marine carbonate cements, biofilms, biomineralization ... - CiteSeerX

3 downloads 87 Views 1MB Size Report
... 1976, Facies and fabric specificity of early subsea cements in shallow Belize (British Honduras) Reefs: Journal of. Sedimentary Petrology, v. 46, p. 523–544.
MARINE CARBONATE CEMENTS, BIOFILMS, BIOMINERALIZATION, AND SKELETOGENESIS: SOME BIVALVES DO IT ALL COLIN J.R. BRAITHWAITE1, JOHN D. TAYLOR 2, AND EMILY A. GLOVER 2 1

Division of Earth Sciences, Gregory Building, The University of Glasgow, Lilybank Gardens, Glasgow G12 8QQ, UK e-mail: [email protected] 2 Department of Zoology, The Natural History Museum, Cromwell Road, London SW7 5BD, UK

ABSTRACT: Recent years have seen a growing interest in the direct and indirect roles of organisms, and more explicitly of organic matter, in the crystallization of minerals in sediments. Two bivalve genera form extensive biofilms that are apparently responsible for the growth of a variety of isolated crystals together with a marine cement. Granicorium indutum and Samarangia quadrangularis are infaunal species that cover their shells with a cemented coating of sand, sculpted to mimic the surface ornament typical of many bivalves. Granicorium has an exceptionally mobile mantle margin that produces large volumes of mucus, harboring a diverse microbial community. The sand grains surrounding both species are initially bound by a biofilm that provides structural integrity but also acts, like others, as a template for the crystallization of a variety of carbonate polymorphs. These include varied prismatic crystals, rice-grain and wheatsheaf forms and, more importantly, large volumes of acicular crystals indistinguishable from typical marine cements. The distribution of these crystals is related to positions accessible to the mantle of the animal and in Samarangia appears to result from the active emplacement of grains and mucus several times a year. Summing the evidence, it seems that these species are responsible not only for the biomineralization involved in the formation of their shells but also for crystallization mediated through biofilms and for the generation of cements that are morphologically indistinguishable from typical marine cements. The boundaries between these three strategies may be closer than we think.

INTRODUCTION

Over the last decades a substantial body of literature has accumulated describing the morphology and locus of rapidly formed marine carbonate cements (Ginsburg 1957; Schroeder 1972; James et al. 1976; Ginsburg and James 1976; Dravis 1979; MacIntyre 1985; Harris et al. 1985; Friedman 1998). Almost in parallel has been a widening interest in the role of organisms and organic matter in driving the formation of cements in a variety of marine and freshwater environments (Golubic´ 1973; Simkiss 1986; Pedley 1992; Thompson and Ferris 1990). Dominated by microbes, the organisms implicated in this process include bacteria, cyanobacteria, and diatoms, as well as polychaetes, fungi, and higher plants (De´farge et al. 1996). Most examples imply the mediation of organisms in generating polysaccharide biofilms that in turn bring about crystallization. The issue of biomineralization of invertebrates, by which a large proportion, including Protozoa, Porifera, all three groups of Mollusca, Brachiopoda, Cnidaria, and Echinodermata, generate calcium carbonate skeletons, seems unrelated, yet here also there is a direct and causative relationship between the presence, composition, and function of organic compounds and the character of crystal growth (Weiner and Traub 1984). Thus, crystals may form abiogenically, organisms may mediate in extracellular crystallization, or mineral matter may be secreted, with protein, as a skeletal material. Here we describe two bivalve molluscs, Granicorium indutum (Hedley 1906) and Samarangia quadrangularis (Adams and Reeve 1850), together able to exploit the entire gamut of mineralizing strategies with, appropriately for bivalves, a foot in all three camps. They raise the important question for sedimentary rocks of what crystallization is truly organic or inorganic. JOURNAL OF SEDIMENTARY RESEARCH, VOL. 70, NO. 5, SEPTEMBER, 2000, P. 1129–1138 Copyright q 2000, SEPM (Society for Sedimentary Geology) 1073-130X/00/070-1129/$03.00

MATERIAL

Granicorium and Samarangia Granicorium indutum and Samarangia quadrangularis are presently classified in separate subfamilies of the Veneroidea (Keen 1969; FischerPiette and Vugadinovic 1977), the Tapetinae and the monospecific Samaranginae, although Harte (1998, supported by Taylor et al. 1999) has suggested that they should be placed together in the Samaranginae. Granicorium indutum is an infaunal bivalve with a simple ovoid shape (Fig. 1A) living a few millimeters below the sandy substrate surface in tropical and subtropical waters 30–150 m deep. The samples described here were collected from around Rottnest Island, off Freemantle, West Australia (Glover and Taylor, in press). The siphons are short and there is no pallial sinus and no fusion of the mantle margin except around the siphons. The true surface of the shell is ornamented only by narrowly spaced growth lines. In living material and untreated museum specimens, however, this surface is concealed beneath a tightly adherent sand coating that forms a series of low concentric ribs. This is not a secondary incrustation but is effectively a primary constructional feature of the shell. The grains forming the coating are derived from the surrounding sediment and include siliciclastic and a variety of bioclastic particles. The material binding them is of two kinds. Close to the shell margin a mucus film is secreted by the mollusc in contact with the periostracum and adherent sand grains. When dry, as in museum specimens, this is hard and brittle. Within a few millimeters of the margin, however, it is replaced by a crystalline cement. Samarangia quadrangularis is also infaunal but is rare in museum collections, and its biology is entirely unknown. We have been able to examine material from the Muse´e National d’Histoire Naturelle in Paris. Records suggest that Samarangia occurs in sandy substrates in tropical waters 3– 70 m deep. Like that of Granicorium the shell has a simple ovoid form and there is no pallial sinus, implying that the animal lives close to the sediment–water interface. The shell is again covered by cemented sand that forms a regular and reproducible surface ornament of up to ten prominent radial ribs and nodes (Fig. 1B). These resemble the ornament on many other bivalves and are clearly a primary constructional feature but, and significantly, they again have no counterpart on the smooth surface of the shell beneath. METHODS

After collection, fresh individuals of Granicorium were relaxed using a few drops of propylene phenoxetol in seawater and then preserved in 80% alcohol. Pieces of the mantle margin were sliced with a razor blade and taken through ascending concentrations of acetone before critical-point drying using liquid CO2. They were sputter-coated with gold–palladium alloy and examined by scanning electron microscopy (SEM). Dried museum specimens provided polished thin sections for polarized light, ultraviolet fluorescence, and cathodoluminescence examination, or were fractured and coated with gold–palladium for SEM examination. Polished sections were coated with carbon for examination using a CAMECA electron microprobe. Because crystals are small and difficult to separate, mineralogy was determined on the basis of crystal morphology, EDX, and trace-element mapping by microprobe.

