] Marine Biology 49, 147-159 (1978) MARINE BIOLOGY © by Springer-Verlag 1978

Siliceous Spicules in Coral Reef Sediments

K. Riitzler and I.G. Macintyre

National Museum of Natural History, Smithsonian Institution; Washington, D.C., USA

Abstract

Experimental etching with hydrofluoric acid indicated that silica deposition oc­ curs in a recognizable pattern in common sponge microscleres. The postdepositional alteration of these spicules has previously been generally unrecognized or misin­ terpreted in the literature. Early stages of postdepositional etching of sponge spicules were observed in the acid insoluble fraction of sediments from the West Atlantic barrier reef near Carrie Bow Cay, Belize. Preliminary data on silica dis­ tribution in the Belize barrier reef show that concentrations in fine sediment «0.25 mm) increase landward of the main reef tract. Sponge spicules are the main component of particulate silica in sediments of the reef and fore-reef where sponge populations abound, whereas grains prevail in the back-reef lagoon deposits. Recycling of locally dissolved silica appears to be important for the growth of many off-shore reef .

Introduction Antarctic diatoms or deep-sea radio­ larian oozes. In coral reefs, sponges "The spicule which has lived, has to de­ are known to be quantitatively important cay, and may live again in another form." (Rlitzler, 1978), yet no information is (Duncan, 1881). available about their participation in Sponges contribute to reef sediments cycling silica. Therefore, we examined in two ways: (1) excavating species (e.g. the distribution and origin of silicious clionids) produce quantities of carbon­ sponge spicules in a modern coral reef ate silt by eroding coral rock, mollusk complex near Carrie Bow Cay, Belize (Cen­ shells and similar substrates (Neumann, tral America) and studied the effect of 1966; Rlitzler, 1975; Moore and Shedd, postdepositional alterations on their 1977); (2) the majority of sponge taxa morphology. We attempted to reproduce (main exception, order Keratosa) form patterns of suspected dissolution in the spicules that are liberated when the ani­ laboratory and searched the literature mals decay. Sponges use two mineral ma­ on recent and fossil deposits for docu­ terials to produce skeletal spicules. In mentation of similar occurrences. Final­ the class Calcarea, the simple diactines, ly, we looked for sources of dissolved triactines and tetractines are composed silica needed for mineralization pro­ of calcite. Calcareous sponges make up cesses in growing sponges. only a small percentage of the sponge biomass in reefs. Members of all other sponge classes (Demospongea, Sclero­ Materials and Methods spongea, Hyalospongea) produce spicules of opaline silica. Clean siliceous spicules were separated In the sea, diatoms, radiolarians, from sponge tissues by boiling in con­ sponges and silicoflagellates partici­ centrated nitric acid and thorough rins­ pate in the mineralization of silica. Al­ ing in distilled water. Sterrasters were though substantial sponge spicule mats obtained from Geodia cydonium (Jameson) 2 m thick and larger have been observed (Adriatic Sea) and G. neptuni (Sollas) in Antarctica (Koltun, 1968; Dayton et (Belize), spherasters from Chondrilla nucu­ al., 1974), on a worldwide basis such de­ la Schmidt (Belize) and selenasters from posits rank behind accumulations such as Placospongia carinata (Bowerbank) (Gulf of

0025-3162/78/0049/0147/$02.60 148 K. Rlitzler and I.G. Macintyre: Sponge Spicules in Reef Sediments

