Geodia barretti, BOWERBANK, 1858

Cárdenas, Paco, Rapp, Hans Tore, Klitgaard, Anne Birgitte, Best, Megan, Thollesson, Mikael & Tendal, Ole Secher, 2013, Taxonomy, biogeography and DNA barcodes of Geodia species (Porifera, Demospongiae, Tetractinellida) in the Atlantic boreo-arctic region, Zoological Journal of the Linnean Society 169 (2), pp. 251-311 : 263-271

GEODIA BARRETTI BOWERBANK, 1858

DEPRESSIOGEODIA P BARRETTI (PhyloCode SPECIES NAME)

( FIGS 7–10 , TABLE 2)

Geodia barretti, Bowerbank, 1858: p. 290 ; 1862: p. 768, 794; 1864: p. 168, 171; 1872a: p. 198; Sollas, 1880a: p. 247; 1888: p. 250; Vosmaer, 1882: p. 23; Norman, 1893: p. 349; Lundbeck, 1909: p. 455; Burton, 1930: p. 490; 1959: p. 9; Filatova, 1938: p. 28; Alander, 1942: p. 73; Blacker, 1957: p. 27; Burdon-Jones & Tambs-Lyche, 1960: p. 6; Koltun, 1964: p. 147; 1966: p. 54; Dyer et al., 1984: p. 669; Boury-Esnault et al., 1994: p. 38; Klitgaard, 1995: p. 2; Klitgaard & Tendal, 2004: p. 57; Nichols, 2005: appendix A; Cárdenas et al., 2010: p. 89; Cárdenas et al., 2011: table S1; Murillo et al., 2012: p. 842; Guihen et al., 2012: p. 1; Schöttner et al., 2013: p. 2; Cárdenas & Rapp, 2013. Geodia baretti (misspelling), Schmidt, 1866: p. 11 –12; Fristedt, 1885: p. 43; 1887: p. 463; Breitfuss, 1911: p. 213; Brøndsted, 1914: p. 526; Lidgren et al., 1986: p. 3283; Warén & Klitgaard, 1991: p. 52; Reitner & Hoffmann, 2003: table 1; Rosenberg et al., 2005: p. 45; Purser et al., 2013: p. 37. Geodia barreti (misspelling), Rezvoj, 1928: table 1.

Cydonium barretti, Gray, 1867: p. 548 .

Sidonops barretti, von Lendenfeld, 1903: p. 101 ; Hentschel, 1929: p. 919. Sidonops baretti (misspelling), Breitfuss, 1930: p. 277.

Sidonops sp. , Hentschel, 1929: p. 867 (synonymy by this study).

Geodia simplicissima, Burton, 1931: p. 2 ; Oug & Rapp, 2010: p. 189; Cárdenas & Rapp, 2013 (synonymy by Cárdenas & Rapp (2013)).

Not:

Geodia barretti, Boury-Esnault et al., 1994 (in part): p. 38 (CP63-E5 = Geodia megastrella ); Voultsiadou & Vafidis, 2004: p. 593 (= Geodia conchilega ); van Soest et al., 2007 (in part?): table 2 (= Geodia atlantica ).

Geodia barretti var. nodastrella Carter, 1876: p. 397 ; Sollas, 1888: p. 247; Topsent, 1892: p. 48 (= Geodia nodastrella ).

Geodia barretti var. senegalensis Topsent, 1891: p. 15 (= Geodia barretti var. senegalensis ).

Geodia barretti var. divaricans Topsent, 1928: p. 110 (= Geodia divaricans ).

Misidentification:

Geodia cydonium, Burton, 1959: p. 9 .

Type locality and deposition of holotype

Geodia barretti , collected by Robert McAndrew (1802– 1873). South side of Vikna Island (formerly called Vigten or Vikten Island), North-Trøndelag, Norway, 183 m, BNHM 1877.5.21.1399 (dry specimen), BNHM 1877.5.21.1400 (one slide of surface and one spicule preparation), BNHM 1877.5.21.1401 (slide of section). Geodia simplicissima, Foldenfjord , northern Norway, 10–75 m, TSZY 10 (wet specimen). Spicule preparations made during this study are now stored at TSZY.