1130

C.J.R. BRAITHWAITE ET AL.

FIG. 1.—External appearance of accreted (cemented) surfaces of A) Granicorium and B) Samarangia.

CEMENT PETROGRAPHY

The characteristics and distributions of the varied crystalline carbonate cements associated with Granicorium and Samarangia show some degree of overlap, but two broad groups can be differentiated. (1) A variety of crystals form on or within mucus biofilms at mantle margins. (2) Crystal coatings resembling marine cements are associated with these at the margins but are also more widely distributed, apparently without biofilms, within sand agglutinated on shell surfaces. (1) SEM examination of the shell margin of fresh Granicorium reveals the presence of sheets of mucus, apparently secreted by the mantle, that are associated with a diverse bacterial community (Fig. 2A). Small crystals of a variety of morphologies are laid down within and project from the surfaces of these sheets (Fig. 2B), or are apparently in contact with quartz and feldspar grains. In Samarangia similar thin mucus sheets, or biofilms, resembling those described and figured by Pedley (1994), Braithwaite and Zedef (1996), De´farge et al. (1996), and others, are also present at the margin of the shell. These are well preserved, although the shells examined were air-dried and had not been specially treated. No bacterial cells were identified in Samarangia and, because we are unable to differentiate mucus secreted by the bivalves from any that might have been generated microbially, we cannot here prove or disprove bacterial influence in mucus secretion. (2) The adherent sand grains in Granicorium are arranged in an imbricate manner with their long axes inclined and converging towards the shell surface, away from the margin (Fig. 3A). The mineral cement binding

grains consists of sheets of closely packed fibrous crystals with their long axes normal to the sheet surfaces. There is a marked asymmetry in the distribution of the cement, and growths are significantly thicker on grain surfaces facing towards the ventral margin, whereas those facing away carry little or no cement. In some areas the cement is multilaminar (Fig. 3B), but laminae taper laterally and some are discontinuous. Total thicknesses are typically 20–40 mm. In a few areas the cement surfaces are strongly convex, forming botryoids that grow towards each other from pore surfaces and may terminate abruptly, leaving part of the grain surface exposed. Electron microprobe element mapping indicates that this cement contains relatively large quantities of strontium but little or no magnesium, and crystals are therefore interpreted as aragonite. Sand grains are also imbricate against the shell surface in Samarangia and are associated with discontinuous fibrous, spherulitic, cement sheets inclined outwards from the surface and away from the aperture (Fig. 3C). Each cement sheet consists of packed radial bundles of elongate fibrous crystals (Fig. 3D) which, like those in Granicorium, are arranged with their long axes normal to the sheet surface. The growth directions of the crystals face inwards towards the pores enclosed by the sheets. Crystals also appear to have nucleated on the shell surface and on attached detrital grains, both of which may originally have had organic coatings. Cement thicknesses vary and, as in Granicorium, there is a marked asymmetry, with thicker growths facing towards the ventral margin of the shell. Along the shell surface, cement growths commonly form L-shaped accumulations in the angles between the shell and overlying grains, tapering outwards (away

FIG. 2.—SEM photomicrographs of A) bacteria on the surface of a mucus biofilm on the middle mantle fold of Granicorium, and B) mucus sheet in Samarangia with projecting carbonate crystals.

SOME BIVALVES DO IT ALL

1131

FIG. 3.—Cement sheets. Thin-section photomicrographs (crossed polars) of A) cementing of imbricate siliciclastic and bioclastic grains to the shell surface of Granicorium and B) multilaminar fibrous cement crust (arrows) on the surface of a bioclast cemented to Granicorium. Note in both of these the asymmetry of the cement distribution. SEM photomicrographs of the shell margin in Samarangia illustrating C) inclined cement sheets binding siliciclastic grains and D) the surface of a ‘‘marine’’ cement sheet.

from the shell) and forwards along the shell surface towards the ventral margin. In some areas groups of crystal fibers are crossed by contiguous growth lines, reflecting successive stages of accretion of the cement. However, growth lines within the sheets locally converge towards the shell surface and layers thin rapidly to their terminations. Because successive sheets are spaced at intervals along the shell surface and are systematically enclosed, they must be produced periodically. At a distance from the shell margin small amounts of epitaxial calcite cement are developed locally on some foraminiferal surfaces and small amounts of ‘‘typical’’ marine cement are present within some intraskeletal pores. Whereas in Granicorium the adherent sediment remains open to circulation, in Samarangia seawater is apparently trapped, and flow is reduced within the sequence of inclined cement sheets (Fig. 3C). Viewing crystals within these two groups as a whole, four morphological categories can be recognized: (1) Hexagonal Prisms.—Hexagonal prismatic crystals with smooth planar faces are common within the mucus of Granicorium. They are approximately 2 mm long with a length-to-width ratio of from 1:1 to more than 5:1 and are simple prisms with planar basal pinacoid (0001) terminations (Fig. 4A). They occur both as isolated crystals and as loose aggregates. Similar prismatic crystals in Samarangia, with length-to-width ratios close to 1:1, are 2–3 mm in diameter and are characterized by prominent screw dislocations on their terminal pinacoidal faces (Fig. 4B). (2) Tabular Discs.—Tabular hexagonal discs in Granicorium are 2–3 mm in diameter with a length-to-width ratio of about 1:3. In contrast to

the prisms, these bear minute crystal projections on their surfaces and are terminated by a basal pinacoid with a prominent screw dislocation (Fig. 4C). Some discs, apparently because growth has been dominated by additions to the exposed dislocation step that were sometimes incomplete, consist entirely of thin, presumably spiral, sheets (Fig. 4D). Discs may again occur as isolated crystals or loose aggregates. (3) Rice-Grain and Wheatsheaf Crystals.—In both Granicorium and Samarangia the mucus coating extending from the mantle margin locally bears or encloses roughly hexagonal crystals up to 20 mm long. Some of these taper to a pinnacled rice-grain shape (Fig. 5A), but others form slightly divergent sub-grains (Fig. 5B), approaching a wheatsheaf morphology (Fig. 5C). In Samarangia veils of mucus are locally studded with isolated hexagonal rice-grain crystals (Fig. 5A) about 20 mm long with a lengthto-width ratio of approximately 3:1. The terminations on these commonly carry subsidiary growth projections that become increasingly well defined, dividing to resemble the pinnacled surfaces described by Folk et al. (1985, fig. 9). In Samarangia, wheatsheaf-form crystal units are 5–25 mm long with a length-to-width ratio of approximately 3:1 and are randomly arranged on grain surfaces. They consist of laterally linked bundles of subcrystals (Fig. 5B), inequalities in length producing ragged ends and a characteristic divergence of sub-crystal axes (Fig. 5C). Groups of crystals radiating from a common nucleus form radial clusters as growth fills the gaps between divergent axes. More widely divergent crystals in Granicorium have a caltrop-like shape, with four or five axes radiating from a common nucleus (Fig. 5D).