Mexico). Scanning electron microscope cause of their low specific gravity, preparations were made by drying spic­ they are readily washed away and eventu­ ules on small circular cover slips which ally, after periods of weeks or months, were mounted on stubs, coated with car­ they are degraded by bacterial action. bon and, subsequently, with platinum. A Spicules, on the other hand, after hav­ Cambridge Stereoscan Mark II A was em­ ing been freed from organic binding sub­ ployed at 100 x to 8000 x primary magni­ stances, sink to the bottom and become fications. part of the sediment. Sections of spicules were made by The great variety of forms exhibited grinding and polishing material embedded by sponge spicules has led to a complex in epoxy. These were etched for 30 sec terminology of some 100 names that ex­ or 5 min in 6% hydrofluoric acid (HF), press size, shape and symmetry condi­ with gentle agitation. Whole spicules tions (Reid, 1968). Comparatively large were also etched for 5 min and 30 min in megascleres, which form the main skele­ 24% HF. tal architecture, are distinguished from small microscleres scattered throughout Sediment samples were taken at 10 lo­ the tissue or concentrated in an ecto­ cations near Carrie Bow Cay (Fig. 6), somal layer (cortex). Megascleres are using a 5 cm diameter hand corer to a grouped according to the number of their depth of 5 cm. Triplicate cores from axes: monaxons, triaxons and tetraxons. each station were pooled, the intersti­ Monaxons and tetraxons occur in reef tial sea water drained through filter pa­ sponges, whereas triaxons are restricted per, and sediment and water carried sep­ to the glass sponges (Hyalospongea) of arately to the laboratory in plastic con­ the deep sea. Microscleres have been tainers. Sand was rinsed in fresh water separated into three major groups: poly­ and dried (600 C) for storage. For fur­ actines (astrose), diactines (sigmatose) ther processing, 3 to 5 g portions were and microrhabds (Dendy, 1921a; Brien, soaked in 10% sodium hypochlorite over­ 1973); all occur in reef sponges. night to remove organic matter. After thorough rinsing, they were oven-dried Spicules are intracellular secretions to constant weight (600 C), sieved by single scleroblasts. Ultrastructural through 250 urn mesh, and the two frac­ studies of their formation have been tions weighed again to 1 mg. The samples confined to monaxon forms in which si­ were then decalcified in 10% hydrochlo­ licic acid (H2Si307) is deposited in con­ ric acid and the insoluble portions centric layers around a non-collagenous rinsed, dried and weighed. Percentage proteinic axial filament. The concentric estimates of the siliceous main compo­ stratification has been explained by the nents were made under a high power ste­ presence of organic lamellae (Garrone, reomicroscope. 1969) and by differences of hydroxyla­ tion caused by periodic pauses in growth Reactive silica in interstitial sea (Schwab and Shore, 1971). Factors such water was determined in the field using as pH, temperature, Si02 and C02 concen­ commercial test kits with 0.1 to 1.0 mg tration and available surface area are and 1 to 40 mg Si02 1-1 ranges (Hach thought to govern the rate of spicule Chemical Company, USA), calibrated growth (J¢rgensen, 1944; Elvin, 1971; against a graded series of standards. Pe, 1973). The standards were prepared from synthet­ ic sea water with dried sodium silico­ Microscleres exhibit an almost unlim- fluoride (Strickland and Parsons, 1968). i ted variety of shapes and, where present, are of great diagnostic value for demo­ sponge classification (Reid, 1968). Un­ fortunately, because of their small size Results and Discussion and delicate structure they are less well preserved in sediments and also Structure and Origin of Sponge Spicules Common more easily overlooked than most of the in Reef Sediments more uniform megascleres. Exceptions are two types of microscleres: sterrasters Classification of sponges, from classes and selenasters. Although their struc­ to subspecies, is based primarily on ture and origin differ, they are both their , which is the only compo­ spherical or subspherical, massive and nent preserved in most fossils and many comparatively large. Unlike most micro­ older museum collections. Spongin fibers scleres they play an important part in and mineral spicules, separate or in skeleton formation and, subsequently, in combination, are the main constituents the composition of reef sediments. This of the sponge skeleton. After death, the study is concerned principally with the collagenous fibers resist decay much structure and postdepositional alter­ longer than the cellular tissue but, be- ation of these two types of microscleres. K. Rutzler and I.G. Macintyre: Sponge Spicules in Reef Sediments 149

Sterrasters wise solidly cemented rays. The spines are interconnected by ridges, or trabe­ Sterrasters, which range from 40 to culae (Fig. 1: 5). Selenasters differ 300 ~ in diameter, are characteristic basically from sterrasters in their de­ of the choristid genus Geodia (Geodi­ velopment from a monaxial microsclere idae), in which they form a densely with a straight axis covered by minute packed cortex several millimeters thick. spines (Fig. 1: 6 and 7). They have also They represent an extreme development of been named sterrospirae (spinispirae, astrose spherical or subspherical spic­ for early stages) to express the super­ ules of almost solid silica, densely ficial resemblance of sterrasters and covered by short projections that termi­ the common origin with spirasters, a nate with recurved spines (Fig. 1: 1 and spiny sclere with spirally twisted axis 2). The sterrasters originate as eu­ (Vosmaer and Vernhout, 1902). Although asters with globular centrum and numer­ concentric layering of silica can be ous pointed rays (Fig. 1: 3). Our pol­ made visible by etching (Fig. 2: 7,8,9), ished and slightly etched cross-sections there is no trace of radiating axial ca­ (Fig. 2: 1-6) illustrate, by the pres­ nals. Apparently the mineral is depos­ ence of radiating axial canals, that sil­ ited directly onto a primordial spiny ica deposition started on the surface of rod of silica. Selenasters are the only a multirayed proteinic structure compa­ sponge spicules known in which a varying rable to the axial filament of monaxon percentage is colored (reddish to brown) . megascleres. Silica continues to be de­ The of the stable pigment, compa­ posited around the rays and eventually rable to that found in calcareous spic­ fills the space between them to form a ules of alcyonaceans, and the way it is bluntly spined spheraster. Deposition of incorporated in the opal, are not known the final 5 to 15 ~ layer of opaline (Vosmaer and Vernhout, 1902). silica, including the ornamentation of the spines by star-like projections, is a secondary process apparently unrelated Silica Dissolution ana its Effect on Spicule to the axial filaments of the rays (Fig. Morphology 2: 3 and 5). At one point on the surface the pattern is interrupted by a depres­ With an average concentration of 1 mg Si sion (hilum) (Fig. 1: 1) that marks the 1-1 sea water, the oceans contain only position of the scleroblast nucleus (Sol­ 2% of the soluble amount of this element. las, 1888) • Values are higher in nearshore waters A related type of spicule is the where terrestrial run-offs are rich in spheraster, a common microsclere among , and in other areas where par­ members of the choristid suborder Astro­ ticulate silica abounds, e.g. intersti­ phorina and the only spicule type of Chon­ tial water of siliceous sediments and drilla spp. (Chondrosiidae), a group of deep-sea habitats (Riley and Chester, uncertain systematic position close to 1971). the Hadromerida. It is similar in shape The unstable skeletal opal, if not to an earlier stage of sterraster, but protected, dissolves quickly after death the spherical center supports only a few of the producing organisms, with the low rays rather than numerous spines more delicate forms like silicoflagel­ (Fig. 1: 4). Aspidasters of Erylus spp. lates and diatoms disappearing first (Ar­ (Geodiidae) are also homologous to sterr­ rhenius, 1963; Hurd, 1972). This process asters, with the same surface ornamenta­ is delayed in sediments with intersti­ tion, but have flattened oval or lozenge tial water that is enriched with respect shapes. SpiCUles similar to spherasters to dissolved silica, a condition that of sponges are also produced by didemnid depends on the presence of particulate , but should not be mistaken silica and on the rate of water exchange. for sponge spiCUles since they consist Friedman et al. (1976) observed that of calcium carbonate. quartz particles are corroded during ce­ mentation in Red Sea reefs and that high micro-environmental pH values which Selenasters largely control carbonate precipitation mi~ht well be responsible for quartz These spicules, which at maturity mea­ etching. They estimated that physiologi­ sure 75 ~ x 60 urn in average, occur ex­ cal activity of could create thin clusively in the family Placospongiidae alkaline water films exceeding pH 10. (Hadromerida) and, like sterrasters, Similarly, Macintyre (1977) noted that form a cortical armature. They are ovoid the extent of alteration of the surfaces to bean-shaped, have a hilum on the con­ of selenasters collected from cores in cave surface, and are covered by low the fringing reef off Galeta Point (Pa­ spines - the free distal points of other- nama) is related to the degree of sub- 150 K. Rutzler and I.G. Macintyre: Sponge Spicules in Reef Sediments