External morphology and cortex: Irregularly massive, up to at least 80 cm in diameter, and up to a weight of c. 38 kg (wet); young specimens are usually spherical to subspherical. Mostly with an obvious attachment area, sometimes formed as several stilt-like projections each attached to a piece of gravel. The surface colour (alive) is usually white ( Fig. 7 A-C, E), but with sometimes various shades of light yellow ( Fig. 7D ) or light brown ( Figs 7F , 8 ). The choanosome alive is light brown ( Fig. 7C ) and becomes whitish in ethanol. The surface is usually clean and smooth but shallow specimens (30–50 m) can be slightly dirty and hispid. Some NWA specimens were very hispid over their entire surface. One to many (more than 30) preoscules (i.e. a depression protecting the true oscules), more or less deep, more or less narrow, with a circular to irregular opening (up to several cm wide) ( Fig. 7A–F ). Preoscules are generally on top. The preoscule contains uniporal oscules ( Fig. 7G ). Each oscule (1 mm in diameter) has a sphincter. Cribriporal pores are scattered over the entire body surface ( Fig. 7H ); single pores are 50–80 Mm, and poral sieves are c. 0.5 mm. The cortex is 0.4–0.6 mm (ectocortex: c. 250 Mm, endocortex: c. 750 Mm) ( Fig. 9A ). In the preoscule, the cortex is without sterrasters and triaenes ( Fig. 9B ), and ridges of microxeas and strongylasters surround the uniporal oscules ( Figs 7G , 9B ).

Description of type material: Medium-sized oval specimen (length: 12 cm, width: 8 cm) from Bowerbank (1872a: plate XI) which has been cut into five pieces (the main specimen and four smaller pieces); three Bowerbank slides including spicule preparation, section, and cortex surface.

Spicules ( Fig. 9 , Table 2): Megascleres: (a) oxeas I, straight or bent, length: 1075–4450 Mm; width: 15–75 Mm. (b) Oxeas II, straight or bent, rarely modified to styles, sometimes slightly centrotylote, length: 190–900 Mm; width: 4–16 Mm. (c) Dichotriaenes, rare orthotriaenes, rhabdome length: 620–4600 Mm; width: 20–150 Mm; orthotriaene clad length: 240–500 Mm; protoclad length: 100–400 Mm; deuteroclad length: 45–450 Mm. (d) Anatriaenes, rhabdome length: more than 7.4 mm; width: 9–40 Mm; clad length: 50– 250 Mm. (e) Meso/protriaenes (rare), rhabdome length: up to 2640 Mm; width: 7.5–15 Mm; clad length: 25–115 Mm; central clad length: 25–98 Mm. Microscle- res: (f) sterrasters, spherical to elongated, length: 65–130 Mm, width: 51–105 Mm, thickness: 60–80 Mm; hilum diameter: 12–23 Mm. Rosettes are made of 3–7 rays, covered with warts; rosette diameter: 4–7 Mm. (g) Strongylasters, rough actines, 3–11 Mm in diameter. (h) Oxyasters I (only in very deep specimens> 1000 m), rough actines, diameter: 30–80 Mm. (i) Oxyasters II, rough actines, diameter: 6–32.5 Mm.

The spiculogenesis of shallow specimens (30–50 m depth) being somewhat disrupted, their spicule measurements have not been included here but they are shown in Table 2 and discussed in Cárdenas & Rapp (2013).

DNA barcodes: We found two haplotypes for the COI Folmer marker. GenBank accession nos. HM 592679 , HM 592695 , and EU442195 : haplotype 1 from Spitsbergen (5), southern, western, and northern Norway (12) , Sweden (1) , off western Ireland

Means are in bold; other values are ranges; N = 30 unless stated otherwise in parentheses, or unless measurements come from other studies. A dash indicates that this measurement is not given in the literature. n.f., not found; n.o., not observed in the specimen in our possession (usually because the sample was too small).