1132

C.J.R. BRAITHWAITE ET AL.

FIG. 4.—SEM photomicrographs of A) hexagonal prismatic crystals in the fibrillar remains of a degraded biofilm of Granicorium; B) a screw dislocation step on 0001 face of a prismatic hexagonal crystal in Samarangia; C) tabular hexagonal prisms with a prominent terminal screw dislocation associated with ?cyanobacterial filaments in Granicorium (note rough surfaces); and D) tabular crystals of lamellar habit associated with bacteria and cyanobacterial filaments in Granicorium.

All of these crystal forms are intimately associated with mucus sheets and in some cases are partially embedded in them. All are found together with bacterial cells, and some overgrow ?cyanobacterial filaments (Figs. 4C, D) although most do not have this especially close relationship. By contrast, the fourth group shows a more tenuous association. (4) Fibrous or Acicular Crystals.—In both Granicorium and Samarangia the mucus sheets decrease in volume with increasing distance (several millimeters) from the aperture margin, and there is a concomitant increase in the volume of a carbonate cement apparently coating grain surfaces. The component crystals of this cement are morphologically distinct from those embedded within the mucus biofilms at the mantle margins and are characteristically fibrous or acicular, locally forming isopachous crusts or isolated radiating clusters and botryoids. Some crystals are roughly rectangular in section and occur in open radiating clusters with divergent axes (Fig. 6A, B). Locally, spherular bundles of closely aligned fibers form dense botryoidal masses in which successive growth stages are marked by concentric growth lines and terminations are generally concordant (Fig. 6C). Locally, acicular epitaxial calcite cement forms on some foraminiferal surfaces and small amounts of randomly oriented acicular crystals occur within a few intraskeletal pores (Fig. 6D). The distribution of this cement is important. The growth directions of crystals within the cement sheets imply growth relative to a substrate that initially supported them and provided structural integrity but that no longer exists in the specimens examined. The cement is thicker and more consistently developed on surfaces facing towards the ventral margins of the shells. Towards the umbones the volume of adherent sand decreases and

the sand coating becomes thinner. This indicates that although the cement initially increases in thickness away from the aperture within a short distance no further cement is added. Close to the shell margin in Granicorium thin mucus sheets alternate with layers of acicular crystals (Fig. 7A) and tenuous nets of mucus are draped across some crystal surfaces (Fig. 7B). Evidently, while crystals formed on mucus surfaces in both Granicorium and Samarangia other mucus sheets were subsequently added, laying the foundations for renewed crystal growth, but sometimes unrelated to the form and distribution of the crystals beneath (Fig. 7B, C). (5) Other Crystals.—A variety of additional crystals have been noted in both Granicorium and Samarangia. In Granicorium a few small bladelike crystals have been identified by EDX analysis as gypsum. In Samarangia others, of roughly cubic form, with surface perforations that may indicate dissolution, are probably halite (Fig. 8). Both may reflect the drying and evaporation of seawater after collection, but it is possible that gypsum is a primary feature (discussed below).

DISCUSSION

Granicorium and Samarangia both demonstrate a close relationship between a variety of crystal forms and the presence of biofilms, and both are characterized by large volumes of crystals that strongly resemble typical marine aragonite cements. What mechanisms can account for the formation of these crystals and for this diversity of form?

SOME BIVALVES DO IT ALL

1133

FIG. 5.—SEM photomicrographs of A) rice-grain crystals in Samarangia, B) parallel crystal bundles projecting from a biofilm surface in Samarangia, C) crystal terminations divided to give wheatsheaf form in Samarangia, and D) radiating crystal bundles in Samarangia.

The Marine Cement Hypothesis The acicular crystals forming distal to the mantle margins in Granicorum and Samarangia resemble marine cements, normally viewed as resulting from inorganic processes. Surface seawater is commonly regarded as supersaturated with respect to calcium carbonate, but although Broecker (1974) suggested that in the tropics it may reach seven times saturation, Morse and Mackenzie (1990) indicate that at present it is about six times supersaturated with respect to calcite and four times supersaturated with respect to aragonite. However, saturation is not the sole factor controlling precipitation, and changes in pCO2, and in the Mg:Ca ratio in seawater, coupled with the relatively low activity products of calcium and carbonates, cause significant decreases in these values and also determine that aragonite is commonly the first phase to form. The fact remains that in warm climates surface waters are saturated with respect to both calcite and aragonite. Thus, any question concerning the relationship between seawater chemistry and the growth of cements in sediments is not so much why sediments become cemented, but why they do not. The answer to this problem lies in the matter of rates. Rates of nucleation and of crystal growth of the commonly occurring carbonate polymorphs are sufficiently slow that if sedimentation rates are too high, or sediments too mobile, grains do not remain in contact with the water body for sufficient time for an organized cement to develop. Sea-floor cementation is therefore confined to stable areas where tidal flow or some other mechanism provides a pump to circulate waters but flow is not so large as to disturb the sediments involved. The sediment must be sufficiently coarse grained to allow effective flow rates to be achieved. From this premise it is reasonable to suppose that, in areas and conditions

where they are otherwise absent, Granicorum and Samarangia create an environment in which marine cements are able to form. How might they do this? Two factors are involved. (1) Granicorum and Samarangia differ from most bivalves in producing relatively large volumes of mucus at the mantle margin. It is this mucus that initially binds the grains to the shell surface, providing structural stability, and which also enables a succession of biologically mediated and physicochemical processes to be set in train. The chemical role of the mucus is considered below. (2) Both species provide an effective pump, respiratory movements forcing relatively large volumes of water from the aperture back through the sediment towards the surface. In both Granicorum and Samarangia, the volume of mucus decreases away from the aperture margin, and there is an initial reciprocal increase in the volume of acicular carbonate cement. This cement locally forms isopachous crusts and closely packed radiating clusters and botryoids, and is morphologically identical to marine cements figured from recent and Holocene carbonates (e.g., Lighty 1985). Crystals are distinct from most of those more intimately associated with the mucus biofilms. The distribution of the cement is significant. It is thicker and more consistently developed on surfaces facing towards the ventral margins of the shells. Towards the umbones the volume of adherent sand decreases. This implies that although the cement thickness increases away from the aperture it does not do so indefinitely, and a short distance from the aperture no further cement is added. One explanation for the growth of acicular cement on the shell surfaces

1134

C.J.R. BRAITHWAITE ET AL.

