Skeletons in confusion: a review of astrophorid sponges with (dicho–)calthrops as structural megascleres (Porifera, Demospongiae, Astrophorida)

Abstract We present a review of astrophorid species possessing calthrops megascleres as structural megascleres (including species with dichotriaene modifications, but excluding mesotriaene and trichotriaene bearing species). Radiating oxeas characteristic of most astrophorids are lacking in such sponges, but auxiliary oxeas are apparently present in some species. These sponges are currently assigned to two families, Pachastrellidae with four nominal genera Dercitus, Stoeba, Dercitancorina, Halinastra (the latter two generally considered junior synonyms of Stoeba), and Calthropellidae with nominal genera Calthropella, Corticellopsis (usually considered a junior synonym), Pachataxa and Pachastrissa. Our review of many original specimens and extensive new material demonstrates the existence of considerable morphological similarity in megasclere shape and arrangement, and possible overlap of some microscleres of these sponges: pseudasters in Dercitus resembling euasters in Calthropella, ataxasters and other modified types in Calthropellidae resembling rhabds in a species of Dercitus. Pachastrellid representatives are proposed to be united in a single genus Dercitus, recognizable as (dicho–)calthrops bearing sponges with sanidaster–like microrhabds. Three subgenera, Dercitus s.s., Stoeba and the revived Halinastra are distinguished to accommodate species groups sharing additional characters. Dercitancorina is merged with Dercitus (Stoeba), because the type species, Dercitus lesinensis was found to be barely distinct from Dercitus (Stoeba) plicatus. Similarly, the calthropellid representatives are proposed to be united in a single genus Calthropella recognizable as calthrops bearing sponges with oxyasters. Three subgenera, Calthropella s.s., Pachataxa and Corticellopsis are distinguished to accommodate species groups sharing additional characters. The calthropellid genus Pachastrissa is considered a junior synonym of Calthropella s.s. because its type species, Pachastrella geodioides, is barely distinct from the type species of Calthropella, Calthropella simplex. Two species previously assigned to Dercitus or Stoeba (Dercitus loricatus and Stoeba natalensis) are excluded from the genus Dercitus as they do not fit with the emended and improved definition of the genus. One species assigned to Calthropella, Calthropella digitata, is excluded from that genus and assigned to the ancorinid genus Stelletta. Based on the similarity of the megascleres and their structure-less arrangement, we predict that the two reviewed genera will be found to be closely related and eventually will need to be accomodated in a single family, but independent molecular evidence is awaited before changes in the family classification, including a verdict on the validity of the family Calthropellidae, will be proposed. Our review included 38 valid species among which fourteen species and one subspecies appear to be new to science. Four of these were represented by insufficient material for a proper description, but the remaining ten are described as new species: Dercitus (Stoeba) senegalensis sp. n., Dercitus (Stoeba) verdensis sp. n., Dercitus (Stoeba) fijiensis sp. n., Dercitus (Stoeba) bahamensis sp. n., Dercitus (Halinastra) berau sp. n., Dercitus (Halinastra) japonensis sp. n., Dercitus (Halinastra) arubensis sp. n., Dercitus (Halinastra) sibogae sp. n., Calthropella (Calthropella) xavierae sp. n., and Calthropella (Pachataxa) pyrifera sp. n. The new subspecies, Dercitus (Dercitus) bucklandi lusitanicus ssp. n. is described for southern East Atlantic populations of the nominal species. Several specimens assigned to existing species were found to be deviating to the extent that they are likely members of further undescribed species. These are briefly discussed to facilitate further studies of the diversity of the two genera. Species of both genera and the six subgenera, including deviating or insufficiently characterized specimens belonging to species not yet properly described, are keyed out. Distribution patterns are discussed.


Introduction
Th e present study addresses the taxonomy of some genera of the order Astrophorida (Dercitus Gray (1867) and Calthropella Sollas (1888) and their relatives), that appear or seem to share characters which do not justify their present allocation in diff erent families. Th e main character shared is the structural position of calthrops megascleres (and derivates thereof ) and the absence of structural oxea megascleres. Skeletons built from that megasclere complement lack a radiate structure making member species thickly or massively encrusting or insinuating. Such sponges are easily missed in trawl or dredge surveys and their diversity appears underestimated. Below, known members of these taxa will be reviewed and new members will be described and extensively illustrated.
A further purpose of this paper is to propose a slightly altered generic allocation of species, facilitating easier recognition of the morphological classifi cation, but the familial allocations will be left as they are in the Systema Porifera  until such time that suffi cient molecular systematic investigations have been made for a reshuffl ing of genera and families in the Astrophorida. Such studies are well on the way (Cárdenas 2010;Cárdenas et al. 2010), but face considerable problems of DNA extraction, especially from old type material, making the present study a timely contribution to astrophorid systematics. Before entering the systematic descriptions we present an overview of the current status of the target genera to explain why the generic allocations need to be improved. For explanation of the terms for spicules and other skeletal features the reader is referred to the Th esaurus of Sponge Morphology (Boury-Esnault and Rützler 1997), which may be freely downloaded from the World Porifera Database ( Van Soest et al. 2008).

Astrophorida possessing calthrops megascleres as structural spiculation
Astrophorida Astrophorida Sollas (1888) is a well-defi ned order of Demospongiae comprising sponges with the combination of triaene megascleres and aster microscleres, usually complemented with large oxea megascleres. Th e oxeas and long-shafted triaenes are arranged perpendicularly to the surface providing a radiate structure -at least in peripheral regions -which is shared with members of the order Spirophorida . Astrophorida and Spirophorida show a clear overlap with 'Lithistida' and future studies will undoubtedly demonstrate that this polyphyletic assemblage will have to be subdivided over these two and possibly other orders. Notwithstanding a general agreement over the non-lithistid Astrophorida as a well-founded ordinal taxon, family group distinction within this order remains tentative, with fi ve families currently defi ned, showing overlapping morphologies and lacking proper synapomorphies (Cárdenas 2010;Cárdenas et al. 2010), instead of which they are loosely defi ned on combinations of characters. Two such families seemingly showing overlap in spicular characters are Pachastrellidae Carter (1875) and Calthropellidae Lendenfeld (1907), as recently redefi ned by respectively Maldonado (2002) and . Both families encompass a few genera lacking the usual structural oxeas and long-shafted triaenes which prevail in most Astrophorida, although persistent reports for the occurrence of auxiliary or accessory oxeas of variable sizes are found in the descriptions of several species of these genera (see below).

Dercitus and its relatives
Th e monospecifi c genus Dercitus (for Dercitus bucklandi (Bowerbank, 1858) and its closely related 'sister'genus Stoeba are uncharacteristic Pachastrellidae, with none of the usual streptasters, long-shafted triaenes and long oxeas found in the majority of the genera. Th e acanthomicrorhabd microscleres are usually considered derivates of sanidasters, and will be further called as such below, but they are characteristically irregular and variable. Small incipient sanidasters appear to be microxea-like with only occasional spines. In addition to these microscleres, Dercitus possesses toxa-like spicules, which are lacking in Stoeba. Th is is the only known diff erence between these 'genera'. Many authors prior to the appearance of the Systema Porifera did not distinguish between Stoeba and Dercitus, because the occurrence of toxas in D. bucklandi was not considered of generic value. Th ere is only a single species with this unique character, and defi ning separate genera Dercitus and Stoeba is not possible without referring to the absence or presence of this same unique character (so called 'A-not A' classifi cation). Dercitus and Stoeba possess this limited set of calthrops and (dicho-)calthrops megasclere spicules, but some species allegedly have a complement of (rare) oxeas. For such species, the genus Dercitancorina was erected by Topsent (1902), type Pachastrella lesinensis Lendenfeld (1894), but the reexamination of the type specimen of P. lesinensis did not reveal the presence of oxeas (see below). Possession of oxeas in species of this group is controversial against the possibility that the often broken and variously sized, 'auxiliary' megascleres may be contaminations. A further genus Halinastra de Laubenfels (1936) was erected for Dercitus exostoticus (Schmidt, 1868 as Pachastrella exostotica) which has peculiar aster-like microscleres hypothesized to be compressed sanidasters (Maldonado 2002). It was synonymized with Stoeba by Maldonado (l.c.), but we demonstrate here that several more such species with 'compressed' sanidasters exist, making resurrection at the subgenus level a logical consequence. In summary, we propose to employ a modifi ed classifi cation of Dercitus s.l. into three subgenera: Dercitus (Dercitus), Dercitus (Stoeba) and Dercitus (Halinastra).

