environmental and biological controls on ... - Wiley Online Library

499

ENVIRONMENTAL AND BIOLOGICAL CONTROLS ON BIVALVE SHELL MINERALOGY BY WILLIAM JAMES KENNEDY Dept. of Geology and Mineralogy, Oxford JOHN DAVID TAYLOR Dept. of Zoology, British Museum (Natural History), London ANTHONY HALL, Dept. of Geology, King’s College, London AND

(Received 10 February 1969) CONTENTS

. . I. Introduction . 11. Shell formation in bivalves . 111. Shell structure types (I) Nacreous structure .

.

. . . . . . . .

. . .

.

.

.

structure . (3) Prismatic structure . . (a)Simple prisms . . (6) Composite prisms . (4)Crossed-lamellar structure (5) Complex crossed-lamellar structure (6) Homogeneous structure . (7)Myostracal layers . . . (8) Layer relations . . . . (9)Other calcified elements . . IV. The distribution of aragonite, calcite and the shell structure types in recent bivalves . . (2) Foliated

V. Controls on shell mineralogy (I) Introduction . (2)Biological controls . (a) The mantle cells (6) The extrapallial fluid (c) The periostracum . (d)T h e organic matrix (e) Discussion . (3) Environmental controls (a) Temperature . ( 6 ) Salinity . . VI. Summary . . . VII. References . .

.

.

. .

. . . .

. . . . . . .

.

.

. .

.

523

.

.

526

*

527

. . . . .

.

516 516 516 516 517 517 517 519 519 519

510

I. INTRODUCTION

The bivalve shell is a complex organic/inorganic system consisting of two valves joined by a largely organic ligament. Each valve consists of an outer organic layer, the periostracum, and within this, the calcified shell. The calcified part of the shell consists of aragonite alone, or aragonite and calcite together. The ligament, which also has an outer periostracum coat, always contains aragonite when mineralized. All of these structures are secreted by the mantle of the animal, a specialized enveloping outer fold of the body wall. The occurrence of both aragonite and calcite in bivalve shells has been known since the days of Sorby (1879) and has subsequently been investigated by many workers, in particular Bsggild (1930). In recent years, interest has been stimulated by workers in two fields: biologists working on mechanisms of calcification (as reviewed by Wilbur, 32-2

WILLIAM JAMES KENNEDY AND

500

OTHERS

1964), and geologists studying shell mineralogy (Lowenstam, 1954a, b, 1963, 1964; Kennedy & Taylor, 1968) and diagenetic fabrics in molluscan carbonate rocks (Bathurst, 1966; Dodd, 1966b; Hall & Kennedy, 1967; Kennedy & Hall, 1967, etc.). For a few years we have been studying the structure of the bivalve shell, with a view to understanding the mechanism of calcification. We have also tried to use shell structure and mineralogy as a key to bivalve phylogeny, and to the relationship between molluscan classes. Both of these interests have led to consideration of the origin of calcified exoskeletons in general. We discuss here the mode of formation of the bivalve shell, the various modes of aggregation of calcium carbonate in bivalve shells, and the controls, both biological and environmental, of this distribution.

riostracum

pallial myost racu m

-

inner nacreous layer

mucous cells

Fig.

I.

Diagram of the relationships in a radial section of the mantle and shell edge of Anodonta cygnea. Based on Beedham ( 1 9 5 8 ~ )and our own observations.

rI. SHELL FORMATION IN BIVALVES

In bivalves the cells responsible for the secretion of the shell are not in direct contact with the shell, except at sites of muscle attachment and in the periostracal groove (Fig. I ) . The process of calcification is fundamentally different from that in some other phyla, e.g. echinoderms and brachiopods. Both the organic and inorganic constituents of the shell are secreted into a thin fluid layer, the extrapallial fluid, which lies in the extrapallial cavity, between mantle and shell (Fig. I). From this fluid the wholly organic periostracum is deposited and then, upon this, the organic matrix and calcium carbonate of the shell. All the constituents of the shell are thus secreted into the extrapallial fluid, whence they finally precipitate and polymerize. The initial event in shell deposition is the formation of the periostracum, which is produced at the margin of the bivalve mantle. The mantle edge is divided into three

