts cored by 1-6 spicules in cross section bound by spongin, but this is not obviously enclosing the tracts.Interconnecting spicules are single oxeas, arranged loosely and irregularly.The surface has a tangential arrangement of single spicules differentiated from the choanosomal reticulation.

Distribution and ecology
Seychelles, epizoic on sponge in sandy bottom beyond reefs, 50 m depth.Also, if synonymy is correct, Madagascar, Kenya, Maldives, and possibly Saya de Malha (Marine Ecoregions Seychelles, East African Coral Coast, Western and Northern Madagascar, Maldives), 37-229 m.

Remarks
The description by Dendy (1922) of a fairly large encrusting specimen (5.5 × 5 × 1 cm) from Saya de Malha (98 m depth) with the name Gellius flagellifer Ridley & Dendy, 1886 possibly conforms to the present species.The flagellosigmas were described as having an upturned curve on the long ending and the presence of visible spongin was also noted.However, the sizes of the oxeas were given as 370 × 20 µm, well in excess of the Seychelles specimen, and no data on sizes of flagellosigmas and normal sigma were provided.This meagre information is not sufficient to be certain of conspecificity.Burton (1959b) reported Haliclona flagellifer from the Southern Red Sea (26 m) and the Maldives (229 m).The specimen from the Southern Red Sea had oxeas only 170 × 10 µm, clearly smaller than the above measurements.The Maldives data appear closer, with oxeas 320 × 19 µm, flagellosigmas 90 µm, and sigmas 30-60 µm.Vacelet et al. (1976) recorded Gellius flagellifer from Southwestern Madagascar (at 37 m depth, beyond the reefs) and this description matches the above description in most aspects (color, skeleton, sizes and shapes of oxeas and flagellosigmas), except for the normal sigmas, which were given as 30-40 × 1.2-2 µm.However, their drawing of these spicules (fig.62c) shows considerable size variation.There is little doubt that the Madagascar and Amirante material are conspecific.Pulitzer-Finali (1993) reported Sigmadocia flagellifer from deeper water (117-138 m) off the coast of Kenya.Oxeas were somewhat larger (310-370 × 13-17 µm), but flagellosigmas and the larger normal sigmas were similar in size.No mention was made of a smaller sigma category, rendering conspecificity uncertain.However, as the specimen also encrusted a sponge (Asteropus), this material has more similarities than differences.
Haliclona (F.) hentscheli nom.nov.as described above is quite similar to the Seychelles species in shape, oxea length, shape of the (large) flagellosigmas and presence of two size categories of normal sigmas.The major difference is the lack of a differentiated small flagellosigma category and the size of the larger normal sigma category, which is clearly smaller (average 54 µm) than that of H. (F.) hentscheli nom.nov.(av.71 µm).Haliclona (F.) flagellifera (Ridley & Dendy, 1886) subgen.et comb.nov.from Marion Island (see below), differs from H. (F.) amirantensis subgen.et sp.nov. in the shape of both the larger and smaller flagellosigmas, the presence of upturned long endings in many of the large flagellosigmas, the presence of two normal sigma categories, and the smaller sizes of the oxeas.
A specimen of H. (F.) flagellifera reported from Kerguelen by Boury-Esnault & Van Beveren (1982) does have flagellosigmas with upturned endings, but is otherwise (oxea sizes, shape of the flagellosigmas, normal sigma sizes) clearly different from H. (F.) amirantensis sp.nov.Additional comparisons are given below.
Skeleton.Haliclona-like, confuse, anisotropic, with primary lines consisting of 1-4 spicules, single spicules interconnecting at all angles, but mostly rectangular, no spongin visible.No special ectosomal arrangement of skeleton and spicules.