Calthropella and its relatives
Calthropellidae (with type genus Calthropella) as recently defi ned have only four recognized genera ( , but there is considerable confusion over the generic defi nitions, which may possibly overlap. At least one genus, Chelotropella Lendenfeld (1907), appears misapplied as a member of Calthropellidae because the family is based on the possession of short-shafted triaenes in combination with euasters, while Chelotropella has long-shafted triaenes and in combination with oxeas forms a radiate skeleton similar to genera of the Ancorinidae such as Stelletta. A further genus that is in dispute is Pachastrissa , based on a type species, Pachastrella geodioides Carter (1876), that seems to diff er from the type species of Calthropella (C. simplex Sollas, 1888) only in the alleged presence of proper oxeas in addition to habit, skeletal structure and spiculation identical to that of C. simplex. A fourth genus, Pachataxa de Laubenfels (1936) will be demonstrated below to be so close to Calthropella that it is proposed to merge it with that genus. It is defi ned as having in addition to euasters peculiar irregular microspined silica-bodies, called ataxasters. SEM studies of these ataxasters demonstrate that they are not far removed from hypersilicifi ed aster-derived microscleres found in several Calthropella species. Corticellopsis Bergquist (1968) was erected to replace the preoccupied Corticella Sollas (1888) with type species Corticium stelligerum Schmidt (1868) and subsequently synonymized with Calthropella by  because of similarity of spicule complement (calthrops and euasters). However, this similarity is mostly on paper as its euasters diff er clearly in shape from those of Calthropella simplex and C. geodioides: the former has spherasters with tuberculate rays, while the latter has commonplace strongylasters. As a consequence of these observations, the calthropellid genera Calthropella and Pachataxa are proposed to be recognized only at the subgenus level, to which the formerly synonymized Corticellopsis is joined as a third subgenus, resulting in the taxa Calthropella (Calthropella), Calthropella (Pachataxa) and Calthropella (Corticellopsis).

Similarities of Dercitus and Calthropella
Th e calthropellid Calthropella (Pachataxa) with peculiar ataxaster microscleres and several species of the pachastrellid genus Dercitus with compressed aster-like sanidasters or smooth or faintly acanthose microrhabds (see below) appear to some extent to bridge the gap between the two family groups. Euasters of Calthropella (Pachataxa) appear peculiar in having a thick centre and numerous short and irregular rays, whereas the ataxasters are often malformed and irregularly spined, occasionally smooth.
Dercitus and its relatives (including Stoeba, Dercitancorina and Halinastra) could be envisaged to reside more comfortably in the same family as Calthropella as a genus devoid of true euasters, but otherwise similar to other genera of that family. Current assignment to Pachastrellidae is not clearly warranted by pachastrellid apomorphies, as the microsclere complement of sanidasters is of astrophorid-wide occurrence including the Ancorinidae. We pose here the question, whether there is suffi cient evidence to suggest these genera should be united within the same family. Ultimately, such a decision should be based on independent molecular and morphological evidence, so we refrain from making formal proposals for change and for the time being the family assignment of the genera is left as it appears in the Systema Porifera (Maldonado 2002; .

Triptolemma and Th rombidae
We will not include in our study the pachastrellid genus Triptolemma de Laubenfels (1955) and the thrombid genera Th rombus Sollas (1886) and Yucatania Gomez (2006), despite the fact that these genera all have short-shafted triaenes and lack proper radiating oxeas. Triptolemma is similar to Dercitus and Calthropella in having calthrops in a large size range and the insinuating habit conforms to that of some species of Dercitus. It is not considered very likely that Triptolemma is closely related as it has a dominance of mesotriaene (dicho-)calthrops and its microsclere complement includes true streptasters (amphiasters, spirasters). Th rombus and Yucatania have their (tricho-) triaenes entirely spined and microscleres are peculiar birotule-like amphiasters.

'Phantom' oxeas
Several species of Dercitus s.l. and Calthropella s.l. have been persistently attributed with the possession of long oxeas as part of the megasclere complement. Th e report of smaller or longer diactines (oxeas, strongyles) in species of Dercitus (Stoeba) is a recurrent point of discussion. Authors either consider them proper but so far have failed to indicate a structural position of such spicules in the skeleton, or they are presumed contaminations. In favor of the latter point of view is that most Dercitus s.l. are either insinuating or consolidating, thus coming into close contact with sediment including loose spicules. Diactines are ubiquitous spicules, produced mostly by Haplosclerida and Halichondrida. Longer oxeas have frequently been mentioned in descriptions of Dercitus and Calthropella alike and such spicules are common in Astrophorida in general, so are more likely to be proper. Nevertheless, it is clear in most cases that they do not take up a structural position in the skeletal architecture, so if they are to be considered proper, they are at best 'auxiliary' and are here regarded as vestiges of their astrophorid ancestry. In many cases, authors expressly mention that the long oxeas are invariably broken or rare, which strengthens the assumption they are not proper.
Species with structural oxeas in a radial position are here excluded and referred to other astrophorid genera.

Biogeography
A recent paper by Moraes and Muricy (2007), describing a new Stoeba from Brazilian waters, suggested in a world distribution map of species of Stoeba that records of the genus would be absent from the Caribbean region. Strictly speaking this was true, but there were previous records of Dercitus sp. from various parts of the Caribbean (Kobluk and Van Soest 1989: Bonaire;Djura andFaulkner 1980 andRützler et al. 2000: Belize;Kohmoto et al. 1988;Burres et al. 1989: Bahamas). Th e new classifi cation of Dercitus s.l. (including Stoeba) will thus remove this distributional anomaly (lack of a circumtropical taxon from the Caribbean region). Similarly, the new assignments of calthropellid genera result in a continuous circumglobal distribution of Calthropella s.l. whereas this was formerly confi ned to an irregular circum-African pattern plus an outlier in New Zealand (Calthropella s.s.) and a disjunct Caribbean -New Caledonian occurrence (Pachataxa).

Contents of the present study
Below, we review the species of the (sub-)genera Dercitus (Dercitus), Dercitus (Stoeba), Dercitus (Halinastra), Calthropella (Calthropella), Calthropella (Pachataxa) and Calthropella (Corticellopsis) based on lists obtained from the World Porifera Database (Van Soest et al. 2008). Although our paper has monographical ambitions the current state of our knowledge of these sponges forces us to refrain from making comprehensive descriptions of all extant species. Almost invariably, species delimitation becomes problematic when more than a single specimen of a nominal species has been described. For this reason, a fairly large number of suspected species remain unnamed and inadequately described, whereas some species suspected to be synonymous with others remain accepted. In addition, poor material of several species and diffi culties in obtaining all relevant type specimens also contribute to this study failing to reach a full blown revision of Dercitus and Calthropella. We treat here a total of 38 species, 24 species of Dercitus s.l. (and one new subspecies), and 14 species of Calthropella s.l.