FIG. 6.—SEM photomicrographs illustrating A) radiating acicular crystal bundles in Samarangia, B) surface of cement showing poorly organized radiating crystals masses in Samarangia, C) the concordant surface of cement formed by parallel crystal growth in Samarangia, and D) random crystals in intraskeletal pore of foram attached to Samarangia.

of Granicorium and Samarangia is that both species live in areas with rapid water circulation where sea-floor cementation might be expected. The fact that sediments in this habitat are uncemented reflects, at least from observations on Granicorium, rapid tidal flow and a parallel sediment mobility that maintains a loose surface. Granicorium and Samarangia both provide an organic binding that cements grains together and holds them in place for several years. The anisotropy in distribution and the imbrication of agglutinated grains result from movements of the animals themselves. These include gross movements within the sediment, as a reflection of the infaunal habit, and also ventilation of the mantle cavity that drives water from the margin of the shell through the pedal gape. The flow generated by ventilation passes upwards through sand adhering to the shell surface and through adjacent loosely packed grains. From this perspective the distribution of the cement in both species is a reflection of the direction of the flow. Gonza´lez et al. (1992) suggest that the fibrous habit of crystals is a reflection of a rapid flow rate and continued renewal of reactant. Areas lacking cement are effectively stagnant. This reasoning leads to the conclusion that although these animals actively accrete sand grains at the shell margin, they do not themselves secrete the bulk of the cement. Instead they provide a milieu in which inorganic cementation progressively adds strength to the material accreted. In this scenario the buildup and subsequent decline of cement away from the margin is a reflection of downflow depletion of calcium and carbonate and progressive undersaturation. The thicker coatings on more recently formed parts of the shell reflect the agerelated increase in size of the animal and the corresponding increase in the

volume of water passing through the shell. The data from Samarangia suggest, however, that there is an additional factor. Biomineralization and Biofilms As indicated, it could be argued that the extensive fibrous cements formed within sediment on the surfaces of these two bivalves are entirely ‘‘inorganic’’. There is no evidence that organic matter is necessarily adsorbed on grain or shell surfaces to promote crystallization. There is, however, a clear relationship between crystal growth and biofilm development at the mantle margins of both species and also a parallel appearance of multiple mucus sheets and multilaminate cement. In Samarangia the growth of the inclined cement sheets indicates the former existence of a substrate that is no longer present. Although seawater may be saturated with respect to calcite and aragonite it shares the characteristics of other saturated solutions and does not precipitate spontaneously. Some input of energy, the activation energy, is required for precipitation to begin. The principal effect of this is widely recognized in sediments where most cements form by heterogeneous nucleation on existing surfaces. Nucleators such as the organic matrix in shells and the surfaces of biofilms lower the activation energy so that crystal formation can take place at much lower saturation levels than would be required for spontaneous precipitation (Stumm 1992). They are more effective if they share chemical or structural characteristics with the nucleating crystal. Shell Secretion.—Although the secretion of the shell in Granicorium is

SOME BIVALVES DO IT ALL

1135

FIG. 8.—SEM photomicrograph of crystals of varying form in Samarangia. including gypsum (g) and halite (h).

FIG. 7.—SEM photomicrographs illustrating A) the interlayering of mucus sheets and cement layers in Samarangia, B) a tenuous (?degraded) mucus net draped over crystal terminations in Samarangia, and C) a mucus sheet added after crystal growth in Samarangia.

not at issue, the mechanism by which it is achieved in this and other molluscs is relevant. The shell structure follows that in other Veneroidea, comprising an outer crossed-lamellar layer, a middle homogeneous layer, and, within the pallial line, an inner layer consisting of homogeneous structure with complex crossed-lamellar patches, conforming to the Group III type defined by Shimamoto (1986, fig. 9c). Among the mechanisms proposed for shell secretion in bivalves, Wilbur (1976) and Weiner and Traub (1984) suggested that the macromolecules of the organic matrix of the shell,

and particularly of the periostracum, provide a structural framework in which the ions of the mineral phase are ordered and nucleate (Weiner and Traub 1984). The nucleation sites have been suggested to be constructed of sheet-like acid-sulfated proteins and polysaccharides that involve amino acid sequences in which aspartic acid groups are separated by a single amino acid group (Weiner and Traub 1984). Aspartic acid, glycene, and serine are all present in the soluble fractions of shell proteins (Crenshaw 1972, 1990), and the high-affinity sites on these molecules responsible for binding calcium appear to be sulfate ester groups (Greenfield et al. 1984). The binding of calcium to protein residues allows both carbonate and calcium ions to be drawn by ionotropy to fixed positions in the structure (Thiele 1967; Crenshaw and Ristedt 1975). Mineral growth may then proceed by epitaxy (Posner and Betts 1981; Weiner et al. 1983; Weiner and Traub 1984) in which there is structural continuity between the calciumbinding sites of the protein and the calcium ions of the mineral, the lattice arrangement of the protein surface forming a template precisely matching that of the mineral phase. However, Crenshaw (personal communication) considers that in reality, calcium could not be bound in this way in seawater because of the large excess of sodium and magnesium present. Relationships superficially resembling those reflecting the adsorption of ions to specific sites may be generated by the microtopography of the substrate and the influence of this on the architecture of the mineral produced (Crenshaw 1990). Irrespective of these arguments, well-defined relationships between the organic matrix and the mineral component of shells have been observed in several bivalves, a gastropod, and Nautilus (Weiner et al. 1983). Sulfated glycoproteins apparently function as templates (Crenshaw and Ristedt 1975), and Crenshaw (1972) isolated a glycoprotein with a high content of aspartic acid and glycine from the shell of Mercenaria. It is assumed that there must be a transfer of calcium and an active transport of bicarbonate within the pallial fluid of bivalves (Wilbur 1976). The mineral may not, however, be nucleated as a result of any structural match between ionic groups of the substrate and those of the crystal lattice (Crenshaw 1990). The repeat spacing of potential binding sites in proteins is at least an order of magnitude greater than that of calcium in mineral carbonates (Crenshaw, personal communication). Whatever the precise mechanism, the organic matrix is implicated in the growth of the mineralized shell, and, for whatever reason, structure and composition are important. The uncertainties surrounding the crystallization mechanism are one reason why there has been keen interest in the mechanism of attachment of oysters (Harper 1997), which seems to be one stage removed from organic influence. Cranfield (1973) was able to show that glands in both the foot and the mantle of Ostrea edulis secrete a tanned mucopolysaccharide–ar-