Fig. 1. Scanning electron micrographs of fresh sponge microscleres. 1: Sterraster from cortex of Geodia cydonium (1000 xl. 2: Sterraster, G. neptuni (1000 xl. 3: Immature sterraster (euaster-stagel, G. neptuni (1000 xl. 4: Spheraster, Chondrilla nucula (1200 xl. 5: Selenaster from cortex of Placo­ spongia carinata (1000 xl. 6: Immature selenaster, P. carinata (1000 ~l. 7: Two early stages (spinispirael of selenasters, P. carinata (1500 xl Fig. 2. Ground and polished cross-sections through sterrasters and selenasters (1-3, 7 and 8: etched 30 sec in 6% hydrofluoric acid, HF; 4-6, 9: etched 5 min in 6% HFl. 1: Sterraster, Geodia neptuni; the crack is an artifact from grinding (1100 xl. 2: Enlarged center and radiating canals (3000 xl. 3: Enlarged spine with central canal showing stratified silica deposition and separate surface ornamentation (6500 xl. 4: Enlarged radiating canals by increased etching (1300 xl. 5: De­ tail of distal portion with distinct stratification; most of the surface ornamentation is lost (6100 xl. 6: Tangential section through sterraster; note two fused rays with central canals and stratified silica filling spaces between rays (9000 xl. 7: Selenaster from Placospongia carinata sectioned through hilum; cracks and pits are artifacts from grinding (1000 xl. 8: Stratified silica deposition; note absence of axial canals (3800 xl. 9: Strong etching and mechanical stress during grinding caused the silica to crack along one layer, outlining shape of original spine (7600 xl; inset: tangential section through similar area (5600 xl 152 K. Rutzler and I.G. Macintyre: Sponge Spicules in Reef Sediments

marine lithification of the surrounding etched spicules, this canal reef sediments. In carbonate sands of a is triangular in cross-section, and re­ Jamaican reef complex, Land (1976) found flects the shape of the axial filament less biological silica in fore-reef sedi­ as well as potential symmetry develop­ ments (0.3%) than in carbonate mud be­ ment (Reiswig, 1971). The originar diam­ hind the fringing reef (1.6%). He ob­ eter of the canal varies with the sys­ served a higher number of etched spic­ tematic position of the sponge between ules in the fore-reef and suggested that approximately 0.5 and 3 ~, and does not the variation in silica content was re­ exceed 10% of the spicule diameter. Etch­ lated to higher opal destruction in en­ ing occurs both inside and on the surface vironments that favor submarine carbon­ of a fragmented sclere, and the rate of ate cementation. dissolution can be related to the ratio Most sponge spicules are still recog­ of spicule to axial canal diameters. nizable even when extensively eroded by dissolution. In high-energy reef sedi­ As a result of physiological defects ments, megascleres are soon broken up to or certain environmental conditions, spic­ short cylindrical fragments (occasional ule erosion can be observed even in live forked pieces derived from tetraxons) . sponges. Resorption of siliceous spicules They appear symmetrical, glassy clear in within a sponge (Dendy, 1921b, p. 64) or transmitted light, and have an axial ca­ in the dense aragonitic of nal (Fig. 3: 1 and 2). In fresh, un- sclerosponges (Hartman and Gareau, 1970)