Material and methods
Specimens in the collections of the National Centre for Biodiversity (formerly Zoological Museum of the University of Amsterdam (ZMA) and National Museum of Natural History, Naturalis (RMNH), Leiden) were obtained by various expeditions and individual collectors over a large period of time. Additionally, type and other speci-mens were borrowed from several institutions, including the Museum für Naturkunde, Berlin (ZMB), the Smithsonian Institution, Washington (USNM), Museum of Comparative Zoology, Harvard (MCZ), the Natural History Museum (London) (BMNH), the Muséum National d'Histoire Naturelle, Paris (MNHN), the Landesmuseum Graz, Austria (LMJG), the Zoological Museum of Copenhagen (ZMUC), the Harbor Branch Oceanographic Institution, Fort Pierce, USA (HBOI), and the Museu Nacional de Rio de Janeiro (MNRJ). Th rough the courtesy of Dr. Joana Xavier (Research Centre in Biodiversity and Genetic Resources, CIBIO, Department of Biology, University of the Azores) we were able to study specimens in her care from Portuguese shallow-water and deep sea localities. Full specimen data are provided with the description of the species.
Specimens were studied from thick sections cut at right angles to the surface and from dissociated spicules, using light microscopy and a JEOL Scanning Electron Microscope. Digital images of the spicules were assembled on a black background and aligned and cleaned up using Adobe Photoshop CS3.
Th e left over part of the spicule suspension was used for light microscopy measurements. Measurements of megascleres include for calthrops cladus length × width, cladome size, i.e. the distance between the apex of a cladus and the imaginary line between two opposing cladi endings; for dichocalthrops, protocladus length × width, deuterocladus length × width, rhabdus length × width and cladome size. Measurements of microscleres include for sanidasters length × width including the spines, for toxa-like microscleres length × width, for euasters greatest diameter including rays. Unless otherwise stated, minimum-mean-maximum from 25 of each spicule type encountered are given. To facilitate comparisons of cladome sizes of our own spicules and those from the literature, in case cladome sizes were not provided by the original authors these were artifi cially standardized by multiplying cladus length ×1.5 (calthrops) and the sum of protocladus and deuterocladus length ×1.9 (dichocalthrops), based on average ratios obtained from random measurements of various specimens. Such artifi cially standardized data are indicated by an asterisk.
In the descriptions and defi nitions, the word calthrops will be used in singular and plural sense (in accordance with Sollas 1888: lx). Th e Th esaurus of Sponge Morphology (Boury-Esnault and Rützler 1997) erroneously referred to the singular of this spicule as 'calthrop', possibly because  proposed 'chelotrop' as term for the same spicule. Th e dichotriaene modifi cation will be termed dichocalthrops below.
Comments: Many authors have pointed out the similarities of Dercitus and Stoeba, the distinction of which rests entirely on the presence of unique toxa-like microscleres in the type species of the former. Th e similarities are the shape and size of the calthrops/short-shafted triaenes, the shape of the irregular sanidasters, the absence of any structure in the skeleton, the compressible liver-like texture, and the presence of large cells (60 μm) with inclusions reported for various Stoeba species as well as Dercitus bucklandi. Th e majority of past authors recognized only Dercitus, but Maldonado (2002) insisted on retaining the distinction at the genus level. Since there are dozens of species conforming to the above given defi nition and only a single North Atlantic species possesses the toxas, it appears unnecessarily formal to keep the use of two genus names for such similar species. It also would confound biogeographical analysis having a single unique character place an endemic genus in an area of the world (Celtic Seas and Lusitanian waters) where such higher taxa endemism in sponges is virtually nonexistent. Ultimately, we need independent molecular evidence to demonstrate that both are members of the same clade, but pending such results, we propose here for practical reasons to lower the status of the genera Dercitus and Stoeba to the level of subgenera. Dercitus being the senior name, the subgenera will be Dercitus (Dercitus) and Dercitus (Stoeba). A third subgenus proposed is the suppressed Halinastra de Laubenfels (1936). Th e subgenera will be defi ned below and are keyed out in Fig. 1.
Calthrops are characteristically variable in size (cladi measuring from 25 -648 μm), and shape, with conical-straight, curved, stunted, deformed, bifi d cladi, frequently one of the cladi being longer than the others, sometimes lacking one ('tripods') cladus or with one or more reduced cladi, occasionally with fi ve or more cladi. Short-shafted dichotriaenes (dichocalthrops) are present in twelve of the currently named and recognized species, and in fi ve of these have replaced the calthrops entirely. When both megasclere types are combined in a single species, dichocalthrops are often distinctly smaller than the calthrops and the proportion of the two varies among individuals of the species.
Key to the subgenera of Dercitus (see Fig. 1 Description (amalgamated from various descriptions of material from the British Isles and NW France). Cushion-shaped to massive-lobose (Figs 2A, 3A), fi lling crevices in vertical rock faces. Size frequently over 5cm 2 . Black to dark greybrown externally. Surface smooth but often has ridges looking like stretch-marks. Th e surface is usually concave. Apertures (oscules?) are fl ush with the surface, variable in size and usually collected into groups towards the centre of the sponge. Consistency moderately fi rm but compressible and spongy when in situ, liver-like in preserved condition.
Skeleton: Main skeleton of confused, randomly arranged calthrops, a layer of sanidasters and many toxas occur near the surface, but these also occur in the choanosome. Th ere are many large pigment cells (?) with black inclusions evident in sections. Habitat. In crevices in vertical rock faces in clean water. Particularly common on limestone substrata. From the extreme low water mark to a few meters subtidally in rock pools and caves.
Distribution. British Isles (SW coast of England and N and W coast of Ireland), and France (Roscoff , Iles de Glénan). Th ere is a record from the northern Gulf of Mexico (Teerling 1975), but this has not been verifi ed and is unlikely to be correct.
Etymology. Named after Mrs. Buckland of Guernsey, who collected one of the type specimens (in fact the species should probably have been named "bucklandae" since the named person is female).
In view of some of the discrepancies we considered it worthwhile to investigate whether spicular diff erences of specimens within the known range of the species would yield a pattern that could explain some of the diff erences. We compared spicular data from the northern samples cited above with those of samples from areas in the southern part of the range, from the coasts and off shore localities of Portugal obtained from Dr Joana Xavier and examined spicule sizes and shapes. Th is led us to the conclusion that there are consistent diff erences between samples from these areas which should be recognized at the subspecies level. Dercitus (Dercitus) bucklandi subsp. lusitanicus ssp.n. urn:lsid:zoobank.org:act:9152BD64-067E-4F21-AF03-6B84E671E2A6 Figs 4A-E, 5f-j, Table 1 Dercitus bucklandi; Ferrer-Hernandez, 1918: 17;Rodriguez Babio & Gondar, 1978: 34;Acuña et al. 1984(in Solórzano, 1991; Templado et al. 1986: 96 ( Description (Fig. 4A). Alive blackish outside and greyish inside (in alcohol chocolate brown throughout). Shape massively encrusting, fl at with smooth surface, no apparent oscules; inhalant openings in sieve plates. Consistency cheesy, slightly compressible. A representative size of preserved specimens is 5×4×0.7 cm.
Skeleton: a layer of microscleres overlying a loose mass of calthrops megascleres. Spicular density appears lower than in the specimens from the British Isles.
Habitat. Encrusting, typically bridging crevices and gaps in the substrate, sublittoral down to 6-38 m.
Remarks. Th e spicule measurements of the present specimens are generally signifi cantly smaller than those provided by most previous authors (Table 1), which is especially clear in the sizes of the toxiform microscleres: 51-111 × 1.5-3 μm in the subspecies bucklandi, 42-69 × 0.5-1.5 μm in the samples of the new subspecies, compared in Fig. 5: with bucklandi bucklandi toxas in Figs 5a-e and bucklandi lusitanicus ssp.n. toxas in Figs 5f-j. Possibly, the southern samples may be distinguished at the species level from the northern samples, but in view of the fact that there is an overall strong similarity with Dercitus bucklandi, we prefer to recognize only subspecies.
Th e toxa-like microscleres of Dercitus bucklandi s.l. are unique in the Astrophorida. Th ey have been called 'toxa' because of the similarity in shape to similar microscleres in microcionid and mycalid Poecilosclerida. However, the resemblance is probably superfi cial and it is unlikely that these are homologous spicule types. Th ere are subtle diff erences mostly only clearly visible in SEM images, such as the tendency to become apically swollen, and an overall faint 'polytyly' (Fig. 4D), and -at least in the Bowerbank type material of the nominal subspecies -the not infrequent presence of peculiar side branches or single long spines near the curved part in the middle. Possibly, this indicates that the microsclere derives from an end to end fusion of two incipient smooth microxeas -a more likely assumption than Topsent's (1895) suggestion that the toxas are modifi ed asters -, but fi rm proof for this hypothesis is wanting. Smooth microxeas are relatively common in several ancorinid genera. Dendy, 1905 Defi nition: Dercitus with a single microsclere category in the form of irregular sanidasters.

Subgenus Stoeba
Type species: Samus simplex Carter, 1880.    Remarks. Th e exact properties and variation of this species have not been established so far, so we can only provide a discussion pending a proper revision. Th e type material has not been reexamined and its precise whereabouts are uncertain. Annandale (1915) described specimens from the Mergui Archipelago (now Myanmar) as similarly 'excavating' (they are insinuating dead corals and fi lling spaces presumably made by clionaid sponges) and having dichocalthrops, and a single calthrops, as megascleres (sizes not given); sanidasters 16.2 μm.
If all these specimens are members of a single species, then it occurs on both sides of the Gulf of Bengal as well as on the islands in the middle of it and to the east into Indonesia. However, specimens need to be reexamined.
Additionally, we observed some oxeas of uniform size, approximately 150 × 3-4 μm, assumed to be foreign to the sponge.
Description of the BMNH Schmidt's type slide (Fig. 6D, Table 2). Th e spicules (calthrops, dichocalthrops and sanidasters) are dissociated, but even in that condition it can be concluded that the slide is almost certainly taken from the holotype as the frequencies of occurrence and the sizes of the spicules are very similar to those of the holotype. Calthrops dominate the megascleres (Fig. 6C); most are regular, but occasionally some are three-claded or rarely two-claded. Endings of the cladi may occasionally be abruptly bent, bifi d, indicating incipient dichocalthrops. Th ey occur in a large size range, cladi 36-183 × 4-28 μm, cladomes 62-258 μm, almost identical in range to the spicules measured from the holotype. Dichocalthrops rare, as the slide contained only four measurable dichocalthrops, protocladi 15-21 × 3.5-10 Th e only non-type specimen available to us, ZMA Por. 15101 from Banyuls has spiculation closely similar to the type material (Table 2). Only the sanidasters appear on average slightly longer and more robust.
Habitat. Encrusting and insinuating in crevices, large depth range down to 100 m. Distribution. Originally reported from Algeria. Elsewhere reported with certainty from Banyuls, Naples and the Adriatic. Possibly some of the records from the adjacent North Atlantic (Lévi and Vacelet, 1958;Boury-Esnault and Lopes, 1985) also refer to the present species, but see below.
D. (S.) plicatus is apparently quite variable and this may have led to widespread reports of the species from various East Atlantic localities, but also from Malaysia (Sollas 1902), Gulf of Mannar (Th omas 1970), Maldives (Calcinai et al. 2000), and Fiji (Tendal 1969).
We believe alleged records of this species outside the Mediterranean (Table 2) need to be reviewed critically on the basis of reexamination of the specimens. We have done so for material available to us from West African and Fijian localities (see below). In these specimens we found important diff erences with Mediterranean Dercitus (Stoeba) plicatus, which led us to assign them to three new species. It is likely that further records from e.g. Indo-West Pacifi c localities which we could not verify are part of a complex of D. plicatus-like sister species distributed over most of the warmer parts of the oceans.
Several species of Dercitus (Stoeba) were synonymized with the present species by various authors, including the type species D. (S.) simplex. In most cases these synonymies are not accepted by us. Description. Encrusting and insinuating among stones and coralline algae, with oscular elevations. Th e fragment examined by us was a thin rounded papilla-like extension (probably from the larger of the two type specimens described by Lendenfeld). Consistency slightly rubbery. Colour (alcohol) orange.