1136

C.J.R. BRAITHWAITE ET AL.

omatic protein. Harper (1992) demonstrated that in Crassostrea gigas and Ostrea edulis this protein cement includes crystalline carbonate with a spherulitic texture similar to those of supposed inorganic marine cements and resembling the cements described here. Indeed, in Saccostrea they are described as ‘‘classic cavity-fill structures’’ and nucleate both on the underside of the shell and on the substrate (Harper 1997). Todd (1993) referred to bivalves, both living and in the geological record, that have been able to immure other organisms as a result of their cementing activity (crystallization) outside the direct influence of the shell or mantle. This indirect relationship between the organism and an associated cement is of particular relevance to Granicorium and Samarangia. Biofilms and Reactions.—The occurrence of extensive acicular and fibrous cements in Granicorium and Samarangia on surfaces with a specific relationship to the mantle margins, and the close correlation between the appearance of multiple mucus sheets and multilaminate cement, provides clear evidence of the influence of the animals on cement growth. In Samarangia the generation of the inclined cement sheets, and the orientation of crystals within them, requires the presence of a substrate that no longer exists. A more direct relationship between the development of biofilms and more diverse crystals is found on the mantle margins. Silve et al. (1992) drew attention to similarities between the physicochemical behavior of the organic matrices of bone and shell nacre in mineralization, and there are compositional, structural, and functional parallels between these skeletal materials and crystals grown on or in biofilms. While biofilms may have similar properties, however, they may not all function in the same way. Verrecchia et al. (1995) described the precipitation of carbonate within the polysaccharide sheaths of cyanobacterial cells. Carbon dioxide and bicarbonate ions diffuse to the cells of the trichome, where they are utilized in photosynthesis, with the result that hydroxyl groups are released into the sheath. This acts as a diffusion barrier and within it the hydroxyl reacts with bicarbonate diffusing from the exterior to form carbonate ions. These in turn react with calcium adsorbed by the polysaccharide of the sheath, causing calcium carbonate to be precipitated (Merz 1992). This process provides a mechanism for precipitation within the sheath, but essentially the same reactions may occur without a sheath, diffusion to the outer medium allowing crystallization outside the bacterial structures (Thompson and Ferris 1990). Similar reactions may take place on or within proteinaceous biofilms. However, while the molecules of microbial sheaths and biofilms may bind calcium (Pentecost and Riding 1986) in a manner similar to that suggested for the organic matrix in bivalve shell mineralization, there is no formal arrangement of the growing crystals that implies epitaxy. A final additional relationship is described by Verrecchia et al. (1995). In this, spherulites nucleate and grow within the mucilage sheath of cyanobacteria but are subsequently expelled into the medium. The importance of this is that it provides a source of crystalline carbonate in an environment that to casual observation appears to have no biological connection. Both Granicorium and Samarangia secrete large volumes of mucus, and in Granicorium at least this contains a rich and diverse microbial flora. It is not clear whether this flora contributes additional mucus, but it may be no coincidence that similar intimate associations of bacteria, cyanobacteria, and mucus with precipitated crystalline carbonates are found in biofilms in sediments (Burne and Moore 1987; De´farge et al. 1996), stromatolites (Braithwaite and Zedef 1996), tufas (Pedley 1994), and hot-spring deposits (Guo et al. 1996). The range of environments in which these occur include marine, brackish (Braithwaite et al. 1989), hypersaline (Bauld 1984; Friedman et al. 1973), freshwater (Pedley 1994; De´farge et al. 1996), and subaerial systems (Verrecchia et al. 1995). In Samarangia, mucus sheets are secreted at intervals by the bivalve and crystal growth is initiated on these surfaces. The distribution and morphology of the sheets suggests that they are actively emplaced by the mobile middle fold of the mantle. The sheets form essentially enclosed cells, trapping fluid (seawater) against the original shell surface and providing

relative isolation from the seawater body. These secretion events are not equally successful, and a number of cement sheets fail to grow significantly from the shell surface. Estimates of the average life of the shells examined indicate that several events per year are able to form a contiguous coating. As in Granicorium the asymmetry of cement distribution implies an anisotropy in the natant fluid, reasonably explained as a response to flow. If mucus and cement sheets do indeed enclose and to a degree isolate water against the shell surface, however, we might expect a progressive depletion in both calcium and carbonate in response to crystal growth and effectively stagnant conditions in which there would be no net flow and therefore no asymmetry in the cement forming. Thus it seems likely that cement growth is entirely dependent on the presence of the mucilage film. Only surfaces accessible to the mantle margin and coated by such films are able to grow cement. Growth seems to cease after some interval but may be reinitiated if the cement surface is reactivated with mucus, generating multilaminar cement crusts. CONTROLS ON CRYSTAL FORM

The variety of crystal morphologies in Granicorium and Samarangia requires some additional discussion. It has long been known that ‘‘foreign’’ ions in a solution may control the mechanism of growth of crystals and thus their morphology. This was convincingly demonstrated by Buckley (1951), who produced a variety of bizarre forms from a range of compounds by essentially ad hoc experimentation. The importance of foreign ions is that, although they may not enter the structure of the growing crystal, their adsorption to specific sites on the crystal surface, or to specific faces of the growing crystal, may profoundly modify crystal habit and lead to the development of sectoral or oscillatory zoning. In this context, Suess (1970) demonstrated the direct inhibiting effect of organic contaminants on carbonate crystal growth. The habits of crystals formed in the presence of a soluble organic matrix differ significantly from those formed from control solutions (Wheeler and Sikes 1984). Addadi and Weiner (1985) noted that nucleation related to the presence of proteins occurs only when the protein is adsorbed onto a rigid surface and there is a stereochemical relationship whereas proteins in solution have the effect of inhibiting growth by binding to crystal faces. Wheatsheaf forms are attributed to growth divergence, but these and ricegrain crystals are not well differentiated and both resemble end members of ‘‘split crystals’’ described by Grigor’ev (1965). Ferna´ndez-Dı´az et al. (1996), and Morse et al. (1997) investigated the role of magnesium in the growth of calcite and aragonite crystals. At high levels of supersaturation it may promote the formation of spherules, whereas at lower levels dumbbell or wheatsheaf forms appear. Given and Wilkinson (1985) illustrate crystals of both high-magnesium and low-magnesium calcite of rice-grain form that they ascribe to abiotic kinetically controlled growth. However, Buczynski and Chafetz (1991) illustrate crystal and aggregate forms of both calcite and aragonite, including spherular and wheatsheaf bundles, that they attribute to bacterially induced precipitation. While these seem to be conflicting explanations, it may be that the influence of the bacteria, which includes the release of organic compounds to the environment, generates kinetic effects identical to those in abiotic systems and which are not therefore specifically microbial. The divided terminations in rice-grain crystals resemble surfaces described by Folk et al. (1985). While these authors were unable to determine whether such terminations reflected either growth or dissolution, Jones and Pemberton (1987) referred to similar pinnacles as resulting from organically mediated dissolution. However, these, and the tabular discoidal crystals in Granicorium, show a surface roughening. Roughening effects may be a reflection of rapid growth where the free energy of reactive sites becomes less important. This is commonly a result of increasing temperature (the critical roughening temperature of Jackson 1958) but is equally