Fig. 3. Sponge spicules in reef sediments. 1: Typical acid-insoluble residue from Carrie Bow Cay fore-reef, Station 1; Geodia sp. sterrasters, some smaller asters and fragments of monaxon and tetraxon megascleres are evident (1300 x). 2: Broken monaxon spicule showing fair amount of etching (enlarged axial canal, enhanced stratification) (3500 x). 3: Surface detail of etched sterraster; surface ornamentation is removed, axial canals are exposed (3200 x). 4: Similar stage of spheraster (4000 x). 5: Similar stage of selenaster (3200 x) K. Rutzler and I.G. Macintyre: Sponge Spicules in Reef Sediments 153 occurs within the living organism and is ridges and become increasingly pitted unrelated to degradation of loose spic­ (Fig. 3: 5; Fig. 5). They decrease in ules in reef sediments. size as the surface is eroded but, be­ Among the microscleres only spher­ cause they lack radiating axial canals asters, sterrasters and selenasters sur­ to promote dissolution, they are proba­ vive to serve as a measure of opal de­ bly the sturdiest of all biogenic opal struction in a given sediment or facies in the sea. As mentioned earlier, Mac­ because their shape and size render them intyre (1977) found both slightly cor­ quite resistant to mechanical destruc­ roded Placospongia sp. selenasters in un­ tion. We have determined morphological lithified sediments and strongly etched changes caused by etching in fresh spic­ ones in cemented sections in a modern ules exposed to dilute hydrofluoric acid. reef off Galeta Point (Panama). Similar­ The resulting erosion patterns can be ly, Weaver and Beck (1977, their Figs. compared with those commonly found in 52 and 82) showed unaltered and partial­ spicules from reef sediments (Fig. 3: 3, ly dissolved sponge spicules from Middle 4, 5; Fig. 4; Fig. 5). The true asters Miocene clay sediments (SE United first lose their surface ornamentation States), including fragmented mega­ and tips of rays, thus exposing axial ca­ scleres and "bean-shaped microscleres" nals that become enlarged as erosion pro­ which are Placospongia sp. selenasters. gresses (Fig. 3: 3, 4; Fig. 4). Heavily These etched and corroded Placospongia etched sterrasters are easily crushed in spp. selenasters have, however, given shallow agitated waters (Fig. 4: 3, 4). rise to some taxonomic confusion. Al­ Selenasters soon lose their surface though recent Placospongia species and

Fig. 4. Corrosion patterns of Geodia neptuni sterrasters. 1: Three stages of HF etched spicules; total exposure time was 5 min in 24% HF, different degree of etching is due to inadequate mixing (800 xl. 2: Similar stages from reef sediment, Carrie Bow Cay fore-reef Station 1 (300 xl. 3: Ad­ vanced stage of etching from sediment (700 xl. 4: Heavily etched fragment from sediment (800 xl 154 K. Rutzler and I.G. Macintyre: Sponge Spicules in Reef Sediments

Fig. 5. Corrosion patterns of Placospongia carinata selenasters. 1: Etched 5 min in 24% HF (850 xl. 2: Similar stage from reef sediment, Carrie Bow Cay fore-reef Station 1 (950 xl. 3: Etched 30 min in 24% HF (1400 xl. 4: Similar stage from sediment (1400 xl

their characteristic microscleres had roded surface ornamentation. Schinde­ already been described (Gray, 1867), wolf's (1967) study of em­ Hinde (1890) introduced a genus Rhaxella bedded in craspeditid ammonites from the for Jurassic sponge fragments containing Upper Jurassic of Russia reported mainly bean-shaped microscleres with smooth to sponge spicules, spherasters (possibly verrucose surfaces. Schrammen (1936) il­ carbonate and therefore almost certainly lustrated rock-forming rhaxes, together belonging to didemnid tunicates), and with spherasters, from Swalian bedded radiolarian tests. Of the sponge spic­ facies (his Plate 19, Figs. 13, 15, 18), ules reported by Schindewolf (1967: his and speculated that they might be car­ Plate 3, Figs. 2-4; Plate 4, Figs. 1-3; riers of reproductive material. Working Plate 9, Fig. 14; Plate 12, Fig. 2a), on the same facies, Reif (1967) con­ rhaxes are certainly selenasters or im­ sidered rhaxes a special form of sterr­ pressions left by this spicule type. A asters, without rays, and noted particu­ smooth sphere (Schindewolf, 1967: his larly the radial structure and verrucose Plate 9, Fig. 15) could be similar to surface of comparative material from the the spheres that occur in the recent ge­ Eocene of Barbados. He also illustrated nus Caminus (Geodiidae), or a corroded spherasters (his Plate 15, Figs. 14 and spheraster. Other "spheres" (his Plate 15) and a "sterraster" (his Plate 15, 9, Figs. 16 and 17) are etched sterr­ Fig. 18). The latter looks more like a asters or spherasters. Spherasters (his verrucose ball, or sterrospheraster, as Plate 10, Figs. 6-12) closely resemble it occurs in recent representatives of those of recent chondrillids or didem­ the genus Aurora (Ancorinidae), but it nids (the nature of the original materi­ may be a true sterraster having a cor- al is unknown). Spherical forms with hol- K. Rutzler and I.G. Macintyre: Sponge Spicules in Reef Sediments 155

:, C) Cross ," "'" Coy c:SJ '-.,V' t? ·'.J9Garbutts /' Coy c "0 E 1? g o ___,Toboecoy'Range : Tobacco Coy ~. Tobacco Coy COCY/'Plum:".... '" Entrance Ca,Y# E o ~ :~Rogged /"/ Coy 16°50'N ,..- TwinA:.'I'l:--+'H+----- ~. , Bluer) Ground', 5 6 Corrie Bow Coy Range::' , ...... --===;-- /. ""'~::::\"~ Curlew Bonk ... o~/ .:.: , ell" .", Stewart '01 i.../" :' .' Coy {I ..~Wee Wee o'''' !': Coy ,:: / ...:' :';. ~ :' :' /,: ( ,'~