Dercitus (Stoeba) lesinensis (Lendenfeld, 1894)
Skeleton: at the surface consisting of a dense crust of microscleres, subdermally the skeleton consists of a mass of calthrops, embedded in a fi brous-organic groundmass.
Habitat. No data. Distribution. Adriatic. So far there is only a single record of this species.

Remarks.
A major diff erence with the original description is the lack of giant oxeas (4000 × 70 μm), some centrotylote, reported and pictured by Lendenfeld (1894); also tylostyles and strongyles were reported. Th e fragment of the type specimen received from ZMB did not contain any such spicules, except for some broken fragments. For Maldonado (2002) the alleged presence of oxeas in a species otherwise considered by him a typical Stoeba was the reason to accept species with structural monaxone megascleres within the defi nition of Stoeba. We maintain that such spicules, if they would be proper, are auxiliary, not structural.
Th is species is similar in most respects to D. (S.) plicatus, but the orange colour has not been reported for that species (although Pulitzer-Finali (1983) mentions brown specimens of D. (S.) plicatus). Based on our observations of the sanidasters, it is possible that these are slightly diff erent from D. (S.) plicatus in spination, with the latter having more profused spines. In view of this, we are forced to retain D. (S.) lesinensis as a valid species for the time being.
Description (from Sarà 1959). Numerous specimens (but only one was preserved, the holotype GG907, present whereabouts unknown) in a cave, depth 0-7 m, 30 m from the entrance. Whitish encrustations fi lling interstices between oysters and barnacles. Individual size 1-6 cm2, total size of all specimens 43 cm 2 .

Remarks.
Because no material has been examined and previous authors assigned this to Stoeba we retain this as a valid species of Dercitus (Stoeba) for the time being on account of the reported oxeas. However, apart from the oxeas the description perfectly fi ts that of Dercitus (Stoeba) plicatus (Schmidt, 1862) (see above). Especially the white colour, the Mediterranean occurrence and the combined presence of dichocalthrops and calthrops are telltale signs that they could be conspecifi c.
Etymology. Named after the locality of the holotype. Habitat. Consolidating coarse limestone fragments on rubble bottoms. Distribution. Senegal, Mauritania, possibly (Eastern) Mediterranean. Remarks. Th e new species was formerly identifi ed as Dercitus (Stoeba) plicatus, but it diff ers clearly by the much larger upper size of the calthrops and the apparent absence of dichocalthrops. Th ere are some discrepancies between the two type specimens in the average size and thickness of the spicules, but in view of the generally large size variation this is considered intraspecifi c variation. It is likely that Lévi's Senegal record of D. plicatus also fi ts with this species, as it was similarly white coloured, and also lacked dichocalthrops. Th e size of the cladi of the calthrops were given as maximum 250 × 30 μm which is clearly larger than in Mediterranean D. plicatus and within the range of our new species. Th e ZMA collection contains a tiny fragment of a specimen from the Aegean Sea d onated by E. Voultsiadou (ZMA 21697) which has calthrops ( Fig. 9E) with cladi up to 295 × 48 μm (cladomes up to 475 μm), and which lacked dichocalthrops. Sanidasters similar in size and shape to those of D. (S.) senegalensis sp. n., especially to the paratype. Apart from the wide geographic separation, this fi ts with our new species.
Th e record of Pansini (1987) of Dercitus plicatus from the Alboran Sea possibly also belongs to this species as it was noted to lack dichotriaenes. However, since no further data were provided this remains undecided.
Etymology. Named after the type locality.
Habitat. Deeper water, consolidating coarse sediment. Distribution. So far with certainty known only from the type locality. Possibly, Topsent's 1928 record of D. plicatus from Boavista also belongs to this species.
Th e two slides contain thin cross sections of the peripheral regions. Skeleton (Fig. 11A): irregular mass of short-shafted triaenes covered by a thin ectosomal crust of microscleres.
It is likely that the record of Halina plicata of Th omas (1970) belonged to the present species as most descriptive remarks fi t except for the upper size of the cladome ('chord') of the dichocalthrops which was quoted as 330 μm. Vacelet and Vasseur (1971) reported this species from Madagascar. Live colour was noted as black, which was maintained in alcohol. Dichocalthrops measurements provided were: protocladi 20-40 μm, deuterocladi 60-80 μm, rhabds 150-180 μm. Sanidasters were 25-30 μm. No mention of calthrops. No further spicule data. The colour difference, lack of calthrops, and slight differences in spicule sizes and shapes render it doubtful whether this material belongs to D. (S.) extensus. Description. According to Tendal (1969) the sponge formed an irregular mass of 2×5×4 mm, situated in a crevice in a dead piece of coral. Smooth surface. Colour dark grey, oscules and pores not visible. Consistency hard. Th e material available to us consisted of a tiny fragment, approx. 1 mm 3 , half of which was sacrifi ced for SEM and a residue spicule slide.
Skeleton: unknown but presumably confused, with an ectosomal cover of microscleres.
Etymology. Named after its type locality.
Habitat. Shallow-water, in crevices in dead corals. Distribution. Known only from the Fijian type locality.

Remarks.
Th e distinctive feature of this new species, apart from details of spicule sizes, is the characteristic shape of the larger dichotriaenes with very short protocladi and long conical deuterocladi (represented most clearly in the dichocalthops at the right of Fig. 12B). Tendal (1969) assigned this material to Dercitus (Stoeba) plicatus (as Halina), but this Mediterranean species has much smaller dichocalthrops (cladomes only up to approx. 200 μm) in a lower quantity than the calthrops, and is white in colour. Tendal's measurements of the spicules diff ers slightly from our measurements. We did not see any sanidasters thinner than approx. 3 μm, whereas Tendal mentions 1-4 μm in thickness and protocladi of the dichocalthrops were never longer than 30 μm, whereas Tendal gives 24-66 μm. Possibly, the limited size of our fragment and the few spicules we could measure explains the diff erences. Even with Tendal's measurements included, the diff erences with Dercitus (Stoeba) plicatus in the size of the dichocalthrops and calthrops remain clear. Furthermore, Tendal stressed that the deuterocladi are always longer than the protocladi and gave an excellent drawing of the characteristically shaped dichocalthrops. Th us the diff erences with D. (S.) plicatus are distinct and unambiguous.
Spicules: calthrops, dichocalthrops, sanidasters. Some spicule types are remeasured to provide additional data and to correct a remarkable error in the original description.
Sanidasters: short and slim, with blunt apices. Remarkably, they were quoted to be 15-74 × 2-8 μm in size, but this is an obvious error. Remeasured they appear to be highly uniform in size and shape: 9-12.8-16 × 1-1.9-2.5 μm, which is also in accordance with the size and shape of the fi gured sanidaster (fi g. 52).
Habitat. Shallow water. Distribution. Only known from the Galapagos Islands.
Remarks. By its large calthrops this species stands out among Dercitus species with a complement of small dichocalthrops. None of the extant species are morphologically close. Description. Slippery smooth encrustation on dead coral (Fig. 13A); the present material (upper part of Fig. 13A) consists of three fragments of 12×10, 8×8 and 8×5 mm, of approx. 1-2 mm in thickness. Th e holotype (lower part of Fig. 13A) is a larger fragment of 7×6 cm (not examined). Colour bright red when alive, darker red in alcohol. Consistency gum-like, easy to cut.
Etymology. Named for its type locality.

Remarks.
Th e red colour is shared with D. (S.) bahamensis sp. n., but this has dichocalthrops and the spicule sizes are also diff erent. Below we report an undescribed Dercitus (Stoeba) spec. from Bonaire for which insuffi cient material was left after making preparation. It diff ers in habit (insinuating) and spicule size (calthrops cladi smaller, only up to 186 ×16 μm, and sanidasters slightly larger, up to 18 × 1.5 μm). Cladi of calthrops of the Bonaire material are frequently bifi d and some ?auxiliary oxeas were present in the slides. Nevertheless, similarities are close enough to reckon with the possibility that the two are conspecifi c members of a variable species.

Remarks.
De Laubenfels (1932) mentions the possible presence of toxas, judged to be foreign. Lee et al. (2007) provide SEM photos of the spicules. Th e species stands out by its small calthrops, which is only shared with D. (S.) xanthus, but in that species there are no normal calthrops, only three-claded ones, and the sanidasters are considerebaly longer. Spicules: three-claded calthrops and sanidasters. Calthrops small, divisible in two size classes with means of approximately 25 and 72 μm. Only 20% of the 163 specimens recorded possessed calthrops, usually in high densities, in the remaining 80 % these spicules were lacking. Sanidasters universally present in all specimens, displaying a wide variation in length and width, 10-20 × 1-2.5 μm, densely spined with relatively short spines up to 1 μm.