SOME BIVALVES DO IT ALL explicable in terms of growth at a high chemical potential such as in supersaturation (Sunagawa 1982). The final issue is that of the formation of gypsum. Although present in only small quantities, this appears anomalous, given the evaporite context in which it typically forms. Gypsum has been reported associated with microbial mats (Krumbein 1986) and as a skeletal material in Charophyceae (Raven et al. 1986), in the Gamophyta, and in some Cnidaria, but it is rare. Lowenstam and Weiner (1989), in a table compiled to illustrate the distribution of mineral phases in various organisms, do not indicate the presence of gypsum in any mollusc. Sulfated compounds are present in the shells of bivalves (Clarke and Wheeler 1922; Kaplan et al. 1963) and have been implicated in the initiation of mineralization (Crenshaw 1972; Wada and Furuhashi 1970). Although sulfur is present in shell amino acids and in some protein fractions, sulfate generated by bacterial activity in pore waters is a far more potent source, and there remains the possibility that both gypsum and halite result from the evaporation of seawater after collection. CONCLUSIONS

The infaunal bivalves Granicorium and Samarangia employ a range of strategies to bring about the crystallization of carbonates. These have important implications for the growth of cements associated with organic matter. The boundaries between organically mediated and inorganic crystal growth may not be sharply defined. Both Granicorium and Samarangia secrete copious mucus from the inner fold and outer surface of the middle fold of the mantle margin, forming a biofilm. In Granicorium the biofilm also contains abundant bacteria, which probably contribute to its bulk. Unusually mobile mantle folds manipulate sand grains within the environment, together with mucus, and pack them against the periostracum and existing cement crusts, fixing them in position. These manipulations take place several times a year. Bacteria may modify the chemical environment within the mucoid curtains, encouraging the initial nucleation of carbonate and (perhaps) sulfate. Crystals of aragonite and calcite nucleate and grow on or within the mucus, and the molecular structure and composition of this may provide both a chemical latch and a structural alignment for calcium and carbonate ions. At a distance from the mantle margin the nucleation and growth of the more extensive marine cements takes place on the surfaces of mucus sheets that coat grain surfaces facing the aperture and provide structural stability during crystal growth. Variations in the composition of the biofilm and in the soluble organic molecules circulating in the fluids surrounding the two bivalves are probably responsible for the diversity of crystals formed. ACKNOWLEDGMENTS

We are grateful to Philippe Bouchet, of the Muse´e National d’Histoire Naturelle in Paris, who kindly lent specimens of Samarangia and allowed us to section one. We would also like to thank Chris Jones for making the geological thin sections and for help and advice, with Alex Ball, in the NHM electron microscopy unit. Harry Taylor assisted with macrophotography. The manuscript has been significantly improved by helpful comments from SEPM referees R. Pamela Reid and Miles Crenshaw, whose help is greatly appreciated. REFERENCES

ADAMS, A., AND REEVE, L., 1850, The Zoology of the Voyage of H.M.S. Samarang. Mollusca: London, Reeve, Benham and Reeve, p. 1–87. ADDADI, L., AND WEINER, S., 1985, Interactions between acidic proteins and crystals: stereochemical requirements in biomineralization: U.S. National Academy of Sciences, Proceedings, v. 82, p. 4110–4114. BAULD, J., 1984, Microbial mats in marginal marine environments: Spencer Gulf, South Australia, and Shark Bay, Western Australia, in Cohen, Y., Castenholz, R.W., and Halverson, H.O., eds., Microbial Mats: Stromatolites: New York, Alan R. Liss, p. 39–58. BRAITHWAITE, C.J.R., CASANOVA, J. FREVERT, T., AND WHITTON, B., 1989, Recent stromatolites in