Remarks
Although cosmopolitan distribution is unlikely to occur in sponge species with a depth range limited largely to continental and upper bathyal waters, it is still possible that the present species from an oceanic island could have been capable of covering large distances.Thus, the reported occurrence of H. (F.) flagellifera from circumglobal southern ocean localities could be consistent with the occurrence of a single species.Here the literature data from the non-tropical southern ocean records of the species is reviewed.Despite numerous reports (see Remarks below), we consider that the only reliable record of H. (F.) flagellifer originates from the type locality at Marion Island (South African administration).
Boury-Esnault & Van Beveren (1982) reported Gellius flagellifer from Kerguelen Islands at a comparable depth of 195 m.The shape of their specimens was also massive encrustations, up to 4.3 × 3 × 0.8 cm.The oxeas were reported as 474-540-589 × 13-14-24 µm, clearly longer than those of the type.The European Journal of Taxonomy 351: 1-48 (2017) flagellosigmas had their longest axis 88-98-129 µm (= length of long ending) and their shorter axis 45-56-67 µm (= width), close in measurements to those of the type specimen, but no small flagellosigmas were mentioned.The shape of the flagellosigma in their paper is more narrow-elliptical than in the type, and the long ending has a faint upturned hook.Normal sigmas have a wider range, 40-83-131 × 2-5 µm, and in the illustrations clearly appear to be divisible into two sigma size categories unlike the normal sigmas in the type.Therefore it is uncertain whether the Kerguelen material is conspecific with the type, and for now is considered to belong to an unnamed Haliclona (Flagellia) spec.until the specimen can be examined.Burton (1938) reported the species (as Adocia) from Eastern Antarctic Wilkes Land, directly south of Australia, at a depth of 36 m.He provided no data, other than remarking that the flagellosigmas reached 120 µm in the longest axis, a similar length to those of the type, but not sufficient to conclude that the Antarctic specimen is conspecific.Göcke & Janussen (2013) reported this species from the Eastern Weddell Sea, Antarctica, at a depth of 602 m.Oxeas were 570-643-715 × 22-26-29 µm, clearly considerably longer and thicker than the type.Flagellosigmas measured 80-106-140 µm in the longest axis, 60-109-155 µm in the shorter axis, also larger than the type specimen.The normal sigmas were 17-24-31 µm, smaller than in the type.Although generally similar to the type, the spicule size data and the lack of differentiated larger and smaller flagellosigmas indicates a likely specific difference.
It is not possible to judge whether specimens reported by Pansini & Sarà (1999) from Magellan Strait are similar to the type specimen, because no description was provided.Dendy (1924) and Bergquist & Warne (1980) reported specimens from northern New Zealand waters (Three Kings Islands), at depths of 200 m and 60-120 m, respectively.The specimens differed in the size of the oxeas, with Dendy's specimen possessing oxeas of only 210 × 8 µm, while those of Bergquist & Warne were close to those of the type in size, 460 × 15 µm.Both specimens had small flagellosigmas of 46 and 64 µm respectively, and a single size of normal sigmas (20 and 28 µm respectively).Neither specimen appears very close to the type morphologically.
Uriz (1987,1988) described Gellius flagellifer from Namibia, SE Atlantic (at depths of 183-290 m).The oxeas were given as 420-570 × 16-30 µm, much larger than the type.In addition, the flagellosigmas and normal sigmas were larger, suggesting that the Namibian material could be specifically different.These comparisons lead to the conclusion that Haliclona (Flagellia) flagellifera is so far endemic to Marion Island of the Prince Edward Islands archipelago in the Southern Indian Ocean.Specimens reported as Haliclona (Gellius) flagellifera from other ocean basins are regarded as likely different species.

Description
Because of the substantial difference between the two localities, each specimen is described separately.
ZMA 20742 (Fig. 12A-F) is a slide only, made from a dried fragment apparently no longer present in the collection.

Distribution and ecology
Arctic waters, NW Pacific, Gulf of Biscay, Gulf of Saint Lawrence, off Mauritania (Marine Ecoregions West Greenland Shelf, East Greenland Shelf, North and East Barents Sea, South European Atlantic Shelf, Gulf of Saint Lawrence, Sahelian Upwelling).Depth occurrence 90-400 m.