Dercitus
Habitat. Sandy bottoms between 16 and 86 m depth. Distribution. Great Barrier Reef, occurring over the entire range.
Description (from Vacelet and Vasseur, 1971). Encrusting the underside of coral blocks, 2-3 cm 2 in expansion; consistency soft; live colour yellow, white in alcohol.
Habitat. Shallow reefs. Distribution. Known only from the Tuléar area, Madagascar.

Remarks.
Live colour and the details of the spicules make it likely that this is an undescribed species. Th e strongyles are considered foreign as the authors themselves suggested.

Dercitus (Stoeba) spec. Bonaire
Dercitus sp. Kobluk & Van Soest, 1989 Description. Insinuating in coral crevices, in reef caves between 10 and 40 m depth. Th e available sponge material has been 'used up' entirely for the initial slides from which the genus identifi cation (as Dercitus) was made. Properties of the sponge and structure of the skeleton cannot be accurately described. Th e skeleton is a confused, dense mass of calthrops megascleres on the outside covered by a layer of microscleres. Th ese are also common throughout the sponge body. Spicules are simple calthrops and sanidasters.
Habitat. In reef caves at the deeper parts of fringing reefs. Remarks. Kobluk and Van Soest (1989) reported a few oxeas of 600 μm length, which are here considered foreign, and 'dichotriaene-like variations' by which they meant the calthrops with bifi d cladi, not to be confused with true dichocalthrops. Since no proper specimen was left, no SEM observations of the spicules of this species could be made. It is likely that it is an undescribed species. It diff ers from Central West Atlantic D. (S.) latex (Moraes & Muricy, 2007) in habit (encrusting in latex), and spicule size (calthrops cladi larger, up to 212 × 25 μm, and sanidasters smaller, up to 15 × 1 μm). D. (S.) bahamensis sp. n. likewise is encrusting and it has dichocalthrops and thicker sanidasters (see above).
Type species: Pachastrella exostotica Schmidt, 1868 (by original designation). Description. Black mass (in alcohol) of 3×3×0.5 cm (Fig. 14A). Surface irregular granulated, grooved, no apparent oscules. In cross section, only the surface region is darkly pigmented, but the choanosome is much lighter coloured. According to Keller choanocyte chambers are crowded in the lower parts of the choanosome, size 25 μm.

Key to the species of Dercitus (Halinastra)
Skeleton: a confused, dense mass of calthrops with a cover of microscleres at the surface. A few oxeas of widely diff erent sizes were noted by Keller, but these were limited between approx. 100-120 × 2.5 μm of a typical haplosclerid type, in the slide examined by us.
Habitat. Deeper water, beyond the reefs Distribution. Known only from the Southern Red Sea. Remarks. Schmidt (1868) gave hardly any description (some vague remarks on the megascleres of Pachastrella monilifera grading into a statement on the Red Sea specimen), but the drawings of the microscleres show a sanidaster and a globular aster. Keller claims to have had access to Schmidt's type specimen, but he faithfully reports that there is a discrepancy between Schmidt's original label and the label of his redescribed specimen, e.g. the original collector was stated to be Ehrenberg by Schmidt, but Keller states it is Siemens, which is probably the correct one. Maldonado (2002) assigned Pachastrella exostotica to the genus Stoeba and declared Halinastra a junior synonym of that genus. However, he possibly overlooked that Pulitzer-Finali's (1986) Pachataxa lutea possesses similar spiculation as P. exostotica (see below). With at least two species sharing the peculiar aster-like compressed acanthorhabds, (and several further species with diversifi ed aster shapes, see below) there is good reason to maintain Halinastra as a distinct taxon, proposed here to be a subgenus.  Description. Blackish grey lobate mass (Fig. 15A), approx. 20×15×10 cm in size, surface in places speckled white by ?coral sand and at the base the specimen is encrusted by a bluish sponge (Haliclona). Oscules prominent on the summit of lobes, approx. 0.5 cm in diameter. Consistency fi rm, compressible. Th e preserved specimen is broken into two similar sized masses (Figs 15B, 16A). Th e black colour is maintained in alcohol and has strongly darkened the fl uid and the labels.
Skeleton: diffi cult to study in the preserved condition due to the intense black colour, but structure is dense and unorganized, with a dense cover of microscleres at the surface.
Etymology. Named after the Berau region, East Kalimantan.
Habitat. Shallow open reef localities. Distribution. So far known only from the type locality E of Kalimantan. Remarks. Despite the long distance between the locality of D. (H.) exostoticus (Red Sea) and the present locality, there are compelling similarities in shape, colour and spicule sizes between the new species and exostoticus. Th e major diff erence between the two is the length and robustness of the sanidasters (up to 42 μm in D. (H.) berau sp. n. against max. 27.5 μm in D. (H.) exostoticus). We also believe that the habit and the choanosome pigmentation will prove to be distinct in the  Description. Holotype (Fig. 17A) is a blackish brown fragment of 3.5×3×1 cm (live colour also noted as black). Surface lumpy but smooth, shiny. Consistency fi rm, compressible.
Skeleton: at the surface a crust of microscleres, overlying a dense confused mass of calthrops.
Etymology. Named after the Japanese type locality. Habitat. Dredged from deeper water, at approx. 150 m. Distribution. Both specimens came from nearby locations, so the distribution so far appears limited to off shore Japan.
Remarks. It is with some hesitation that these two Japanese sponges are assigned to a new species, rather than to D. (H.) berau sp. n. Th e two species appear generally similar, but sanidasters are shorter and slimmer than in D. (H.) berau sp. n. and the compressed sanidasters are few and consist mostly of oval rather than globular shapes. Combined with the clearly deeper occurrence and the considerable geographic distance we believe it is justifi ed to consider both distinct at the species level. Th e two are obviously closely related to D. (H.) exostoticus (see Table 3) with which they form a species complex. Description. Th e larger specimen (10100) is a thick liver-like mass (Fig. 19A) with fl at smooth surface, 8×6×3 cm in size, consistency compact, cheesy to rubbery. Live colour bright yellow, in alcohol dark brown. Th e preserved specimen shows a darker outer rim of 1-2 mm in thickness over a lighter coloured choanosome. Th is is not evidence of a cortex, it merely shows that the post-mortem discolouration does not penetrate evenly throughout the sponge. Our specimen conforms in all macroscopical aspects to Pulitzer-Finali's type. Th e second specimen is smaller and thinner, approx. 1×1×0.2 cm, dark glistening red, with a slightly bumpy surface. Th e Brazil specimen is sized 4×4×4 cm, cut off from a much larger sponge.
Skeleton: the ectosome has a dense layer of sanidasters and 'spherasters' carried by a similarly dense mass of calthrops. In the choanosome the calthrops and sanidasters are less densely and confusedly distributed. Th ere is no zonation or other structural arrangement of the skeleton.
Compressed sanidasters (Fig. 19D) are peculiar, irregular, multirayed, with a very thick center and very short rays, which appear spine-like rather than issuing from a common center, 5-7.7-9 μm.
Habitat. Open substrate in deeper parts of the reef, more shallow in caves. Distribution. Widespread Central West Atlantic: Puerto Rico, Southern Caribbean, Northeastern Brazil.
Th rough the courtesy of Prof. Hajdu we received a fragment (Fig. 20A) of a fl eshy specimen from the oceanic island groups east of North East Brazil, originally bright yellow, now dark brown in alcohol and staining the alcohol considerably. Th e fragment showed several diff erences with the Curaçao and Puerto Rico specimens (see also Table 4): calthrops (Figs 20B-C) were often more like short shafted triaenes with one cladus considerably longer than the others; they were generally much more irregular and some appeared to have mesotriaene modifi cations. Length of cladi 33-82-123 × 4-7.5-11 μm. Sanidasters (Fig. 20D) were only 10-14-19 × 0.5-1.1-2 μm, and compressed sanidasters (pseudasters) (Fig. 20E) 5-7.2-9 μm. Th ese diff erences may eventually lead to a distinction of a separate Brazilian species but for the time being we emphasize the similarities.
'Pachataxa' lutea does not belong to Calthropella (Pachataxa) for two reasons (see also below): (1) it possesses sanidasters, (2) there are no ataxasters, and (3) the spherasters do not resemble true euasters, they are compressed derivations of the irregularly spined sanidasterlike acanthomicrorhabds. In any case, they seem to be morphologically very close to the Red Sea species Dercitus (Halinastra) exostoticus Schmidt (1868 as Pachastrella) (see above). Description. Small soft patches of a brown-red colour insinuating in three coral fragments (Fig. 21A), occupying spaces less than 5 mm in diameter. No remarkable macroscopical features.
Microrhabds (Figs 21F) short and thick, almost smooth, with only scattered very short spines, 5-7.7-9 × 3-4 μm. Th e microrhabds very likely derive from the sanidasters as there are some stages (Fig. 20E)  Distribution. So far known only from the island of Aruba, Southern Caribbean. Remarks. Th e microrhabds are unique among species of the genus Dercitus s.l. Th e genus Pachastrella also possesses short and thick acanthose microrhabds, but in that genus the microrhabds are much more regularly spined, not of the irregular type found in the present species. Furthermore, Pachastrella species possess oxeas as megascleres next to calthrops or short-shafted (dicho-)triaenes and various types of streptasters such as amphiasters and spirasters. Calthropella (Pachataxa) species (family Calthropellidae) have ataxasters which may look similar in shape, but in addition these possess euasters (see below). Description. Th in leathery-rubbery encrustation (Fig. 22A) cementing coral rubble and fi lling crevices in coral material. Colour (alcohol) pale orange brown. Surface smooth, not encrusted, no oscules apparent; interiorly with a diff erent more mushy texture. Size 2.5×2×0.5 cm.