(2), Davis Strait (1), Flemish Cap (1), and the Mediterranean Sea (1). No. KC574389 : haplotype 2 (1-bp difference with haplotype 1 in position 382: ‘A’ instead of ‘T’) was found in two specimens from the Flemish Cap ( UPSZMC 78262 , UPSZMC 78268 ). Nos. EU552080 , HM 592809 (28S, C1-D2 domains): we have sequenced 28S (C1-D2) from specimens from Spitsbergen (1), western Norway (2), and off Ireland (1): we did not observe genetic differences in this marker among NEA specimens. No. KC481224 (18S), obtained from ZMBN 77922 (Korsfjord, Norway) and ZMBN 89722 (Spitsbergen): no variation was observed .

Distribution ( Fig. 10 ): Geodia atlantica , G. barretti , and G. hentscheli may have been confused in the past, especially until the description of G. atlantica by Stephens (1915) and G. hentscheli by Hentschel (1929), and above all when juveniles were found (e.g. Burton, 1949). This should be kept in mind when examining the G. barretti distribution map that includes a few records not verified by us. However, we did check specimens from Fristedt (1887) ( SMNH), Lundbeck (1909) ( ZMUC), Boury-Esnault et al. (1994) ( MNHN), Voultsiadou & Vafidis (2004), Nichols (2005), and van Soest et al. (2007) ( ZMAPOR). Specimens from Voultsiadou & Vafidis (2004) and van Soest et al. (2007) were mis-identifications (cf. Discussion). The record in the Asturias ( Spain) given by Ferrer-Hernández (1918) at 150–300 m depth is based on slides, and it is dubious as it seems too shallow for this species at this latitude, but he unfortunately gives no description. Other identifications could be confirmed by accurate descriptions and plates (e.g. Vosmaer, 1882). Geodia barretti has been found at depths from 30 to 2000 m. Most NEA records are from between 200 and 500 m, at temperatures of 4–8 °C; Grand Banks, Flemish Cap, Nova Scotia, and Davis Strait specimens were found at 410–1852 m, at temperatures of 3–5 °C. Shallow specimens from the western Norwegian coast have been collected at temperatures of 3–9 °C, and possibly experience up to 14–15 °C in September–October ( Cárdenas & Rapp, 2013). The only specimen we identified from the Mediterranean Sea was collected at 167 m where the water temperature is around 13 °C and the salinity usually more than 38 p. p.m. Localities where the species occurs at lower temperatures, down to 0.4 °C, were only found in the Denmark Strait. Breitfuss (1930) reports G. barretti in the southern part of the Kara Sea at –1.75 °C but we have not examined this specimen, and because no other records exist of this species in this area, this record needs to be confirmed and is here considered dubious.

Blacker (1957) only gives the coordinates for his 1949 and 1950 trawls; we could not find coordinates for the 1951, 1952, 1954, and 1955 trawls. Likewise, Dyer et al. (1984) do not give coordinates for their 1978–81 trawls. We therefore manually copied on Figure 10 the G. barretti records between northern Norway and Spitsbergen from figure 3c in Dyer et al. (1984), which also integrates the Blacker (1957) localities.

Biology: Gametogenesis has been well studied as well as the annual reproductive cycle ( Spetland et al., 2007). This study on Scandinavian fjord populations shows that G. barretti is (1) gonochoric and oviparous and that (2) reproduction coincides with phytoplankton blooms. Gametogenesis usually takes place from February to May with a gamete release in early summer; sometimes a second gametogenesis/ spawning event takes place later in the summer ( Spetland et al., 2007). In our only specimen from the Mediterranean Sea, collected on 22 August 2010, spermatogenesis was observed. We found no indications of asexual reproduction in this species.