1137

landlocked pools on Aldabra, western Indian Ocean: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 69, p. 145–165. BRAITHWAITE, C.J.R., AND ZEDEF, V., 1996, Hydromagnesite stromatolites and sediments in an alkaline lake, Salda Go¨lu¨, Turkey: Journal of Sedimentary Research, v. 66, p. 991–1002. BROECKER, W.S., 1974, Chemical Oceanography: New York, Harcourt Brace Jovanovich, 214 p. BUCKLEY, H.E., 1951, Crystal Growth: New York, Wiley, 571 p. BUCZYNSKI, C., AND CHAFETZ, H.S., 1991, Habit of bacterially induced precipitates of calcium carbonate and the influence of medium viscosity on mineralogy: Journal of Sedimentary Petrology, v. 61, p. 226–233. BURNE, R.V., AND MOORE, L.S., 1987, Microbialites: Organosedimentary deposits of benthic microbial communities: Palaios, v. 2, p. 241–254. CLARKE, F.W., AND WHEELER, W.C., 1922, The inorganic constituents of marine invertebrates: U.S. Geological Survey, Professional Paper 124, 62 p. CRANFIELD, H.J., 1973, Observations on the function of the glands of the foot of the pediveliger of Ostrea edulis during settlement: Marine Biology, v. 22, p. 211–223. CRENSHAW, M.A., 1972, The soluble matrix from Mercenaria mercenaria shell: Biomineralization, v. 6, p. 6–11. CRENSHAW, M.A., 1990, Biomineralization mechanisms, in Carter, J.G., ed., Skeletal Biomineralization: Patterns, Processes and Evolutionary Trends, vol. 1, New York, Van Nostrand Reinhold, p. 1–9. CRENSHAW, M.A., AND RISTEDT, H., 1975, The histochemical localization of reactive groups in the septal nacre from Nautilus pompilus, in Watabe, N., and Wilbur, K.M., eds., The Mechanisms of Mineralization in the Invertebrates and Plants: University of South Carolina Press, p. 335–367. DE´FARGE, C., TRICHET, J., JAUNET, A.-M., ROBERT, M., TRIBBLE, J., AND SANSONE, F.J., 1996, Texture of microbial sediments revealed by cryo-scanning electron microscopy: Journal of Sedimentary Research, v. 66, p. 935–947. DRAVIS, J., 1979, Rapid and widespread generation of recent oolitic hardgrounds on a high energy Bahamian Platform, Eleuthera Bank, Bahamas: Journal of Sedimentary Petrology, v. 49, p. 195–208. FERNA´NDEZ-D´IAZ, L., PUTNIS, A., PRIETO, M., AND PUTNIS, C.V., 1996, The role of magnesium in the crystallization of calcite and aragonite in a porous medium: Journal of Sedimentary Research, v. 66, p. 482–491. FISCHER-PIETTE, E., AND VUGADINOVIC, D., 1977, Suite des re´visions des Veneridae (Moll. Lamellibr.) Chioninae, Samaranginae et comple´ment aux Ve´nus: Muse´um National d’Histoire Naturelle, Me´moires, Se´rie A, Zoologie, v. 106, p. 1–186. FOLK, R.L., CHAFETZ, H.S., AND TIEZZI, P.A., 1985, Bizarre forms of depositional and diagenetic calcite in hot-spring travertines, central Italy, in Schneiderman, N., and Harris, P.M., eds., Carbonate Cements: SEPM, Special Publication 36, p. 349–369. FRIEDMAN, G.M., 1998, Rapidity of marine carbonate cementation—implications for carbonate diagenesis and sequence stratigraphy: perspective: Sedimentary Geology v. 119, p. 1–4. FRIEDMAN, G.M., AMIEL, A.J., BRAUN, M., AND MILLER, D.S., 1973, Generation of carbonate particles and laminates in algal mats—example from sea-marginal hypersaline pool, Gulf of Aqaba, Red Sea, Egypt: American Association of Petroleum Geologists, Bulletin, v. 57, p. 541–557. GINSBURG, R.N., 1957, Early diagenesis and lithification of shallow water carbonate sediments in South Florida, in LeBlanc, R.J., and Breeding J.G., eds., Regional Aspects of Carbonate Deposition: SEPM, Special Publication 5, p. 80–100. GINSBURG, R.N., AND JAMES, N.P., 1976, Submarine botryoidal aragonite in Holocene reef limestones, Belize: Geology, v. 4, p. 431–436. GIVEN, K.R., AND WILKINSON, B.H., 1985, Kinetic control of morphology, composition, and mineralogy of abiotic sedimentary carbonates: Journal of Sedimentary Petrology, v. 55, p. 109–119. GLOVER, E.A., AND TAYLOR, J.D., 1999, Diversity and distribution of subtidal macromolluscs around Rottnest Island, Western Australia, in Walker, D.I., and Wells, F.E., eds., Seagrass Communities, Marine Flora and Fauna of Rottnest Island, Western Australia: Western Australian Museum, Perth, p. 101–119. GOLUBIC´, S., 1973, The relationship between blue–green algae and carbonate deposits, in Carr, N.G., and Whitton, B.A., eds., The Biology of Blue–Green Algae: Oxford, U.K., Blackwell, Botanical Monographs, v. 9, p. 343–472. GONZA´LEZ, L.A., CARPENTER, S.J., AND LOHMANN, K.C., 1992, Inorganic calcite morphology: roles of fluid chemistry and fluid flow: Journal of Sedimentary Petrology, v. 62, p. 382– 399. GREENFIELD, E.M., WILSON, D.C., AND CRENSHAW, M.A., 1984, Ionotropy nucleation of calcium carbonate by molluscan matrix: American Zoologist, v. 24, p. 925–932. GRIGOR’EV, D.P., 1965, Ontogeny of minerals: Translation from the Russian, Ontogeniya Mineralov Izdatel’stro, L’vovskii Universitet (1961), Brenner, Y., ed., Israel Program for Scientific Translation, Jerusalem, 242 p. GUO, L., ANDREWS, J., RIDING, R., DENNIS, P., AND DRESSER, Q., 1996, Possible microbial effects on stable carbon isotopes in hot-spring travertines: Journal of Sedimentary Research, v. A66, p. 468–473. HARPER, E.M., 1992, Post-larval cementation in the Ostreidae and its implications for other cementing bivalvia: Journal of Molluscan Studies, v. 58, p. 37–47. HARPER, E.M., 1997, Attachment of mature oysters (Saccostrea cucullata) to natural substrata: Marine Biology, v. 127, p. 449–453. HARRIS, P.M., KENDALL, C.G.ST.C, AND LERCHE, I., 1985, Carbonate cementation—a brief review, in Schneiderman, N., and Harris, P.M., eds., Carbonate Cements: SEPM, Special Publication 36, p. 79–95. HARTE, E.M., 1998, Superfamily Veneroidea, in Beesley, P.L., Ross, G.J.B., and Wells, A., eds., Mollusca: the Southern Synthesis: Fauna of Australia, v. 5, Part B, p. 355–361, Melbourne, CSIRO Publishing.

1138

C.J.R. BRAITHWAITE ET AL.

