Spheciospongia cf. vagabunda, Kelly-Borges & Vacelet, 1998

Marlow, Joseph, Bell, James J., Shaffer, Megan, Haris, Abdul & Schönberg, Christine Hanna Lydia, 2021, Bioeroding sponge species from the Wakatobi region of southeast Sulawesi Indonesia, Zootaxa 4996 (1), pp. 1-48 : 20-23

Spheciospongia cf. vagabunda trincomaliensis ( Ridley, 1884)

Synonymy. Spirastrella vagabunda trincomaliesis Ridley, 1884 (described as variety of Spheciospongia vagabunda in a footnote on p. 468 of original description of Spheciospongia vagabunda ).? Spirastrella vagabunda Ridley, 1884 (original description).? Cliona vagabunda Rosell & Uriz, 1997 (genus transfer, concerning main species).? Spheciospongia vagabunda Kelly-Borges & Vacelet, 1998 (genus transfer, concerning main species).? Spirastrella cylindrica Kieschnick, 1896 (presently accepted as a junior synonym of S. vagabunda ).? Spirastrella vagabunda var. fungoides Dendy, 1905 (variety of S. vagabunda , but name no longer accepted).? Spirastrella vagabunda var. gallensis Dendy, 1905 (variety of S. vagabunda , but name no longer accepted).? Spirastrella vagabunda var. tubulodigitata Dendy, 1905 (variety of S. vagabunda , but name no longer accepted).? Spheciospongia inconstans sensu Calcinai et al. (2000) . Not: Spirastrella vagabunda var. arabica Topsent, 1893 (described as variety of S. vagabunda , but not conspecific). Not: Anthosigmella vagabunda sensu de Laubenfels, 1954 (genus transfer for S. vagabunda , but not conspecific sample).

Material examined. Spheciospongia cf. vagabunda trincomaliensis : B3-OB-01, slide of discarded sponge kept at VUW, sponge from west of Hoga, Wakatobi, Banda Sea, sampled between March and August 2014, 3– 20 m, coll. J. Marlow. B3-OB-02, slide of discarded sponge kept at VUW, sponge from west of Hoga, Wakatobi, Banda Sea, sampled between March and August 2014, 3– 20 m, coll. J. Marlow. B1-OB-01, slide of discarded sponge kept at VUW, sponge from west of Hoga, Wakatobi, Banda Sea, sampled between March and August 2014, 3– 20 m, coll. J. Marlow. Slide of BMNH 1954.3.9.446, Bowerbank reference number 11/31, Spheciospongia vagabunda previously published as Suberites trincomaliensis sensu Carter (1887) . Eastern Sri Lanka, ca. 8°27’60”N 81°12’59”E, November 1881, coll. D. Johnston.

Morphology and erosion. Morphology transcending more than one clionaid growth form; in places as thick beta- or low-relief gamma-sponge, basally eroding ( Fig. 8A ). Oscular mounds at times elevated into coneshaped fistule-like structure, surface generally smooth to moderately lumpy, forming pronounced lamello-digitate projections in more sedimented environments ( Fig. 8B ). Colour of ethanol-preserved specimens retained, but dulled. Alive externally predominantly brown to dark olive. Choanosome light brown or light olive green with embedded debris and sediment ( Fig. 8C ), suggesting functional potential as delta-sponge. Individuals large, with average surface area of 100 cm 2, but occasionally up to 1.5 m 2. Large, conical oscular chimneys with wide openings of 2–7 cm in diameter, rising up to 10 cm above sponge surface, often merging with others into larger individuals. Ostia not detected. Pronounced aquaeous system with wide canals (5–10 mm in larger canals). Average erosion depth 5 cm into substrate, but occasionally penetrating down to 7 cm. Erosion chambers circular, small (cross-sectional area 1.1 mm 2 ± 0.7 SD) and distributed in honeycomb pattern, with little calcareous material remaining between chambers.

Skeletal characteristics and presence of Symbiodiniaceae . Tylostyles in palisade in ecotosome with tips pointing outwards, represented by shorter size class. Dense packing of surface spicules resulting in cortex, tylostyles below in looser, less organised arrangement and representing the longer spicule type. Spirasters largely restricted to ectosomal crusts. No evidence for photosymbionts, neither through surface fluorescence, nor through histology.