Geodia barretti survives well in tanks in open circulation systems. Specimens at Tjärnö Marine Biological Laboratory (University of Gothenburg) and High-Technology Center (University of Bergen) have been kept in tanks for two years and we have even observed release of sperm cells (Rapp & Cárdenas, unpublished results). Cultivation of explants has also been successful, and has led to studies on growth and regeneration ( Hoffmann et al., 2003), stability of the microbial community ( Hoffmann, Rapp & Reitner, 2006), as well as oxygen dynamics ( Hoffmann et al., 2005), the last showing the importance of anaerobic processes within this species. Actually, both aerobic (nitrification) and anaerobic (denitrification) microbial processes were later detected in G. barretti ( Hoffmann et al., 2009; Radax et al., 2012a, b), thus suggesting the complexity of the nitrogen cycle in G. barretti . The microbial community was further studied and understood using conventional bacterial cultivation and 16S rDNA clone libraries ( Graeber et al., 2004) or using a metatranscriptomic approach ( Radax et al., 2012a). In G. barretti , this community seems to be dominated by three prokaryotic groups: phylum Chloroflexi (SAR202 cluster), the candidate phylum Poribacteria, and Acidobacteria; potential eukaryotic symbionts were poorly represented (<1%) ( Radax et al., 2012a).

The sponge-feeding chiton H. nagelfar and the parasitic foraminiferan H. sarcophaga have been found living on G. barretti in the NEA ( Warén & Klitgaard, 1991; Cedhagen, 1994; Todt et al., 2009). Predators such as the aforementioned chiton may cause surface injuries which are later filled with sediments and spicules, and encapsulated in new sponge tissue, thus forming large inclusions ( Hoffmann et al., 2004). Klitgaard (1995) shows that, overall, this species has less associated macrofauna than any of the other boreo-arctic Geodia species : only ten different species of epibionts were recorded vs. 62 for G. macandrewii . The chemistry (elemental analysis, amino acids, sterols and quaternary ammonium compounds) has been investigated by Hougaard et al. (1991a, b). Brominated cyclodipeptides have been particularly studied in G. barretti . Three structurally similar brominated cyclodipeptides (barettin, 8,9- dihydrobarettin, and bromobenzisoxazolone barettin) were isolated and described ( Lidgren et al., 1986; Sölter et al., 2002; Sjögren et al., 2004; Hedner et al., 2008). Experiments have shown that these three cyclodipeptides inhibit settlement of barnacle larvae [ Amphibalanus improvisus ( Darwin, 1854) ] in a dosedependent manner ( Sjögren et al., 2004; Hedner et al., 2008), thus suggesting that these chemicals may play a role in preventing fouling of the sponge surface. It has further been shown that barettin and 8,9-dihydrobarettin act in synergy against foulers ( Sjögren et al., 2011). These compounds may also be involved in defence against grazers or predators [deterrence experiments with the hermit crab Pagurus bernhardus (L., 1758)] ( Sjögren et al., 2011). The spelling of ‘barettin’ with only one ‘r’ is due to a misspelling of G. barretti with one ‘r’ in the original paper describing this molecule ( Lidgren et al., 1986).

Distinctive characters: External morphology: the generally smooth surface (absence of hispidity and epibionts) and white colour. The irregular form, especially in specimens larger than about 15 cm in diameter. The clearly visible sieves in the sometimes numerous preoscular cavities. Spicules: usually dichotriaenes and strongylasters (but these characters are not sufficient as G. hentscheli can also have both).

Remarks: As explained before ( Cárdenas et al., 2010), we stress that G. barretti ’s oscules are not covered by a sieve. There is a depression called a preoscule, in which we find single uniporal oscules (without any kind of sieve). Every oscule has its own unique sphincter, and this is clearly visible with the naked eye ( Fig. 7G ) or in a thick section ( Fig. 9B ). We find the same arrangement in G. hentscheli (cf. below).