Remarks
All slides and specimens in the ZMA and RMNH collections labeled as Gellius vagabundus remaining from the material studied by Vosmaer (1885), including several labeled as the var.γ, and even one specimen (nr.74) indicated by Vosmaer (1885: 29, pl. V figs 38-38) as having the spiculation depicted, were examined.No flagellosigmas were found, all specimens and fragments belonged either to Desmacella, Hymeniacidon or Hemigellius, with sigmas of normal shape, or lacking.There is one dried sample without identification in the ZMA collection, bearing only a small label with text 'Sp.XXX No. 76', the number given to Gellius vagabundus by Vosmaer (1885).The sample consists of three fragments, all of which are Hymeniacidon-like (with larger and smaller styles, as depicted in pl.V figs 32-33).However, the ZMA 20742 slide presumably made from the dried material does have the spicules depicted in Vosmaer's Pl.V figs 36-38.Although the number is 76, not 74, it is clear that this slide was made from a previously present dried fragment and is now is all that remains of Vosmaer's var.γ.
Both the Barents Sea slide and the Mauritanian specimen have been assigned to H. (F.) porosa based on spicule shapes, and the presence of a single category of normal sigmas that are characteristically rare.They resemble Fristedt's description of Desmacella porosa from Davis Strait, although the type specimen -listed to be present as Gellius porosus in the Zoologisk Museum Copenhagen under reg.nr.DEM 107 -was much larger (9 × 6 cm).The size of the oxeas was given by Fristedt as 350 µm, but Lundbeck (1902) re-examined the type and found some of them to be as small as 250 µm.Fristedt did not mention any normal sigmas, but Lundbeck (1902) found several normal sigmas.The length of the flagellosigmas was given by Fristedt as 120 µm across, somewhat larger than the ones of ZMA 20742, but smaller than those of ZMA 06624.Lundbeck (1902), in his description of a specimen from the N coast of Iceland, found oxeas and flagellosigmas in the same size range as those of the above described Barents Sea slide (ZMA Por.20742), and the normal sigmas were 50-80 µm, somewhat larger than those of the present material.Topsent (1896: 281, pl. VIII fig. 4) reported Gellius flagellifer from the nearby Gulf of Biscaye (depth 400 m), with oxeas 350 × 13-14 µm and flagellosigmas up to 90-100 × 2 µm.The drawing of the flagellosigmas closely resembles the present material.He neither discussed, nor figured the normal sigmas, suggesting that they were rare [he admitted to their presence later (Topsent 1904)].The rarity of the normal sigmas makes it likely that the material belongs to H. (F.) porosa.Lambe's (1896) record of Gellius flagellifer is probably also H. (F.) porosa for the same reasons.
Ferrer Hernandez (1918: 22, fig. 3) reported Gellius porosus from the coast of Asturias, N Spain, at 200 m depth (material originally collected by Orueta).His description and figure may be similar to the type of Fristedt.
In addition to H. (F.) porosa, Lundbeck (as well as other authors, e.g., Rezvoi 1928: 91) also described Arctic specimens assigned to Gellius flagellifer Ridley & Dendy, 1886.Lundbeck's specimens differ from his H. (F.) porosa in having larger oxeas (up to 476 µm long), and abundant normal sigmas in a large size range.For these and other reasons, it is unlikely that these Arctic specimens are conspecific with specimens occurring in the Southern Ocean Prince Edward Archipelago.There are subtle small differences in the sizes and shapes of the flagellosigmas and the normal sigmas when compared with the type of H. (F.) flagellifera (see above).As Lundbeck insisted that these specimens were not conspecific with H. (F.) porosa (several authors, e.g., Topsent 1896 and Lambe 1896 thought otherwise), and access to the specimens was not possible, they remain provisionally as Haliclona (F.) spec.until their status as a separate species from H. (F.) porosa can be established (but see also below).Brøndsted's (1933) species Gellius rhaphidiophorus from Greenland at 410 m depth, was described as close to Gellius porosus, with "more or less flagelliform" sigmas of 71-110 µm, in addition to rather rare sigmas of 20-36 µm and raphides of 36-40 µm.Species with flagellosigmas and raphides are otherwise not known, so possibly the raphides are foreign, in which case it could be a junior synonym of H. (F.) porosa.However, conspecificity is uncertain because the presence of true flagellosigmas cannot be verified due to the lack of illustrations and the ambiguous description.Likewise, many records of Gellius porosus and Sigmadocia porosa (cf.above in the historical overview) remain to be substantiated, as they were either not taxonomically described or insufficiently characterized.
The name combination Haliclona (F.) porosa is threatened by previous use of the name 'porosa' in combination with the unaccepted genus name Arcesios and the subgenus name Reniera.Arcesios porosa Duchassaing & Michelotti, 1864 is described unrecognizably and no original material is known to be extant in collections (cf.Van Soest et al. 1983).Schmidt (1870: 40) assigned this species to Reniera Schmidt, 1862, claiming in a two-line comment that Arcesios was a junior synonym of Reniera, but failed to provide any evidence.Schmidt stated he had a specimen from "Crabb Island" [sic] that answered to the descriptions of A. porosa.He gave no description of this material and no specimen from "Crabb Island" is kept in the collections of the Museum of Comparative Zoology, Harvard, or the Musée de Zoologie at Strasbourg, where most of Schmidt's 1870 material is housed.Although Reniera is a subgenus of Haliclona, there is no evidence that Arcesios porosa is a member of the subgenus.This fact precludes a definite and formal decision about the preoccupied state of the combination Haliclona (Flagellia) porosa and does not warrant proposal of a new name for it as a junior secondary homonym.
An additional homonym of the present combination is Reniera cinerea var.porosa Topsent, 1901.This was renamed Reniera topsenti Thiele, 1905 because of Schmidt's (1870) combination Reniera porosa (and was subsequently assigned to the genus Haliclona as H. topsenti by Burton (1940: 99).Because Schmidt's combination has priority, even if it is likely not the same species as Arcesios porosa, the name Reniera topsenti and the current combination Haliclona topsenti remain accepted.