Fig. 6. Map of Belizean barrier reef and lagoon near Carrie Bow Cay. Sample numbers correspond with those in Table 1, and sampling stations are described in text

low centers and radiating tubes (Plate confirm the nature of OST as eroded 12, Figs. 1-4; Plate 13, Figs. 3-5) iden­ geodiid sterrasters. tified by Schindewolf (1967) as "Radio­ laria or spheres of Porifera" cannot be accepted as sponge spicules because they Silica Distribution in Shallow-Water Sediments show a distinct outer double wall. The of the Barrier Reef Shelf near Carrie Bow Cay, taxonomic affinity of rhaxes is still Belize debated among paleontologists (F. Wieden­ mayer, Basel, personal communication). Carrie Bow Cay (formerly Ellen Cay) is a Our present findings leave no doubt that small islet (120 m x 35 m) composed of they are etched selenasters of placo­ reef sand and rubble located on the spongiids. outer edge of the Belizian barrier reef Spheroid opaline bodies, commonly as­ (formerly British Honduras), 22 km south­ sociated with sponge spicules in Miocene east of Stann Creek (Dangriga). The and Recent sediments of Japan and termed nearest land is Sittee Point, 18 km due "OST" shells by Japanese workers, were west, where the Sittee River empties in­ described as a laracariid radiolarian, to the lagoon. The closest island inside Hataina ovata Huang (1967). Inoue and Iwa­ the lagoon is Twin Cays, 2.6km north-west saki (1975) debated the radiolarian na­ of Carrie Bow Cay. Ten samples of bot­ ture of OST shells and pointed to the tom sediments were taken and analyzed striking similarity with sterrasters of for dissolved (reactive) and particulate Geodia spp., although the inner structure silica (Fig. 6, Table 1) from: (1) sand of superficially fractured sterrasters trough, outer fore-reef; (2) spur and appeared solid under the scanning elec­ groove zone, inner fore-reef; (3) reef tron microscope and did not show the ra­ crest; (4) reef flat, between reef crest dial "tubules" of OST. The outer portion and Carrie Bow Cay; (5, 6) outer lagoon, of axial canals in sterraster rays are west shore of Carrie Bow Cay; (7) outer almost completely filled and can only be lagoon patch reef; (8, 9) inner lagoon, made clearly visible by slight etching east shore of Twin Cays; and (10) main­ (Fig. 2: 1 and 3). Our observations now land beach, north of Stann Creek Town. 156 K. Rutzler and I.G. Macintyre: Sponge Spicules in Reef Sediments

Table 1. Distribution of silica in Carrie Bow Cay sediments. Terr.: terrigenous sand; Sterr.: sterrasters, selenasters; Mega.: large fragments of megascleres, >50 ~m x 20 ~m; Misc.: miscel­ laneous small debris

Sample Location Reactive Si Sediment size Particulate Silicious components no. and depth (as mg Si02 (dry weight %) silica (dry (estimated %) 1-1) >250 ~m <250 ~m weight % of Terr. Sterr. Mega. Misc. <250 um fraction)

Fore reef 0.9 79.2 20.8 1.4 o 27 53 20 trough, 24 m 2 Fore reef 0.2 97.3 2.7 1.1 o 33 35 32 groove, 4 m 3 Reef crest, 0.3 96.0 4.0 0.7 o 25 35 40 1 m 4 Reef flat, 0.4 96.5 3.5 0.6 o 23 29 48 0.5 m 5 outer lagoon, 0.3 97.6 2.4 0.8 91 o 8 intertidal 6 outer lagoon, 0.2 98.4 1.6 0.5 60 o 5 35 0.5 m 7 outer lagoon 0.5 77.1 . 22.9 2.0 13 30 30 27 patch reef, 4 m 8 Inner lagoon, 1.5 86.6 13.4 3.0 92 2 5 1 m 9 Inner lagoon, 1.2 81.3 18.7 3.7 87 5 7 intertidal 10 Mainland beach, 3.0 84.8 15.2 98.7 99 o o intertidal 11 outer reef, 0.3 o o o o o o o surface water

For comparison, silica in surface water fusion of silica dissolved from inter­ (11) over the outer reef was also deter­ stitial particles in sediments is con­ mined. sidered an important input also in the Undersaturation of surface and inter­ deep sea (Fanning and Pilson, 1974). stitial seawater in respect to reactive Particulate silica of whole samples silica is evident in all samples (Table varied from 0.02 to 0.29% dry weight for 1). In the lagoon, terrestrial influence reef sands, 0.01 to 0.69% for lagoon sed­ extends to Twin Cays, 16 km from the iments. Microscope examination showed nearest mainland shore. Outer reef sur­ that most fragments (96 to 100%) were face water contained 0.3 mg Si02 per less than 250 urn in diameter. It follows liter, compared to 0.2 to 0.9 mg for in­ that the content of particulate silica terstitial water in the same area, 0.2 in whole samples greatly depends on the to 0.5 mg in the outer lagoon, and 1.2 size distribution of reef sediments and 1.5 mg in the inner lagoon near Twin (i.e., the proportion of grain sizes Cays. Mainland shore water (high-energy above and below approximately 2 phi), quartz beach sand) contained 3.0 mg Si02 which may explain Land's (1976) high con­ per liter. Sample number and frequency centration of silica in the muds of Dis­ were insufficient to determine whether covery Bay (1.6%) and Florida Bay (1.0%). local variations were due to mixing of We therefore determined the opal weight low silica ocean and enriched lagoon wa­ percentages of the sediment fractions ters or if stagnant interstitial water that pass through a 250 urn mesh sieve had picked up silica by dissolution of and consider these to be a more objec­ skeletal silicic acid. The latter mecha­ tive measure of distribution. Qualita­ nism must be considered the cause of the tive composition of siliceous particles comparatively high value (0.9 mg) in the indicates their origin (Table 1). In our deeper fore-reef (Sample No.1). This reef sediments, sponge spicules ac­ possibility was confirmed by exposing counted for most of the siliceous mate­ pure sponge spicules (acid-boiled sterr­ rial. Values were highest in the fore­ asters from Geodia neptuni cortex) to reef (1.4 and 1.1%), where sample loca­ silica-free artificial sea water (pH 5.7) tions were close to large sponge popula­ as a 4% suspension in a plastic bottle. tions. In lagoon deposits, silica in­ The sea water picked up 0.55 mg Si02 1-1 creases shoreward of the main reef to­ h- 1 (a total of 11 mg 1-1 in 20 h). This wards the mainland, but most of the mate­ indicates that even under open cycle con­ rial is terrigenous - except for a patch­ ditions in the sea interstitial water reef station with massive sponges nearby. can be considerably enriched during pe­ Although the postdepositional solu­ riods of low water exchange rates. Dif- tion of siliceous spicules affects the K. Rutzler and I.G. Macintyre: Sponge Spicules in Reef Sediments 157