Bivalve shell mineralogy

50'

folds, an inner muscular fold, a middle sensory fold, and an outer secretory fold. The epithelial cells on the inner side of the outer mantle fold secrete the periostracum. The first calcareous material is laid down by the epithelial cells of the outer part and outer mantle fold. The remainder of the shell is laid down by the cells of the general outer surface of the mantle. It is well known that the cells in the various regions of the outer mantle fold differ in shape, structure and histochemical properties (Trueman, 1949; Yonge, 1953 ; Beedham, 1958a, b ; Dunachie, 1963; Kawaguti & Ikemoto, 1962; Bevelander & Nakahara, 1968). The cells secreting the periostracum are tall and slender; those of the general outer surface of the mantle are broad and short. Little is known of bivalve growth, but it would appear that there is a generative zone in the inner part of the periostracal groove. As the animal grows, the cells change in shape, histochemistry and function. A single cell may thus secrete periostracum, outer shell layer and, perhaps, middle shell layer, become part of the pallial attachment area, secrete inner shell layer, become part of the adductor muscle attachment and then return to secreting inner shell layer. The periostracum is a thin, always entirely organic layer of quinone-tanned protein (Degens, Spencer & Parker, 1967)and has a complex, layered structure which is poorly understood. The periostracum plays a vital role in the development of the calcified shell, in that it acts as the substrate for mineralization and crystal growth. It also has a protective function, especially in preventing shell corrosion by acidic waters, and is thus thicker in species and genera which are liable to be exposed to such conditions, e.g. the fresh water Unionidae, and forms which live in potentially acidic interstitial waters, e.g. the Nuculacea and Nuculanacea. Thick periostraca also characterize bivalves secreting acid as an aid to rock boring, i.e. the mytilid Lithophaga. The remainder of the bivalve shell is deposited as an intimate mixture of protein (organic matrix) and calcium carbonate, either aragonite or calcite. The modes of aggregation of the carbonate and protein can be placed in seven main categories, six of which can be regarded as shell structure types. The seventh is a specialized deposit laid down beneath areas of muscle attachment, and these myostracal layers are invariably aragonitic. The shell structure types are constant in their mineralogy, and in their distribution amongst the extant superfamilies of bivalves. With the exception of myostraca, such shell structures occur as distinct, monomineralic layers within the bivalve shell. 111. SHELL STRUCTURE TYPES

The seven shell structure types and the other calcified elements present in bivalves are briefly described and illustrated below. A more detailed description and discussion of the organic matrix/carbonate relationships, crystallography and development of these shell-structure types is to be found elsewhere (Taylor, Kennedy & Hall, 1969). These descriptions are based on optical and electron microscope investigations of more than 600 bivalve species.

WILLIAM JAMES KENNEDY AND OTHERS

502

Nacreous structure This familiar shell structure type is present in the Nuculacea, Mytilacea, Pinnacea, Pteriacea, Unionacea, Trigonacea, Pholadomyacea and Pandoracea. Nacre is invariably aragonitic. Nacreous structure is built up of minute tablets ranging from 2 to 10p in diameter, and from 0.4 to 3 p in thickness. These tablets may be arranged in sheets more or less parallel to the shell interior (sheet nacre) or in lenses normal to the shell interior (lenticular nacre). Individual tablets may be euhedral or rounded. When euhedral, faces of the forms {OIO}and { I IO}are prominent, the largest, flat faces are of the form (001). Each tablet is surrounded by, and contains inclusions of, proteinaceous organic matrix. (I)

interlamellar and intracrystal Iine organic matrix

Fig.

2.

Block diagram of sheet nacre, showing carbonate and organic matrix relationships. (Based on Grbgoire (1967) and Watabe (1965).)