Burton (1949) identified some very small Geodia specimens as G. barretti ; this identification is probably wrong, as the specimens seem to have been buds, and came from ‘an unspecified point in the Arctic and from an unknown depth’. But he pointed out the similarity of some of his specimens to G. parva . Indeed, what he had in front of him must have been buds from G. hentscheli or G. parva .

Sidonops sp. (ZMB Por 7552) described by Hentschel (1929) has cribriporal pores and uniporal oscules in large preoscules (7– 3.5 cm in diameter), and has microxeas and strongylasters. According to these characters and pictures of the specimen (courtesy of C. Lueter, ZMB), we can be sure that this is G. barretti . Of the two G. barretti specimens reported by van Soest et al. (2007) from Rockall Bank, only one was found in the collection (ZMAPOR 19647) and it turned out to be G. atlantica .

We examined a spicule slide of the specimen identified as G. barretti from the Mediterranean Sea (specimen A183), collected at a surprising 4–6 m depth in the Aegean Sea ( Voultsiadou & Vafidis, 2004). In our opinion, this is a misidentification; the dichotriaenes and oxyasters are typical of Geodia conchilega Schmidt, 1862 , a common Mediterranean shallow species for which we had comparative material (e.g. MNHN DNBE- 846, G. conchilega from Banyuls, France, collected and identified by N. Boury- Esnault). But we did identify a G. barretti specimen collected in the ‘Canyon des Moines’ (south Corsica) at 167 m depth (‘CorSeaCan’ campaign with the ROV Achille). The spicule morphologies match those of our other specimens: there are only oxyasters II (10– 35 Mm); microxeas can be slightly bent and are occasionally centrotylote; the cortex thickness is standard (0.5–0.6 Mm). The main and only difference we could find is that the spherical sterrasters are smaller (56– 59.9– 65 Mm) than in the Atlantic G. barretti ( Table 2), except those from shallow waters. It had the same COI Folmer haplotype 1 as all the NEA and most of the NWA specimens. This is the first true record of this species in the Mediterranean Sea. At least six additional sightings between 167 and 199 m depth (without collection) of G. barretti -like specimens were made during the ‘MedSeaCan’ and ‘CorSeaCan’ campaigns: in the ‘Banc de Magaud’, ‘Banc de Nioularge’ (both off the Côte d’Azur), and ‘Canyon de Cargèse’ (western Corsica) (J. Vacelet & M. Fourt, pers. comm.). Overall, these G. barretti -like specimens seem less smooth than northern ones, with regular small bumps on their surface (where megascleres cross the cortex and retain sediments); unlike the NEA specimens, they always had a single deep, wide preoscule (c. 3 cm in diameter). We cannot be completely sure they are all G. barretti because Geodia megastrella Carter, 1876 can have a very similar external morphology, although it has never been observed in the Mediterranean (although we have found it in the Balgim material collected off Morocco, along with G. barretti ).

The spiculogenesis of shallow specimens (30–50 m) is disrupted so that spicule morphologies are somewhat different ( Cárdenas & Rapp, 2013). Thus, it has been shown that G. simplicissima from northern Norway is actually a G. barretti growing in shallow waters. G. simplicissima has therefore been put in synonymy with G. barretti ( Cárdenas & Rapp, 2013) .

We examined the holotype of Geodia barretti divaricans [ MOM 04-1333 (wet specimen) and MNHN DT-1299 (type slide)]. In the choanosome, it has slightly spined oxyasters which can reach a very large size (25–70 Mm in diameter) and just under the cortex strongly spined oxyspherasters (17–22 Mm in diameter), which are different from the oxyasters II of G. barretti . Geodia divaricans is clearly different from G. barretti .

Geodia barretti senegalensis was elevated to the species rank by Burton (1956) without any explanation. We therefore examined the holotype of G. barretti senegalensis (MNHN DT-3241, dry specimen). This shallow Geodia does not have a preoscule so it is certainly not G. barretti and G. senegalensis is a valid species. The external morphology looks more similar to some specimens of Geodia gibberosa Lamarck, 1815 from the Caribbean reefs.