Description
Small dirty white to greyish brown or greyish beige encrustations (Fig. 14A holotype, A1 paratype), often forming thick cushions or small globular masses, occasionally pear-shaped (grey in alcohol).Color remains unchanged in alcohol.Size variable from tiny, < 2 mm individuals up to 2.5 × 1.6 × 1 cm.Surface undulating to shaggy, sometimes 'hairy' due to protruding spicule tracts, and also appearing clathrate.There may be one or two oscules, only apparent in larger, thicker specimens.Consistency soft.

Remarks
The present deep-water North Atlantic specimens are overall very similar to the type of Gellius flagellifer.The one major difference is the occurrence of two distinct size categories of normal sigmas.Minor differences are thinner oxeas and larger size range of the two flagellosigmas' size categories in the present specimens.Topsent (1904) reported Gellius flagellifer from the Azores (845-1360 m).Oxea size ranges were from 335-345 × 8-10 µm in the more shallow station, to 620-680 × 18-20 µm at the deeper stations, suggesting a relationship between oxea size and depth.In the Azores material flagellosigmas reached sizes of up to 118 µm, and normal sigmas occurred in a large size range of 30-80 µm, suggesting the presence of size categories in both.It is possible that the Azores specimens conform to the Irish material, at least the shallower sample (but there is some doubt over the identity of the deeper sample, see below).
Remarkably, Topsent (1928: 314) took the erroneous view that his earlier reports on Gellius flagellifer were part of what he considered to be Gellius vagabundus (Schmidt, 1870), referring to Vosmaer as the inspiration for this change.In fact, Topsent's (1928) description of 'Gellius vagabundus' from 1378 m near São Miguel, Azores, is likely to be a different species from his other described specimens as the sigmas appear dissimilar to the flagellosigmas discussed here.Possibly, it is a Haliclona (Gellius) species, but it is not the present species.Lundbeck (1902) (and other authors such as Rezvoi 1928 andKoltun 1959) assigned specimens from Arctic waters to Gellius flagellifer, which could perhaps be members of the present species because Lundbeck's drawing (pl.XIV fig.1d) of the normal sigmas shows a large size range.However, without the original material this cannot be decided for certain.

Etymology
The specific epithet refers to professor Eduardo Hajdu (Museu Nacional de Rio de Janeiro, Brazil) in recognition of his many contributions to our knowledge of the sponge fauna of South America.

Description
Encrusting to irregular lamellar with oscular lobes (Fig. 15A-C).The three samples were obtained from the same station, but some were fragmented into small cm-sized pieces making it difficult to describe the overall shape in more detail.The specimen chosen as the holotype is basically an oscular lobe of 1.5-2 cm high and wide, with an oscule of 3 mm in diameter, the paratypes are fragments, partially overgrowing dead parts of associated organisms, including sponges.The surface is irregular, punctate.The color (in alcohol) ranges from shades of pinkish light or darker brown.The consistency is soft and fragile.