distribution of silica in reef sediments, Table 2. Spicule content of 10 species repre­ the degree of control in various parts senting the most common silicious sponges at of the reef habitat could not be as­ Carrie Bow Cay sessed relative to other controlling fac­ Species Spicule content tors - including the removal or introduc­ (% of dry tion of spicule-rich fine deposits by weight) wave and current action, or the in situ Haliclona rubens (Pallas) 20.8 production of carbonate and siliceous Xestospongia muta (Schmidt) 57.6 skeletal material. Gelliodes ramosa (Carter) 22.4 Callyspongia vaginalis (Lamarck) 3.8 Agelas conifera (Schmidt) 19.3 Role of Sponges. in Recycling Reef Silica Neofibularia nolitangere (Duchassaing & Michelotti) 17.1 Abundance of silica in a reef mainly de­ Iotrochota birotulata (Higgin) 2.5 pends on the distance from terrestrial Hemectyon ferox (Duchassaing & Michelotti) 13.4 inputs, the current regime determining Spheciospongia vesparium the dispersal of all dissolved and par­ (Lamarck) 58.7 ticulate matter, and the aggregation of Geodia neptuni (Sollas) 67.1 organisms producing siliceous skeletons. Average 28.3 Diatoms may contribute significantly to some lagoon habitats (P. Hargraves, per­ sonal communication), but sponges are the' chief component in the cycling of Reiswig (1973) for a Jamaican reef. From this mineral. The amount of silicic acid his figures we calculate the standing required to maintain a sponge population crop of Mycale sp. as 0.4 g dry weight depends on sponge biomass, species com­ per m2 acceptable substrate and for Te­ position, and mineralization rates. thya crypta as 13.3 g m-2 (1 m depth) and Standing crop and composition of 25.4 g m-2 (3 m depth). Weights 'of spic­ sponge communities vary greatly in dif­ ules contained in these sponges are ferent parts of the Carrie Bow reef com­ 0.01 g m-2 for Mycale sp., 3.4 g m- 2 and plex. In the shallow lagoon, suitable 6.5 g m-2 for T. crypta. conditions exist where breaks in the Rate and efficiency of silica deposi­ reef barrier permit flow of strong tidal tion in marine sponges is virtually un­ current without the destructive action known. Experimental data from fresh­ of large waves or swell. There we distin­ water sponges (Ephydatia muelleri) show that guish a soft-bottom community, 1 to 6 m 1 mg Si02 1-1 water is sufficient for deep, dominated by Thalassia testudinum sea­ spicule formation, although higher depo­ grass and by a few species of large sition rates are obtained with the "nor­ sponges (Ircinia spp., Spheciospongia vespa­ mal" (ambient) 5 mg concentration (Elvin, rium), and patch reefs, 3 to 5 m deep, 1971). Another spongillid (E. fluviatilis) composed of large scleractinian coral from a 15 mg Si02 1-1 habitat produced heads, gorgonians and sponges. In the higher spicule numbers with reduced di­ high-energy shallow back-reef, reef­ mensions in impoverished (1.5 mg) cul­ crest and fore-reef zones, the sponge ture water, and a smaller number of fauna is restricted to small incrusting, thicker spicules in an enriched (60 mg) excavating or cryptic forms. In the medium (Pe, 1973). These observations deeper part of the inner fore-reef (spur were made during time spans of hours to and groove zones, 4 to 12 m) and on the a few days after germination of gemmules. outer fore-reef (12 to 60 m), the sponge Field studies of older and more estab­ fauna becomes increasingly more diverse lished populations from lakes and streams and important in biomass. A preliminary in Wisconsin (USA) demonstrated that census of standing crop of siliceous silicon requirements are species-depen­ sponges in these habitats shows a range dent (Jewell, 1935). Some forms do not of 1 to 2000 g wet weight per m2 suit­ thrive in water with less than 2 mg Si02 able substrate (Rtitzler, in preparation) . 1-1 , others have strongly reduced skel­ Using conversion factors averaged for 10 etons if growing in concentrations of common species, we estimate dry weight less than 0.7 mg. Conditions in fresh as 20% of wet weight (Rtitzler, 1978), water, however, can be more fluctuating and spicule content as 28% of dry weight and extreme than in the sea because re­ (Table 2). Hence, skeletal silica con­ duction of silicon may well be paral­ tained in live reef sponges can range be­ leled by low levels of other important tween the orders 0.1 to 100 g m- 2 . Land elements. If the requirements of marine (1976) estimated several grams of silica species are similar to those of spongil­ in sponges per square meter of living lids, growth of off-shore coral reef reef. More accurate biomass calculations sponges is highly dependent on recycling for two siliceous species were given by locally dissolved silicic acid. Marine 158 K. Rutzler and I.G. Macintyre: Sponge Spicules in Reef Sediments sponges may also have retained