Specimens from the Ibero-Moroccan Gulf ( Fig. 8A–C ) collected during the ‘Balgim’ campaign (CP63-E2, CP98-47) have a few differences from the rest of the specimens ( Boury-Esnault et al., 1994). The preoscules are covered with sediment and the preoscule cortex is loosely attached to the underlying choanosome ( Fig. 8B, C ). Small anatriaenes II have been found in the preoscule, along with a few pro(meso)triaenes. Their presence may be due to the large amount of sediments in the preoscule, a condition seldom observed in boreo-arctic specimens where preoscules are usually clean of sediments and usually more firmly attached to the choanosome in northern specimens ( Fig. 8D ). We decided to consider these anatriaenes as a second size category as such smaller anatriaenes were never observed in other specimens. The microxeas are essentially straight, usually slightly thicker on one half, and never centrotylote, but the straightness seems to be a common feature of all NEA and NWA specimens collected at more than 1000 m depth (e.g. ZMBN 85202 from Ireland, UPSZMC 78259 from the Flemish Cap). When spicules of Balgim specimen CP98-47 were examined by SEM, we could not find any other differences with our SEM observations of Norwegian and deep Ireland specimens. Together, these differences (dirty preoscule and anatriaenes II, slightly asymmetrical microxeas never centrotylote) do not justify a new species for the time being. More specimens and genetic data would be necessary to understand the status of this southern population, probably related to the Mediterranean Sea populations ( Fig. 10 ). Specimen CP63-E5 had unusually large sterrasters (125–155 Mm), orthotriaenes, and a fairly thick cortex (1 mm thick), so we re-identified it as G. megastrella .

No morphological differences were found between specimens from the NWA (Flemish Cap, Nova Scotia, Davis Strait) and specimens from the NEA. We just note that some NWA specimens were very hispid (e.g. UPSZMC 78268–78269), a feature never observed in NEA specimens. NWA Specimens deeper than 1000 m depth do have oxyasters I but they are much smaller (c. 32–42 Mm) than in NEA specimens deeper than 1000 m (up to 58–80 Mm), and more difficult to consider as a separate size category. Also, NWA specimens deeper than 1000 m have subspherical sterrasters, not elongated like in NEA specimens collected at the same depths. No morphological differences were observed between the two haplotypes ( Table 2: UPSZMC 78269 vs. other specimens). Interestingly, haplotype 2 is closer by 1 bp to the sequence of G. hentscheli than haplotype 1: it is therefore closer to the common ancestor of these sister-species, which would suggest that the common ancestor lived in the NWA. It so happens that UPSZMC 78268 (haplotype 2) was initially identified as G. hentscheli due to its important hispidity, spherical shape, and narrow unique preoscule (see a complete description of UPSZMC 78268 on the Sponge Barcoding Project, http://www.spongebarcoding.org). Further work is needed to see if haplotype 2 has a consistently G. hentscheli -like morphology (as in Fig. 7A ).

Blacker (1957) and Dyer et al. (1984) sampled extensively between northern Norway and Spitsbergen between 1949 and 1981 and found extensive sponge grounds. But when O.S.T. participated in the Meteor 1990 cruise in the same area, very few Geodia were collected ( Barthel, Tendal & Witte, 1991). Eight triangular dredges were made in the southern area off Bear Island, and only one small Geodia was collected. In the northern area off Spitzbergen numerous triangular dredges were made and no Geodia were collected; a single large specimen of G. macandrewii was taken by a hyperbenthic sledge. There is a possibility that the reason for this sampling discrepancy is the sampling method – Blacker (1957) and Dyer et al. (1984) used a trawl while a large triangle-dredge was essentially used on the Meteor – but this does not seem very likely. Alternatively, the large masses of sponges earlier reported may have disappeared since 1981 due to an inflow of very cold water from the north, intensive trawling activity in the area, or disease, although it is difficult to believe that any of these would hit such a large area. It is also possible that the Meteor cruise was rather unlucky in finding sponge grounds. The ‘Ecosystem Barents Sea’ cruise in 2007 collected tonnes of Geodia in station 2562, but this was much closer to the Norwegian coast (c. 80 km).