Spicules. Megascleres—Tylostyles ( Fig. 8D & E ), straight to slightly curved, highly variable in size, longer in choanosome than in surface layer. Large variability in tyle shape; generally longer than wide, mostly subterminal and only subtly developed, with occasional stylar modifications. Dermal tylostyle dimensions (min – mean – max and standard deviation): length 250 – 412.9 – 665 µm ± 112.5 SD; shaft width 5 – 7.2 – 11 µm ± 1.4 SD; and tyle width 5 – 7.2 – 10 µm ± 1.2 SD ( Fig. 8D ). Choanosome tylostyle dimensions (min – mean – max and standard deviation): length 304 – 549.8 – 709 µm ± 91.8 SD; shaft width 5 – 8.0 – 12 µm ± 1.7 SD; and tyle width 6 – 8.7 – 12 µm ± 1.5 SD (means across three Wakatobi specimens, with N = 25 spicules each; Fig. 8E ). Microscleres – Spirasters, somewhat variable in morphology, but broadly falling into two categories; 1) short-helical, or amphiaster-like spirasters with compound spines from the ectosomal crust (common), and 2) more slender, helical spirasters (very rare, likely restricted to the choanosome). Dermal spirasters with terminal clusters of spines and compact, crown-like spine bouquets aligned along convex side of spicule ( Fig. 8F ). Commonly short and straight (amphiaster-like), C-shaped or with two subtle bends. Amphiaster-like spiraster dimensions (min – mean – max and standard deviation): length 7.0 – 9.0 – 11.4 µm ± 1.2 SD; central shaft width 1.0 – 2.1 – 3.0 µm ± 0.4 SD. Choanosomal spiraster dimensions (min – mean – max and standard deviation): length 10.0 – 12.0 – 22.5 µm ± 6.0 SD; central shaft width 2.0 – 2.1 – 2.5 µm ± 0.3 SD (means across three Wakatobi specimens).

Habitat and occurrence in the Wakatobi. Very common sponge; most abundant on steep drop-offs and overhangs, less abundant at high rugosity sites dominated by morphologically complex corals, and absent from highly sedimented site. Growing across bare substrate and often in apparent competition with live corals and other benthic taxa.

Remarks. The conical-fistular shape of the exhalants, the fleshy nature of epilithic body parts and the spicule complement of size-variable tylostyles and spirasters with dense compound-spines, partly distributed as ectosomal crust, is a character combination diagnostic of sponges belonging to the genus Spheciospongia ( Rützler 2002b, and amended diagnosis above). Deciding on Spheciospongia resulted in a large range of less well-described material for comparison. By far most of the presently accepted Spheciospongia species have been reported from the Indo-Pacific, but many names have not been used since their original description, and their taxonomic characters and distributional ranges are not widely known. Moreover, many of the used names may represent erroneous identifications or species complexes (C. Schönberg, unpubl. data; see also Schönberg et al. 2017a). We therefore initially screened a broad spectrum of original descriptions of Indo-Pacific Spheciospongia species and a few Spirastrella and Cliona species , and we excluded most of these largely based on spicule characters and habit properties (Appendix II). Of those that were left, we could not easily dismiss the following species that were insufficiently described or shared some characters with our material: Spirastrella andamanensis , Spheciospongia capensis , Spheciospongia digitata , Spheciospongia florida , Spheciospongia mastoidea , and Spheciospongia vagabunda . We were unable to obtain type material for most of these species, and the resulting comparisons and decisions remain tentative.