Dercitus
Skeleton: an ectosomal crust of microscleres overlying a mixture of microscleres and dichocalthrops embedded in largely organic choanosome with relatively low spicule density.
Etymology. Named after the Dutch naval vessel HMS 'Siboga' which was instrumental to the 1899-1900 collection of a rich deep sea fauna from South East Asian waters.
Habitat. Dredged from a sandy bottom with small limestone rubble and shells. Distribution. Known only from the type locality in Eastern Indonesia. Remarks. Th e present specimen was identifi ed by Maurice Burton (who provisionally identifi ed a large part of the Siboga sponge collection) as Stoeba extensa, but habit and colour diff er from that of Dendy's species, whereas the spicules are significantly smaller. S. extensa has a dense mass of dichocalthrops but also calthrops which are absent from the present new species. Th e density of megascleres is also much different. Th e dichocalthrops appear similar to those of D. (S.) occultus (see above), and also the colour matches. However, that species has an endolithic growth form and the sanidasters are apparently not divisable in two shapes. Th e membership of Halinastra of this species rests on the diff erentiation of the sanidasters into two shapes, but proper pseudasters are not found. If the spicular variation of D. (S.) occultus and D. (H.) sibogae sp. n. will be demonstrated to overlap in a future study it might turn out the two are conspecifi c, but for the time being the emphasis is put on the diff erences. Lebwohl, 1914 Dercitus loricatus Lebwohl, 1914: 84, pl. IX fi gs 21-60. Pachastrella monilifera; Maldonado, 2002: 155 (not Schmidt, 1868.

Dercitus loricatus
Material examined. None. Description (from Lebwohl, 1914). Encrusting a lithistid (Discodermia) and partly overgrown by Pachastrella tenuilaminaris. Colour brownish yellow. Consistency compact in dry condition. Surface covered by dirt, serpulids, foraminifera and bryozoans, amidst two openings with diaphragms supposed to be oscules. Surface skeleton a thick crust of microrhabds and metasters. Th e interior is a mass of calthrops interspersed by metasters.

Remarks.
Th is is a dubious description as it was partly overgrown by a Pachastrella species, which might have contaminated the spicule complement (e.g. the microrhabds which are characteristic for Pachastrella). Since no sanidasters nor euasters were described, we exclude this material from our review as a pachastrellid of uncertain affi nity. Maldonado (2002) assigned this to Pachastrella monilifera, but there is insuffi cient evidence for this.

Remarks.
Th is is a dubious Dercitus (Stoeba) species and it needs redescription. Probably it should be assigned to an ancorinid genus.
Type species: Calthropella simplex Sollas, 1888. Comments: Calthropella species are diverse in megasclere and microsclere composition refl ected in several generic names currently in use or considered synonyms. We propose to reassign Calthropella species as here employed, to three subgenera, including Corticellopsis, formerly a synonym, but now revived to encompass those Calthropella species having only 'normal' euasters, Pachataxa, formerly a valid genus, to encompass species with deformed microspined polyangular microscleres ('ataxasters') along with normal oxyasters, and Calthropella s.s. for species having globular asters with characteristically tuberculated rays next to normal asters. Several new species of all three subgenera have been found in various parts of the world and will be described below.
Th e genus Pachastrissa is subsumed into the subgenus Calthropella (Calthropella) because the type species, Pachastrella geodioides (cf. below) diff ers only from the type of Calthropella, C. simplex, in possessing an unstable complement of single oxeas, without structural position in the skeleton, often broken, and occasionally absent. Below, we maintain separate species Calthropella (C.) simplex and C. (C.) geodioides, because we did not examine the type of the former, but Sollas' description of it makes it clear that it diff ers from C. (C.) geodioides in two doubtful features, lack of dichocalthrops (rarely present in C. (C.) geodioides) and oxeas (unstable presence in C. (C.) geodioides, see below). Especially the shape of the globular asters in both is so similar that conspecifi city is likely. Likewise Pachastrissa pathologica (Schmidt, 1868), and P. inopinata (Pulitzer-Finali, 1983) are returned to Calthropella (C.), close to C. (C.) geodioides.
Th e junior synonymy proposed for Pachastrissa as Calthropella does not extend to Pachastrissa hartmeyeri Uliczka (1929) and Pachastrissa nux (de Laubenfels, 1954) as Jasplakina, both of which fi t better in the ancorinid genus Penares on account of the structural oxeas and the presence of ectosmal microxeas in these species.
Th e newly defi ned genus Calthropella is distinct from Dercitus s.l. in the possession of true euasters, the lack of sanidasters, the lack of large darkly pigmented cells with inclusions, and in the harder less organic texture of most species (except subgenus Corticellopsis). Calthrops are the dominant megasclere type, varying widely in size and shape (cladi 40-800 μm), with dichocalthrops so far known only from C. (C.) geodioides and its Indonesian 'variety'. Asters are basically oxyasters, but they are often variable in shape with a strikingly high frequency of spheroxyasters with irregular shapes culminating in peculiar globular silica balls ornamented with tuberculated protrusions in some species. Th e subgenus Pachataxa has asymmetri-cal, deformed microspined siliceous microscleres which are derivatives of asters or microrhabds (which of these is unclear, so far). Diff erentiation in two aster types is usual. Like with several Dercitus species, oxea megascleres are reported in several descriptions, with widely diff erent sizes and usually in broken condition. Th e oxeas may be foreign, but the persistent reports of their presence in some species indicate otherwise. In any case they do not have a structural position in the skeleton and are here considered 'auxiliary'.
Spheroxyasters, considered a young stage of the tuberculated asters by Sollas, but identifi ed as a proper separate aster-type by Topsent; no size given by either authors but according to Topsent the same size as the truncated asters, so 12-24 μm in diameter. Globular spherasters with tuberculated rays, variable, often irregular, in a wide size range, possibly divisible in two size categories, 24 μm and 12 μm.
Remarks. Th is species diff ers from the closely similar C. (C.) geodioides in the lack of dichocalthrops (but these are small and rare in C. (C.) geodioides) and oxeas (but these are an unstable complement not certainly proper in C. (C.) geodioides). Apart from Sollas (1888) and Topsent (1904) (who expressed doubts whether C. simplex and C, geodioides might be synonyms), no records of this species have been published. It is also closely similar to a new species from Indonesia, C. (C.) xavierae sp. n. (see below). Topsent (1892) distinguished a variety 'durissima' for specimens with globular spherasters becoming entirely irregularly rounded in some specimens; these specimens are here assigned to a separate species of Calthropella (Calthropella) (see below).
Habitat. On rocks in deep water, 599-1222 m. Distribution. Off Cabo São Vicente, Portugal; Azores.  lists the species for Madeira, but the basis for that is not clear.
Remarks. Topsent (1904) and Pulitzer-Finali (1983) both point out the great similarity of Calthropella simplex and C. geodioides. Points of diff erence apparently are the absence of dichocalthrops and oxeas in C. simplex. In our present series of samples we could not fi nd any oxeas in samples Ma012 and Ma005 (see above) and in the other samples they were invariably single, not forming bundles or radiating towards the surface and appeared to be broken generally in all sections and spicule mounts. Possibly this was artefactual. Dichocalthrops were rare to extremely rare in all studied samples. Dichocalthrops that were encountered were all in the small size class of triaene megascleres (less than 200 μm cladome size) and often were not 'complete', i.e. only one or two of the cladi were forked.
We believe that the two species simplex and geodioides could very well be the same species and could eventually be merged, but we will await further studies including molecular sequencing. Topsent's (1897) variety of Calthropella geodioides from Ambon (Indonesia) is not close to the present species. It shares the presence of oxeas and dichocalthrops with C. geodioides but calthrops are much smaller (200 μm) and thick-centred euasters are only 10-12 μm. It is redescribed in Desqueyroux-Faúndez (1981) but remains ill-known (see below).
Skeleton: largely confused interiorly, with a densely crowded outer layer of microscleres.
Habitat. Deep water, 450-1132 m. Distribution. W of Flores and seamounts S of the Azores. Remarks. Topsent's variety durissima (type specimen MNHN DT 833 from deep water off the Azores) is here raised to species rank because of the peculiar modifi cation of the tuberculated asters into small irregular silica-balls, which are also larger than those of C. (C.) simplex and C. (C.) geodioides. Topsent's type material was more extensive (largest specimen had half the size of a fi st) and the surface more elaborate ('rugueux', oscules sometimes on a papilla). In 1904, Topsent retracted his variety durissima as he had become convinced that the lumpy asters were derived from the tuberculated asters and probably subject to considerable variation. Having examined the variation of these asters, we disagree with the retraction of durissima and Topsent's assignment of it to the variability of C. (C.) geodioides. Th e predominance of three-claded calthrops and the absence of dichocalthrops combined with the presence of the lumpy entirely spined 'spherochiasters' is suffi cient evidence for specifi c distinction. Schmidt, 1868: 19, pl. III fi gs 3-4. Calthropella pathologica;Topsent, 1938: 24;Vacelet, 1969: 166, fi g. 3;Pouliquen, 1972: 746, pl. 7 fi g. 3;Maldonado, 1992:  Redescription (partly from Topsent, 1938). Two specimens are present in the Schmidt collection of the Muséum National d'Histoire Naturelle, Paris, the largest (MNHN DT 753) of which is 4.5 cm in widest size and 2-2.5 cm high, here chosen as the lectotype. Th e smaller specimen (MNHN DT 754), here designated paralectotype, size 1×3 cm (now apparently reduced to 1×1.5 cm) was examined by us and we made SEM photos of the spicules. Both specimens have a rough surface (Fig. 26A). According to Vaclelet (1969) living specimens are white, but the type specimens we examined were yellowish in alcohol.
Th e megasclere complement also comprised thin oxeas of 15 μm thickness (Fig. 26B). Cross sections of the type specimens examined by P. Cárdenas show scattered bundles of the thin oxeas running vertically to the surface, but their length still is diffi cult to determine. We are indebted to P. Cárdenas for this information. Th e largest unbroken piece we found in preparations of DT 754 was 2000 × 12 μm, which is in accordance with fi ndings of Voultsiadou and Vafi dis (2004).
Tuberculated asters with ornamented rays (Fig. 26D); compared to C. (C.) geodioides the rays were relatively long; sizes highly variable, but overlapping without clear separation in smaller and larger asters, diameter 9-18.4-24 μm Habitat. Deep water, and in caves in more shallow water, depth occurrence at least 13-250 m. Vacelet (1969) reported that his specimen was insinuating in calcareous algae.
Remarks. Topsent (1938) remarked that the specimens are similar in most aspects to Calthropella geodioides but diff er in the tendency of the calthrops to develop extra cladi. In addition the tuberculated asters of that species have shorter rays (almost entirely consisting of tubercles). Oxeas are stated in Voultsiadou & Vafi dis (2004) to be all broken, whereas Vacelet (1969) notes 'wide axial canals', adding to the conclusion that they are not of structural signifi cance. Nevertheless, the bundles reported by P. Cárdenas (in litteris) appear to indicate these oxeas are proper to the sponge. Possibly, they are a remnant of ancestral radiating oxeas.
Oxeas, with extremely elongated, fl exuous points, all broken in the slide, but at least 2000 × 15 μm.
Habitat. Trawled from 60 m depth. Distribution. Known only from the type locality off the coast of Genoa, 44°N 9°E.