Mass mortality of G. barretti was actually observed in the Kosterfjord area (southern Norwegian and western Swedish waters) and started in the winter of 2006/07; this may be due to unusually high temperature and a deepening of the thermocline in 2006 and 2008 in this region ( Guihen et al., 2012). Maximum temperatures in the autumn 2006/08 at Tisler reef were 12.5 °C instead of 9 °C in other years. However, shallow specimens at 30 m depth on the western Norwegian coast seem to experience up to 14–15 °C in September–October ( Cárdenas & Rapp, 2013). So the dramatic rate of change in temperature in the Kosterfjord (4 °C in less than 24 h) is more likely to be one of the causes of this mass mortality. The population still suffers from the incident and the mortality is still high (P.C. & M.T., ROV observations at c. 80 m depth in Swedish waters of the Kosterfjord in May 2012).

Figures

Figure 7. External morphology of Geodia barretti Bowerbank, 1858. A, specimen PC10 (fixed in ethanol) from Korsfjord, Norway; specimen is 10 cm large. B, specimen ZMBN 89716 just after dredging from north of Spitsbergen, 215 m depth; specimen is 30 cm long. C, specimen cut in half just after dredging from the Polarstern ARK-XXII/1a, 2007, field# 40-4(3); specimen is 14 cm large. D, specimen PC72 just after dredging from Korsfjord, Norway, at 200–400 m depth; specimen is 30 cm long. E, specimen in situ at Traenadjupet, northern Norway, 66°58′N, 11°7′E, 292 m depth, later collected under field# 27-1(11) (picture taken by J. Schauer, Polarstern ARK-XXII/1a, 2007). F, specimen PC359 just after dredging at 1818 m depth off western Ireland; specimen is 16 cm large. G, uniporal oscules inside the preoscule of ZMBN 89722 (from Spitsbergen); each oscule has a separate sphincter. Scale bar: 1 mm. H, cribriporal pores of ZMBN 89715 (from Spitsbergen). Scale bar: 1 mm.

Figure 8. External morphology of Geodia barretti BOWERBANK, 1858. A. Balgim specimen CP63-22 cut in half, showing the preoscule full of sediments. B. Balgim specimen CP98-47 cut in half, showing the preoscule. C. Close-up of preoscule full of sediments (CP63-22). D. Close-up of preoscule (ZMBN 89722 from Spitsbergen). Scales: A: 5 cm; B-D: 2 cm.

Figure 9. Sections and spicules of Geodia barretti Bowerbank, 1858. A, thick section showing the cortex essentially made of sterrasters, and underlying large dichotriaenes. Scale bar: 1 mm. B, thick section in a preoscule showing the uniporal oscules (u.o.) with a single sphincter (s). Between the oscules, bundles of microxeas and strongylasters. Note the near absence of sterrasters. Scale bar: 500 Mm. C, microxeas, rarely modified to microstyles (ZMBN 77922, Korsfjord). Arrow points at centrotylote portion. D, dichotriaene (UPSZMC 78260, Davis Strait). E, anatriaene (UPSZMC 78260, Davis Strait). F, oxyaster II (ZMBN 77922, Korsfjord). G, strongylasters (ZMBN 77922, Korsfjord). H, oxyasters I (PC359, western Ireland). I, oxyasters I (Balgim CP98-47). J, close-up on the hilum of a sterraster. Notice the warty rosettes (ZMBN 77922, Korsfjord). Scale bar: 2 Mm. K, sterraster (ZMBN 77922, Korsfjord).

Figure 10. Distribution of Geodia barretti Bowerbank, 1858 (map made with GeoMapApp, http://www.geomapapp.org). T, type locality;?, dubious records.