Hesitantly we excluded Spheciospongia andamanensis as it was described with smaller oscule width and with spirasters that were noticeably longer than those in the Wakatobi sponge ( Pattanayak 2006; Appendix II). Spheciospongia capensis oscules are apparently arranged in rows, which never occurred in the Wakatobi material ( Carter 1882; Appendix II). Spheciospongia digitata shared a number of characters with our material, including the overall habit, the occurrence of subtylostyles of a similar size range and spirasters with spine clusters ( Hentschel 1909; Appendix II). We nevertheless rejected conspecificity of our material with Spheciospongia digitata , because in the latter species the oscules are in the mm range, the dominant spirasters were longer, and tylostyle tyles were mostly terminal, but subterminal in the Wakatobi sponge. The habit of Spheciospongia florida resembles that of our sponge, it is also basally eroding, has similar bioerosion chamber diameters, shares the spiraster crust, a comparable tylostyle size range, tylostyles with sometimes subterminal tyles and having short spirasters with spine clusters ( von Lendenfeld 1897; Appendix II). However, we did not readily accept conspecificity between Spheciopongia florida and our sponge from Wakatobi as the former has more robust tylostyles, but more delicate spirasters and not the longer, helical type of spirasters. Furthermore, our Wakatobi species lacks conulose fistules with apical groups of concave, blind depressions. Spheciospongia mastoidea from the Red Sea is again a species with several traits matching those of the Wakatobi sponge: fistular habit with a spreading base, wide apical oscula and similar subtylostyles ( Keller 1891; Appendix II). This species is little understood and has not been redescribed or rereported since its original description, which does not mention any spirasters, complicating the comparison. Again, we hesitated to exclude this species, but finally rejected it, based on the pronounced ring structure of the main canal walls in Spheciospongia mastoidea , which were faint in the Wakatobi sponge, if present. Spheciospongia mastoidea produces mostly single-fistule specimens, but our sponge often formed spreading clusters that commonly had more than one low-conical fistule.

The tentative exclusion of the above species left Spheciospongia vagabunda for further consideration. Indo- Pacific sponges with similar taxonomic characters as the Wakatobi sponges have previously been identified as Spheciospongia vagabunda (e.g. Sutcliffe et al. 2010), and Spheciospongia vagabunda has been repeatedly recorded from Indonesia (e.g. van Soest 1990; Barucca et al. 2007; Becking et al. 2013; Haris et al. 2014; Hadi et al. 2016; Calcinai et al. 2017). The Leiden Naturalis Museum lists several specimens from different locations in Indonesia (http://bioportal.naturalis.nl). However, this option is once more quite problematic for a number of reasons. Comparison is again extremely difficult by the different approaches various authors used to assess the tylostyle dimensions (Appendix II). Accounts for Spheciospongia vagabunda trincomaliensis are in reasonably good agreement with the Wakatobi sponge, but the description for the main species Spheciospongia vagabunda seems more dissimilar to the Wakatobi material. Spheciospongia vagabunda sensu Ridley (1884) appears to have a high affinity to Spheciospongia inconstans and was described as olive or greenish, “tending to grow up into large nodular elevations, which may bear one or more vents” and having very robust, ensiform tylostyles with well-formed tyles ( Ridley 1884; also Dendy 1905; Appendix II). Past authors have stressed that Spheciospongia vagabunda can have a highly variable morphology (e.g. Dendy 1905; Kelly-Borges & Bergquist 1988; Kelly-Borges & Vacelet 1998; Sutcliffe et al. 2010), and over time a significant number of varieties were described, which however do not all seem to be in synonymy of each other ( Ridley 1884; Topsent 1893; Dendy 1905). We propose that the present concept of Spheciospongia vagabunda is based on a species complex, likely already representing more than one species in the original description. For example, Kelly-Borges & Bergquist (1988) described coral-encrusting individuals following the contours of corals, juveniles and endopsammic individuals forming steep-sided conical projections and individuals on dead substrate forming mounds, while Ridley (1884) had commented on the consistent appearance of the habit of Spheciospongia vagabunda . Several authors have observed more than one colour morph at their sample site ( Kelly-Borges & Bergquist 1988; Richmond 1997; Kelly-Borges & Vacelet 1998), and accounts differed in whether or not Spheciospongia vagabunda has photosymbionts ( Ridley 1884; Padovan et al. 2012 versus present observations), and whether or not it has more than one tylostyle size ( Kelly-Borges & Bergquist 1988; Kelly- Borges & Vacelet 1998; Sutcliffe et al. 2010 versus historical accounts such as Dendy 1905). Obvisously, detailed re-examination of several historical specimens will be necessary to unravel these inconsistencies and to resolve the putative species complex. As this would be beyond our present aims, we settled for a tentative species identification as Spheciospongia cf. vagabunda trincomaliensis sensu Ridley (1884) .

Figures

FIGURE 8. Spheciospongia cf. vagabunda trincomaliensis. (A) Field image of adult specimen with multiple large, coneshaped exhalants; (B) field image of slightly sedimented specimen; please note the wart-like surface structure at the base of the exhalants as characteristic in Spheciospongia species; (C) incision into sponge tissue revealing a thick cortex and embedded sediment; (D & E) tylostyles; (F) spirasters.