a more ef­ proved identical with rhaxes, an impor­ ficient silicon metabolism in a medium tant of hitherto debated or­ that is more chemically balanced. Re­ igin. duced dimensions and erosion of spicules Because of their size, the sponge inside live marine sponges do occur, but spicules are restricted to the fine sedi­ have yet to be correlated with ambient ment fractions (less than 250 ~). Par­ silicon levels. ticulate silica in this fine fraction In situ pumping activity, growth rate ranges from 0.4 to 2% dry weight in the and skeletal silica in a sponge give a reef habitats, and from 3.0 to 3.7% in measure of silica uptake, but reliable lagoon mud. High values occur in the data are scarce. The best example is a reef near large sponge populations, detailed study of a population of Tethya where silica consists predominantly of crypta in Jamaica (Reiswig, 1973). This sponge spicules, whereas terrigenous species contains spicules amounting to sands make up most of the noncarbonate 26% of total dry weight. Observation of material in the lagoon. 30 individuals over a period of 11 The distribution of dissolved silica months showed that an average specimen in interstitial water samples from the (0.8 1 volume) produces 2 g skeletal sil­ barrier reef near Carrie Bow Cay showed ica per year. An individual of this size Si02 values of 0.2 to 0.9 mg 1-1 for the is able to pump at least 3.5 m3 water seaward reef, and 1.2 to 3.0 mg 1-1 for through its body each day. A calculation lagoon habitats. Differences are attrib­ using the lowest level of reactive sili­ uted to terrestrial inputs, to local dis­ con in Carrie Bow reef waters (0.2 mg solution of solids, and to mixing with Si02 1-1) shows that, during a year, T. impoverished oceanic surface water. crypta circulates water through its body Data on Si02 requirements for spicule containing 126 times the amount of silic­ formation are available only for fresh­ ic acid used for skeleton mineralization. water sponges, but suggest that marine Experimental work with marine species is sponges in off-shore reefs might depend needed to learn how much of this quanti­ on recycling locally dissolved silica ty can actually be converted to spicules. unless they have a higher metabolic ef­ ficiency in extracting Si02 from the wa­ ter they pump. Conclusions Standing crop of the population on the Carrie Bow reef tract varies between Siliceous spicules (megascleres, micro­ 1 and 2000 g wet weight per m2 suitable scleres), important for classification, substrate. An estimated 0.1 to 100 g si­ are produced by sponges as part of their liceous spicules per m2 are contained in skeleton (main exception, Calcarea, Kera­ live sponges. tosa) , and are released to form sediment grains. Etching and dissolution, aided Acknowledgements. Field work for this study by mechanical abrasion of free spicules, was supported by a grant from the Exxon Corpora­ occur after death and decay of the tion. Mr. W.R. Brown and Miss M.J. Mann operated sponge owing to silicon undersaturation the scanning electron microscope. We thank Dr. of shallow reef waters. F. Wiedenmayer, Basel, for drawing our attention Two kinds of sponge microscleres ­ to the literature on rhaxes. Contribution No. sterrasters and selenasters - were found 40, Investigations of Marine Shallow Water Eco­ to be reliable indicators of silica ero­ systems Project, Smithsonian Institution. sion in recent and fossil sediments. Both are abundant, massive, and compara­ tively large, and reflect differential Literature Cited etching by revealing patterns of origi­ nal silica deposition. Structural Arrhenius, G.: Pelagic sediments. In: The sea. changes observed in sediment samples Vol. 3. The earth beneath the sea history, could be replicated with fresh spicules pp 655-727. Ed. by M.N. Hill. New York and exposed to hydrofluoric acid. Sterr­ London: Wiley-Interscience 1963 asters from the cortex of choristid Brien, P.: Les , morphologie et re­ sponges (Geodia spp.) are the extreme de­ production. Traite Zool. 3, 133-461 (1973) velopment of astrose (polyactine) spic­ Dayton, P.K., G.A. Robilliard, R.T. Paine and L. ules. Etched sterrasters, previously mis­ B. Dayton: Biological accommodation in the identified as radiolarians ("OST"), ex­ benthic community at McMurdo Sound, Antarc­ pose radiating axial canals. Selenasters tica. Ecol. Monogr. 44, 105-128 (1974) of hadromerid Placospongia spp. develop Dendy, A.: The tetraxonid sponge spicule, a from a spiny diactin to a bean-shaped study in evolution. Acta zool., Stock. 2, 95­ spicule without radiating canals. Cor­ 152 (1921a) roded selenasters showed a characteris­ Report on the Sigmatotetraxonida collected by tic pitted surface structure and were H.M.S. "Sealark" in the Indian Ocean. Trans. K. Rutzler and I.G. Macintyre: Sponge Spicules in Reef Sediments 159