Remarks.
Th is species is similar to C. pathologica in most aspects. Oxeas are stated to be all broken, so chances are they are not proper or only auxiliary. Nevertheless their great length and peculiar endings are the main reason to distinguish this species as separate.
Pulitzer (l.c.) suggests that C. geodioides, C. simplex, C. pathologica and his new species C. inopinata could be all members of the same species. For us, this suggestion has merit, at least as far as a possible synonymy of C. pathologica -C. inopinata and C. simplex -C. geodioides is concerned. A further North Atlantic species, C. (C.) durissima differs clearly in having peculiar silicifi ed microscleres derived from the tuberculated asters.  Van Soest, Snellius II Exped. Stat. 095, 19 September 1984. Description. Lobate mass (Fig. 27A), size 6×5×5cm, microhispid surface, rough to the touch. Colour pinkish brown. Two oscule type openings are located on the rounded upper surface, 3-4 mm in diameter. Consistency hard.
Skeleton: a dense crust of asterose microscleres at the surface covers a dense mass of calthrops. Asters are also crowded in the choanosome. Th e skeleton is densely confused and there are few canals or cavities; some broken monaxone spicules were present.
Etymology. Th e name of the new species acknowledges all the help we received from Joana B.R.T. Xavier during her stay in Amsterdam, and the generous donation of several sponge fragments important for this study.
Habitat. Dredged from hard bottom at 138 m. Distribution. Known only from the type locality, E of the island of Komodo, Nusa Tengara (southern island chain bordering the Banda Sea).
Remarks. Th is is a clear Calthropella (Calthropella) resembling the type species C. (C.) simplex in all but the fi ner details of spicule sizes and shapes. Th e habitus is more elaborately globular and has pinkish brown colour (in alcohol) in contrast to more fl attened pale yellow-grey C. (C.) simplex; there appear to be two distinct size categories of oxyspherasters in the new species.
No matching descriptions are found in the Indo-West Pacifi c region, but Topsent (1897) reported the occurrence of Calthropella geodioides var. from Ambon Bay, Indonesia (see also below based on a redescription in Desqueyroux-Faúndez, 1981) which shows a few similarities. It diff ers from our new species at least by the possession of dichocalthrops, apparently proper oxeas and white colour.
Description (from Topsent 1897 andDesqueyroux-Faúndez 1981). Th is is described as a small white sponge, similar to the North Atlantic C. geodioides, but differing by possessing calthrops with a fi fth cladus, and by rare microscleres. Calthrops are much smaller than in the North Atlantic material, with cladi length 180-210 μm, whereas the dichocalthrops are much larger than those of C. geodioides proper, being of similar size as the calthrops. Th ere are thick-centred tuberculated asters with truncated rays, 10-12 μm. Further megascleres are oxeas of 750 × 20 μm (not pictured by Desqueyroux-Faúndez).

Remarks.
It is likely that the Ambon material belongs to an undescribed species, as the length of the cladi of the calthrops and the diameter of the tuberculated asters are much smaller than in the type of Calthropella geodioides, and 'normal' oxyasters are apparently not present. In addition, the localities and habitats are widely separated (deep sea off Portugal vs. Ambon Bay).
Type species: Pachastrella lithistina Schmidt, 1880 (by original designation).  (Fig. 28), size 9×6 cm, thickness 2 cm. Surface diff erent on both sides, oscular side ( Fig. 28B) with irregularly scattered oscules of 1-2 mm diameter each elevated on small hillocks; the poral side ( Fig. 28C) with numerous small openings irregularly arranged in groups separated by areas without such openings. Edges of the plate smooth, without pores or oscules. Colour of the dried specimen pale yellow with a pinkish brown tinge. Consistency hard, slightly crumbly (dry condition). In cross section (Fig. 28D), there is a dense interior yellow-white mass, riddled with thin canals; at the periphery there are subdermal spaces underneath a 1 mm thick crust.

Key to the species of Pachataxa
Skeleton: a confused mass of calthrops with at the periphery a dense mass of ataxasters, which are also strewn in the interior.
Habitat. Deep water, 160 m. Distribution. Only known from the type locality, off Grenada. Remarks. Th e type material was assigned by  to a lectotype in the Strasbourg Museum and a paralectotype in the Museum of Comparative Zoology, Cambridge, USA, but it is virtually certain that both are fragments of a single specimen. Th us, both the alleged lectotype and paralectotype are now assigned to the same holotype as schizoholotypes. C. (P.) lithistina diff ers from both other Pachataxa species in the shape (large plate with diff erent oscular and poral surfaces) and in the large variations in ataxaster shapes and sizes.  Description. Massive with rounded fl attened lobate outgrowths (Fig. 30A). Size of holotype 3.5×3.5×1.2 cm, paratypes slightly smaller to about 1 cm 3 . Brownish beige in alcohol. Small slightly raised oscules of less than 1 mm diameter are scattered in vague groups on the upper surface. Consistency fi rm to hard, crumbly.
Habitat. Deep water, 355 m. Distribution. Known only from off New Caledonia. Remarks. Th e New Caledonian species diff ers from the Caribbean C. (P.) lithistina in overall spicule size and particularly in the small size of the oxyasters; that species is much larger and has a plate-shape. C. (P.) pyrifera sp. n. (see below) diff ers likewise in having larger oxyasters up to 24 μm diameter and asymmetrical pearshaped ataxasters.  Description. Type material available to us has a fl attened (Fig. 31A left), rounded shape, size 1×1×0.2 cm. It is a fragment of a larger holotype specimen (3.5×3×2.5 cm) that is massively rounded and may have a large central oscule (Fig. 31A right). Colour very pale green or dirty white. Skeleton: an irregular mass of large triaenes covered at the periphery by a thick layer of ataxasters / microrhabds, which are also strewn in the interior. Loose oxeas of widely diff erent sizes present in moderate quantities, considered foreign.
Ataxasters (Figs 31D-E), typically pyriform (pointed one-sidely), occasionally ovate or rounded, microspined all over, but spines tend to be grouped; no branching shapes were found, making the term 'ataxaster' inappropriate for these spicules, but their homology to the ataxasters of P. lithistina is nevertheless obvious; size 7-20.5-31 μm, measured along the longest axis.
Etymology. From the Greek pyros = pear and fero= carry or bear, referring to the possession of the pear-shaped microscleres.
Habitat. Deep water, 506 m. Distribution. NE of Santa Cruz Island, Galapagos, East Pacifi c.