Remarks
Although the shapes of the flagellosigmas and the sigmas clearly differ from those of the above described H. (F.) hajdui subgen.et sp.nov., this material is not named here because it is too small to allow proper study.Only a few flagellosigmas were found, leaving open the possibility that larger spicules with upturned apices and a more elliptical outline might have been missed.Future study is necessary to delimit the characters of this species against those of H. (F.) hajdui subgen.et sp.nov.Haliclona (Flagellia) edaphus (De Laubenfels, 1930)  A thick plate-like mass (40 × 30 × 20 cm) encrusting stones in an intertidal cave.Color whitish in life and in alcohol.Surface smooth.Oscules of about 1 mm diameter are distributed evenly over the upper surface.Consistency firm to hard.
Skeleton.Dense, confused reticulation of thick oxeas.At the surface single spicules are arranged tangentially.

Distribution and ecology
California, near Carmel (Marine Ecoregion Northern California).Apparently confined to intertidal and shallow subtidal rocks.De Laubenfels mentions a second locality for this species, Point Fermin, near San Pedro (33.7054°N, 118.2938°W) (Marine Ecoregion Southern Californian Bight). of sigma length (which may be a misprint as it is the same as the upper size of the flagellosigmas), their data conform to the present description and are slightly different from De Laubenfels' original description.Lee et al. (2007) provided illustrations of the holotype and of its skeleton.They treated the name edaphus as an adjective (by adjusting the combination with the genus Xestospongia to edapha), but it is a noun from the Greek, meaning bottom or pavement.For that reason, the name edaphus in its original spelling is retained here.Dickinson (1945) reported this species from the (Mexican) Gulf of California (Carmen Island, approximately 25.94° N, 111.09°W) at a depth of 120 m.The oxeas of his specimen measured up Green & Bakus (1994) reported two specimens, one from a depth of 54-63 m and one from 200 m, collected from the Santa Maria Basin off the coast of Southern California.The descriptions are somewhat confused, and apparently the spicule measurements between the two specimens varied widely.Possibly, the specimen from 54-63 m could conform to H. (F.) edaphus, but the one from 200 m deep appears to be different as the authors gave 'sigma' measurements of 13-250 µm, which are not compatible with the sizes from H. (F.) edaphus.Assignment of these deep-water California records awaits proper redescription of the specimens.Characters with little or no differentiation among the species appear to be color and skeletal structure.All species have colors in shades of light brown or dirty white and skeletons are usually confusedpaucispicular.Some species have visible quantities of spongin, partially enveloping spicule tracts, notably H. (F.) indonesiae subgen.et sp.nov., H. (F.) hamata and to a lesser extent H. (F.) amirantensis subgen.et sp.nov.

Remarks
While detailed comparisons of the specimens available in this study reveal small differences between them, it is striking to note that specimens from widely separate localities exhibit a basically uniform skeletal structure and spicule complement.Assuming this basic morphology as evidence for monophyly, it is likely to mean that the subgenus Flagellia subgen.nov.has had a long history, possibly going back to Mesozoic times when interocean corridors were in place and continental platforms and slopes were generally interconnected.Such an age would appear to be in conflict with the status of subgenus and would justify raising it to genus level as ubiquitous distributions are more likely at the genus level.However, currently the status of the family Chalinidae requires a systematic review, including morphological taxonomy in parallel with molecular techniques, the outcome of which may affect the status of many of the genus names presently subsumed under the synonymy of Haliclona s.l.It is prudent to refrain from adding to the genus level content of the family and its order until a molecular tree of genera (based on sequences of their type species) is available.

Fig. 2 .
Fig. 2. 'Flagellated' sigma of Mycale (Naviculina) diversisigmata Van Soest, 1984, a sigma type superficially similar to a flagellosigma, but the morphology is considered non-homologous due to the lack of a strong asymmetry in the longer and shorter endings.
Meristic and descriptive features of the species assigned to specific taxa of Haliclona (Flagellia) subgen.nov.distinguished in the present study.The first line of flagellosigma measurements represents lengths of long and short endings, the second concerns widths and thicknesses.n.a.= not applicable, because of the absence of the spicule type.The first line of the description of the flagellosigma (only the largest (I) category) concerns the overall shape, the second concerns details of the long ending.Descriptive terms are explained in the text.Measurements are in µm.VAN SOEST R.W.M., Haliclona (Flagellia)  subgen.nov.among species, as seen in H. (F.) indonesiae subgen.et sp.nov.and H. (F.) hamata, where they differ in the thickness of these spicules.