Linn. Soc. Lond. (Ser. 2. Zool.) 18, 1-164 production. Proc. 3rd into Symp. coral Reefs (1921b) 2, 500-505 (1977). Ed. by D.L. Taylor. Miami: Duncan, P.M.: On some remarkable enlargements of School of Marine and Atmospheric Sciences, the axial canals of sponge spicules and their University of Miami causes. Jl R. microsc. Soc. (Ser. 2) 1, 557­ Neumann, A.C.: Observations on coastal erosion 572 (1881) in Bermuda and measurement of the boring rate Elvin, D.: Growth rates of the siliceous spic­ of the sponge, Cliona lampa. Limnol. Oceanogr. ules of the freshwater sponge Ephydatia muel­ 11, 92-108 (1966) leri (Lieberkuhn). Trans. Am. microsc. Soc. Pe, J.: Etude quantitative de la regulation du 90, 219-224 (1971) squelette chez une eponge d'eau douce. Archs Fanning, K.A. and M.E.Q. Pilson: The diffusion Biol., Bruxelles 84, 147-173 (1973) of dissolved silica out of deep-sea sediments. Reid, R.E.H.: Microscleres in demosponge classi­ J. geophys. Res. 79, 1293-1297 (1974) fication. Paleont. Contr. Univ. Kans. 35, 1­ Friedman, G.M., S.A. Ali and D.H. Krimsley: Dis­ 37 (1968) solution of quartz accompanying carbonate Reif, W.-E.: Schwammspicula aus dem weiBen Jura precipitation and cementation in reefs: exam­ Zeta von Nattheim (Schwabische Alb). Paleonto­ ple from the Red Sea. J. sedim. Petrol. 46, graphica (Abt. A) 127, 85-102 (1967) 970-973 (1976) Reiswig, H.M.: The axial symmetry of sponge spic­ Garrone, R.: Collagene, spongine et squelette ules and its phylogenetic significance. Cah. mineral chez l' eponge Haliclona rosea (0. S.. ) Biol. mar. 12,505-514 (1971) (Demosponge, Haploscleride). J. Microscopie Population dynamics of three Jamaican Demo­ 8, 581-598 (1969) spongiae. Bull. mar. Sci. 23, 191-226 (1973) Gray, J.E.: On Placospongia, a new generic form Riley, J.P. and R. Chester: Introduction to ma­ of Spongiadae in the British Museum. Proc. rine chemistry, 465 pp. London and New York: zool. Soc. Lond. 1867, 127-219 (1867) Academic Press 1971 Hartman, W.D. and T.F. Goreau: Jamaican coral­ Rutzler, K.: The role of burrowing sponges in line sponges: their morphology, ecology and bioerosion. Oecologia (Berl.) 19, 203-216 fossil relatives. In: The biology of the (1975) Porifera, pp 205-243. Ed. by W.G. Fry. London: Sponges in coral reefs. In: Coral reefs: re­ Academic Press 1970 search methods, pp 299-313. Ed. by D.R. Hinde, G.J.: On a new genus of siliceous sponges Stoddart and R.E. Johannes, Paris: UNESCO from the lower calcareous grit of Yorkshire. 1978 Q. Jl geol. Soc. Lond. 46, 54-61 (1890) Schindewolf, O.H.: Analyse eines Ammoniten­ Huang, T.: A new Radiolaria from the Somachi Gehauses. Abh. math.-naturw. Kl. Akad. Wiss. Formation, Kikai-jima, Kagoshima Prefecture, Mainz 8, 135-188 (1967) Japan. Trans. Proc. palaeont. Soc. Japan Schrammen, A.: Die Kieselspongien des oberen (N.S.) 68, 177-184 (1967) Jura von Suddeutschland. A. Vorwort und all­ Hurd, D.C.: Factors affecting solution rate of gemeiner Teil. Palaeontographica (Abt. A) 84, biogenic opal in seawater. Earth planet. Sci. 149-194 (1936) Letters 15, 411-417 (1972) Schwab, D.W. and R.E. Shore: Mechanism of inter­ Inoue, M. and Y. Iwasaki: A problematic micro­ nal stratification of spic­ organism similar to the sterraster of sponges. ules. Nature, Lond. 232, 501-502 (1971) Proc. Japan Acad. 51, 273-278 (1975) Sollas, W.J.: Report on the Tetractinellida col­ Jewell, M.E.: An ecological study of the fresh­ lected by H.M.S. Challenger, during the years water sponges of northern Wisconsin. Ecol. 1873-1876. Rep. scient. Results Voyage HMS Monogr. 5, 461-504 (1935) Challenger (Zool.) 25, i-clxvi, 1-458 (1888) J¢rgensen, C.B.: On the spicule-formation of Strickland, J.D.H. and T.R. Parsons: A practical Spongilla +acustris (L.) I. The dependence of handbook of seawater analysis. Bull. Fish. the spicule formation on the content of dis­ Res. Bd Can. 167, 1-311 (1968) solved and solid silicic acid of the milieu. Vosmaer, G.C.J. and J.H. Vernhout: The Porifera Biol. Meddr 19, 1-45 (1944) of the Siboga Expedition. I. The genus Placo­ Koltun, V.M.: Spicules of sponges as an element spongia. Siboga Exped. 6A, 1-17 (1902) of the bottom sediments of the Antarctic. In: Weaver, C.E. and K.C. Beck: Miocene of the S.E. Symposium on Antartic Oceanography, Santiago, United States: a model for chemical sedimenta­ Chile, 1966, pp 121-123. Cambridge: Scott tion in a peri-marine environment. Sedimenta­ Polar Research Institute 1968 ry Geol. 17,1-234 (1977) Land, L.S.: Early dissolution of sponge spicules from reef sediments, North Jamaica. J. sedim. Petrol. 46, 967-969 (1976) Macintyre, I.G.: Distribution of submarine ce­ Dr. Klaus Rutzler ments in a modern Caribbean fringing reef, Department of Invertebrate Galeta Point, Panama. J. sedim. Petrol. 47, Zoology, NHB 163 503-516 (1977) Smithsonian Institution Moore, C.H. and W.W. Shedd: Effective rates of Washington, D.C. 20560 sponge bioerosion as a function of carbonate USA

Date of final manuscript acceptance: June 23, 1978. Communicated by M.R. Tripp, Newark