Remarks.
Th e species is assigned to Calthropella (Pachataxa) on the basis of its spiculation. Size and shape of the calthrops is similar to C. (P.) lithistina, but these diff er clearly in the diameter of the spheroxyasters (those of C. (P.) pyrifera sp. n. are twice as large and these are also twice as large as those of P. enigmatica, see above). Especially the shape of the ataxasters renders this species distinct from the other two Calthropella (Pachataxa).

Subgenus Corticellopsis Bergquist, 1968
Corticella Sollas, 1888: 281 (preoccupied by Corticella Ehrenberg, 1872. Defi nition: Calthropella with 'normal' euasters (oxyasters and/or strongylasters). Type species: Corticium stelligerum Schmidt, 1868 (by subsequent designation herein).  (1895), based on a fragment presented to the British Museum (Natural History), BMNH 1910.1.1.863. Encrusting on corals and bridging crevices between coral branches, thickness about 5 mm. Colour white (in dry condition), yellowish inside. Desqueyroux-Faúndez & Stone (1992) picture the Schmidt specimen of which a fragment was studied from Sebenico as a massively encrusting sponge of 4×4 cm, encrusted and riddled with shells. According to Sollas (1888) it has a thickness of 5 mm. Cruz (2002) reports pale yellow or greyish live colour. Th e World Porifera Database (van Soest et al., 2008) has an in situ picture made by Dr B. Picton from a locality just off the coast of Marseille showing a pale yellow colour. No apparent oscules.

Key to the species of Corticellopsis
Skeleton: densely spiculous at the surface, organic with few spicules in the interior. No defi nite skeletal structure.
Microscleres euasters in two categories (although the LMJG fragment available to us only contained a single one: strongylasters 16-20.2-22 μm in diameter (Fig. 32C) with microspined rays in the studied fragment). Schmidt (1868) gives 20 μm as size of the strongylasters, whereas the oxyasters were measured by him as 50 μm. Topsent gives only ray lengths, 30 μm for the oxyasters (which could conform to 50 μm in overall diameter), and 4-7 μm for the 'chiasters' (which seems a bit too short for an overall 20 μm diameter mentioned by Schmidt). Presumably the various type specimens have considerable size variation in the asters.
Habitat. Fairly shallow water, from intertidal caves to 20 m. Distribution. Adriatic; Marseille; Canary Islands; NW Aegean Sea, 39°N 25°E (Algeria is given by previous authors, but this is probably based on a misunderstanding of Schmidt's text).

Remarks.
Th e species is the type of the preoccupied genus Corticella Sollas (1888) replaced by Corticellopsis Bergquist, 1968 (see  for details). It is apparently quite rare as there are only a few records from the Mediterranean (Northern Adriatic, Greece, and unconfi rmed from Marseille). Schmidt (l.c.) mentions Algeria in his text, but very probably refers to another Corticium species (either C. candelabrum or C. plicatum = Dercitus plicatus). Th e species has been reported recently from the Canary Islands (Cruz &Bacallado 1981 andCruz 2002) as pale yellow or grey masses with a skeleton of calthrops with cladi of 60-320 μm, irregular chiasters (strongylasters) of 8-23 μm diameter, drawn with smooth rays, fi ve-rayed oxyasters also drawn with smooth rays of 16-40 μm. Voultsiadou and Vafi dis (2004) give calthrops cladi of 120-240 × 20 μm, 'chiasters' with 5-8 μm ray lengths, and six-rayed oxyasters of 20-28 μm ray length. Th ey also mention the presence of considerable amounts of oxeas, but these were assumed to be foreign. Th e various measurements Strongylasters, rather irregular in shape, with stunted rays, but not with a thickly silicifi ed centre, 11 μm.
Habitat. Sublittoral fringe. Distribution. Rangitoto Island, northern New Zealand. Remarks. Th e drawings and photos of the asters are limited to the rather irregular strongylasters, whereas no illustrations are available of the oxyasters. It is assumed these were regular and unremarkable.  Van Soest, nr. 612/20, 12 December 1992. Description. Cartilaginous crust on a piece of dead coral, surface smooth, size 2×1×0.5 cm. Greyish black alive, pale brown in alcohol.
Habitat. Encrusting dead corals in shallow water. Distribution. Known only from a single locality south of Victoria, Mahé. Remarks. We refrain from naming this material due to the rarity of its microscleres. By its small sized calthrops this is a distinct species, as no matching descriptions appear in literature dealing with Calthropella s.l. Th e oxyspherasters were quite rare, and they could not be detected in any of the thick sections, whereas the small oxyasters were only marginally less rare, so there is a possibility that they are not proper to this sponge. Th e cartilaginous nature and low spicular density of the specimen reminded of Dercitus rather than of Calthropella, but we did not fi nd sanidasters.
Habitat. Deep water, 120 m. Distribution. North Kenya Banks, E Africa.

Remarks.
Th is is in all probability not a Calthropella as it has a dense skeleton with confusedly arranged oxeas. Th is species is best assigned to Stelletta s.l..

Discussion
From the above presented data a close relationship of Dercitus s.l. and Calthropella s.l. has been made plausible, notwithstanding a clear separation in microsclere complement, shape and texture. Th e assignment of these two genera to diff erent astrophorid families appears unnecessary, because Dercitus, although lacking euasters, does not possess clear pachastrellid streptasters other than acanthomicrorhabd-like sanidasters. Several Calthropella species, although possessing euasters, have heavily silicifi ed microscleres similar to some Dercitus species. Th ese microspined silica bodies can be morphologically derived from euasters (e.g. those of Calthropella (Calthropella) or from microrhabds (those of Calthropella (Pachataxa). Th e ovoid microspined microscleres of Dercitus (Halinastra) arubensis sp. n. appear to bridge the gap between Dercitus (Halinastra) and Calthropella (Pachataxa).
Assuming that the similarities will be found to be homologous and that the two considered groups will be found to be closely related, family assignment to one of the currently recognized families is not unequivocal. Calthropella at present is a member of the small family Calthropellidae, which -with the above proposals of generic synonymy -is reduced to virtually a single genus (Pachastrissa, Pachataxa and Corticellopsis are junior synonyms, Chelotropella is an ancorinid because of a radiating skeleton of long-shafted triaenes). Dercitus s.l. could be assigned as a second genus to Calthropellidae, but defi ning such a modifi ed family is problematic in view of the fact that Triptolemma shares similar skeletal structure but has undoubted pachastrellid characters in the form of genuine streptasters (amphiasters, spirasters). Similarity of Triptolemma to Dercitus s.l. appears to be considerable because next to (dicho-)mesotriaene megascleres this genus also has calthrops and acanthomicrorhabds. Some species of the type genus of Pachastrellidae, Pachastrella, although having a skeleton with structural oxeas, appear close again to Triptolemma in sharing calthrops, dichocalthrops, mesotriaenes, streptasters and acanthomicrorhabds. Th is could mean that the above observed similarities are homoplastic rather than homologous, demonstrating independently acquired losses possibly induced by adaptations to the crevice-agglutinating-endolithic habitat. Th us, our results seem ambiguous for eventual changes in the classifi cation. We will await further independent evidence for a revised classifi cation of astrophorid genera.

Further diversity
Including the three unnamed species described above, there appear to be at least seven records of Dercitus and Calthropella that likely concern distinct species in need of further description and diff erentiation before they can be properly named: A further two records (Dercitus (Dercitus) bucklandi sensu Teerling, 1975 from the Western Gulf of Mexico, and Dercitus sp. sensu Rützler et al. 2000 from Belize) await description and assignment to subgenus.

Distribution patterns
Previous analysis of generic distributions of Demospongiae ( Van Soest 1994) indicated that a circum-global temperate and warm water occurrence is the most common pattern. Dercitus and Calthropella are probably following this pattern closely. Dercitus s.l. (Fig. 34)  an oddly limited distribution in the Celtic Seas and South European Atlantic shelf marine ecoregions. Subgenus Stoeba is the most widespread, occurring over almost all provinces mentioned above. Subgenus Halinastra, although less common, likewise is represented in provinces encircling the globe. It is lacking so far in the Eastern Pacifi c and Eastern Atlantic provinces.
Calthropella s.l. (Fig. 35) is less speciose, so far, but is also circumglobally distributed, occurring in most provinces that also contain Dercitus species. However, in the northern hemisphere the genus is lacking from provinces and regions north of the tropical provinces: it is not found in the Celtic Seas, the Gulf of Mexico and the Southern Californian Bight region. However, in the regions south of the tropics one species occurs in the Northeastern New Zealand province. Most species, excepting those of subgenus Corticellopsis, are found in deep-water and are rarely occurring above 100 m. Subgenera Calthropella (Calthropella) and Calthropella (Corticellopsis) so far are not known from the Caribbean. Subgenus Pachataxa is lacking from the Eastern Atlantic and Mediterranean. It is likely that these distribution patterns will turn out to be less disjunct when more information on the distribution of members of this genus is collected.