HEDLEY, C., 1906, The Mollusca of Mast Head Reef, Capricorn Group, Queensland: Linnean Society of New South Wales, Proceedings, v. 31, p. 453–479. JACKSON, K.A., 1958, Mechanisms of growth, in Liquids, Metals, and Solidification: Cleveland, Ohio, American Society of Metals, p. 174–186. JAMES, N.P., GINSBURG, R.N., MARSZALEK, D.S., AND CHOQUETTE, P.W., 1976, Facies and fabric specificity of early subsea cements in shallow Belize (British Honduras) Reefs: Journal of Sedimentary Petrology, v. 46, p. 523–544. JONES, B., AND PEMBERTON, S.G., 1987, Experimental formation of spikey calcite through organically mediated dissolution: Journal of Sedimentary Petrology, v. 57, p. 687–694. KAPLAN, I.R., EMERY, K.O., AND RITTENBERG, S.C., 1963, The distribution and isotopic abundance of sulfur in recent marine sediments of southern California: Geochimica et Cosmochimica Acta, v. 27, p. 297–332. KEEN, M., 1969, Superfamily Veneracea, in Moore, R.C., ed., Treatise of Invertebrate Paleontology, Part N (2), Mollusca 6, Bivalvia, p. N670–N690: The University of Kansas and the Geological Society of America, Boulder, Colorado. KRUMBEIN, W.E., 1986, Biotransfer of minerals by microbes and microbial mats, in Leadbeater, B.S.C., and Riding, R., eds., Biomineralization in Lower Plants and Animals: Systematics Association, Special Volume 30, p. 55–72. LIGHTY, R.G., 1985, Preservation of internal reef porosity and diagenetic sealing of submerged early Holocene Barrier Reef, southeast Florida Shelf, in Schneiderman, N., and Harris, P.M., eds., Carbonate Cements: SEPM, Special Publication 36, p. 123–151. LOWENSTAM, H.A., AND WEINER, S., 1989, On Biomineralization: Oxford University Press, 324 p. MACINTYRE, I.G., 1985, Submarine cements—the peloidal question, in Schneiderman, N., and Harris P.M., eds., Carbonate Cements: SEPM, Special Publication 36, p. 109–116. MERZ, M.U.E., 1992, The biology of carbonate precipitation by cyanobacteria: Facies, v. 26, p. 81–102. MILLIMAN, J.D., 1974, Marine Carbonates, in Milliman, J.D., Mu¨ller, G., and Fo¨rstner, U., eds., Recent Sedimentary Carbonates, Part 1: Berlin, Springer-Verlag, 375 p. MORSE, J.W., AND MACKENZIE, F.T., 1990, Geochemistry of Sedimentary Carbonates: Amsterdam, Elsevier, 707 p. MORSE, J.W., WANG, Q., AND TSIO, M.Y., 1997, Influences of temperature and Mg:Ca ratio on CaCO3 precipitated from seawater: Geology, v. 25, p. 85–87. PEDLEY, H.M., 1992, Freshwater (phytoherm) reefs: the role of the biological film and its bearing on marine reef cementation: Sedimentary Geology, v. 79, p. 255–274. PEDLEY, H.M., 1994, Prokaryote–microphyte biofilms and tufas: A sedimentological perspective: Kupaia, Darmsta¨dter Beitra¨ge zur Naturgeschichte, v. 4, p. 45–60. PENTECOST, A., AND RIDING, R., 1986, Calcification in cyanobacteria, in Leadbeater, B.S.C., and Riding, R., eds., Biomineralization in Lower Plants and Animals: Systematics Association, Special Volume no. 30, p. 73–90. POSNER, A.S., AND BETTS, F., 1981, Molecular control of tissue mineralization, in Veis, A., ed., Chemistry and Biology of Mineralized Connective tissues: Amsterdam, Elsevier, p. 257– 266. RAVEN, J.A., SMITH, F.A., AND WALKER, N.A., 1986, Biomineralization in the Charophyceae, in Leadbeater, B.S.C., and Riding, R., eds., Biomineralization in Lower Plants and Animals: Systematics Association, Special Volume no. 30, p. 125–139.

SCHROEDER, J.H., 1972, Fabrics and sequences of submarine carbonate cements in Holocene Bermuda cup reefs: Geologische Rundschau, v. 61, p. 708–730. SHIMAMOTO, M., 1986, Shell microstructure of the Veneridae (Bivalvia) and its phylogenetic implications: Sendai, Japan, Tohoku University, Science Reports, Second Series (Geology), v. 56, p. 1–39. SILVE, C., LOPEZ, E., VIDAL, B., SMITH, D.C., CAMPRESSE, S., AND CAMPRESSE G., 1992, Nacre initiates biomineralization by human osteoblasts maintained in vitro: Calcified Tissue International, v. 51, p. 363–369. SIMKISS, K., 1986, The processes of biomineralization in lower plants and animals—an overview, in Leadbeater, B.S.C., and Riding R., eds., Biomineralization in Lower Plants and Animals: The Systematics Association, Special Volume no. 30, p. 19–37. STUMM, W., 1992, Chemistry of the Solid–Water Interface:, New York, Wiley, p. 448. SUESS, E., 1970, Interaction of organic compounds with calcium carbonate. I. Association phenomena and geochemical implications: Geochimica et Cosmochimica Acta, v. 34, p. 157–168. SUNAGAWA, I., 1982, Morphology of crystals in relation to growth conditions, in Rodrigues, C.R., and Sunagawa, I., eds., Crystal Growth Conditions in Sedimentary Environments: Estudios Geolo´gicos, v. 38, p. 127–134. TAYLOR, J.D., GLOVER, E.A., AND BRAITHWAITE, C.J.R., 1999, Bivalves with concrete overcoats: Granicorium and Samarangia: Acta Zoologica [Stockholm], v. 80, p. 285–300. THIELE, H., 1967, Geordnete Kristallisation. Nucleation und Mineralisation: Journal of Biomedical Materials Research, v. 1, p. 213. THOMPSON, J.B., AND FERRIS, F.G., 1990, Cyanobacterial precipitation of gypsum, calcite, and magnesite from natural alkaline lake water: Geology, v. 18, p. 995–998. TODD, J.A., 1993, The bivalve shell as a preservation trap, as illustrated by the Late Jurassic gryphaeid, Deltoideum delta (Smith): Scripta Geologica, [Leiden], Special Volume no. 2, p. 417–433. VERRECCHIA, E.P., FREYTET, P., VERECCHIA, K.E., AND DUMONT, J.-L., 1995, Spherulites in calcrete laminar crusts: biogenic CaCO3 precipitation as a major contributor to crust formation: Journal of Sedimentary Research, v. A65, p. 690–700. WADA, K., AND FURUHASHI, T., 1970, Studies on the mineralization of the calcified tissue of molluscs XVII. Acid polysaccharide in the shell of Hydriopsis schlengeli: Japanese Society for Science and Fisheries, Bulletin, v. 36, p. 1122–1126. WEINER, S., AND TRAUB, W., 1984, Macromolecules in mollusc shells and their functions in biomineralization. Royal Society [London], Philosophical Transactions, v. B304, p. 425– 434. WEINER, S., TALMON, Y., AND TRAUB, W., 1983, Electron diffraction of mollusk shell organic matrices and their relationship to the mineral phase: International Journal of Biological Macromolecules. v. 25, p. 325–328. WHEELER, A.P., AND SIKES, C.S., 1984, Regulation of carbonate calcification by organic matrix: American Zoologist, v. 24, p. 933–944. WILBUR K.M., 1976, Recent studies of invertebrate mineralization, in Watabe, N., and Wilbur, K.M., eds., The Mechanisms of Mineralization in the Invertebrates and Plants: University of South Carolina Press, p. 79–108. Received 16 December 1998; accepted 16 September 1999.