These features are summarized in Fig. 2. Details are shown in P1. I , Figs. I , 4-7. The distribution of the two nacre types appears to be related to shell geometry. Lenticular nacre appears in the middle shell layer of bivalves with a high spiral angle and distinctly convex shell. Sheet nacre forms the inner layer of such shells, or the whole of the nacreous part of flat shells with a low spiral angle (i.e. pteriaceans and mytilids) . In flat shells, growth is essentially parallel to the shell interior, and this is equally true for the inner layer of most nacreous shells. This is the growth form of sheet nacre. In the middle layer of strongly convex shells, accretion is mainly normal to, or at an angle to, the inner surface and is the mode of growth seen in lenticular nacre. FoZiated structure This, the ‘ calcitostracum’ or ‘ subnacreous’ layer of many authors, is invariably calcitic. It is present in the Ostreacea, Pectinacea, Anomiacea and Limacea and also occurs in some gastropods (MacClintock, I 967). Foliated structure is built of minute laths 1-20 p or more long, 2-4 p wide and 0.2-0.5 p thick. The crystallographic faces developed on these laths are poorly known; (2)


503

the large flat faces are of the form (0001).Angles between other faces are so variable that, although Tsujii, Sharp & Wilbur (1958) have suggested that the {IOTO)and {I 150) forms are developed, other forms could well be present. Laths are joined in side-to-side contact to form sheets that dip at a low angle to the shell interior, and, as a result, the surface pattern resembles a tiled roof. These features are summarized in Fig. 3, and details are shown in P1. 2, figs. 9, 10;P1. 3, figs. 3, 5. At optical level, foliated structure shows well-developed pseudopleochroism. inner shell surface

/ \

,lath

terminations

/ laths

Fig. 3. ( a ) Block diagram of foliated structure (based on electron micrographs). (b) Folia outcropping on the inner surface of Anomiu ephippiunt (from an electron micrograph),showing growth halts.

Both optical and electron microscope investigations have shown that the foliated layer of the Bivalvia develops as a three-dimensional dendrite system (Gorodetsky & Saratovkin, 1958). Watabe & Wilbur (1961) have described the early development of the dendrites.


504

( 3 ) Prismatic structure This is a familiar shell structure type, described by several previous workers: Carpenter (1844,1848), Schmidt (1923,1924,1925),Beggild (1930),etc. Two main types, simple and composite prismatic structure, are recognized here. These usually form outer shell layers. The simple prisms may be aragonitic or calcitic, while the composite prisms are invariably aragonitic. ( a ) Simple prisms. These are rather similar, irrespective of mineralogy. The most consistent difference is that calcitic prisms have transverse striations (Pl. 2, fig. 3 ;

interprismatic walls

diverg;ng crystallites

accretion lines

first-order prisms

-ventral

margin

shell interior

I (c)

internal flange

Fig. 4. ( a ) Block diagram of simple calcite prisms. ( b ) Block diagram of aragonite simple prisms. (c) Block diagram of composite prisms (based on Nicculu).


505

Fig. 4), whilst aragonitic simple prisms show diverging longitudinal striations and have less conspicuous transverse ones (PI. 2, fig. I ; Fig. 4). Simple calcitic prisms are found in the Pteriacea, Mytilacea, Pinnacea and Ostreacea. What seem to be modified simple calcitic prisms are present in the anomalous heterodont Chama pellucida Broderip. Simple aragonitic prisms occur in the Solemyacea, Unionacea, Trigonacea, Pandoracea, Pholadomyacea and Poromyacea. On shell interiors, simple prisms show a characteristic pattern, with a thick conchiolin wall separating the carbonate units. This conchiolin wall varies from 0-5p to 8 p in thickness in most species, although it is very thin in mytilids and oysters. The prisms have a polygonal outline, with relatively straight walls, meeting at triple points of 120' (PI. 2,fig. 3 ; Fig. 4). Prism diameter is variable, for example from 9 y to 80 y in some oysters (Tsujii et al. 1958), whilst prisms in Pinna are visible to the naked eye, and may be several millimetres long. In vertical sections, prisms appear as columns, usually expanding in diameter away from the periostracum, and separated by conchiolin walls (PI. 2, fig. I). This gross structure seen at optical level is confirmed and expanded by electron microscopy, which has shown that calcite prisms are built up of stacks of disk-shaped carbonate layers separated by thin sheets of organic matrix (Watabe & Wada, 1956; GrCgoire, 1961a, b). Even smaller packages of carbonate, that are each enclosed in protein, are recognizable within these disks. Aragonitic prisms are built of divergent longitudinal blocks 2-3 p wide, separated by organic sheets 0.3 p thick. Growth of simple prisms begins with the appearance of small spherulites on the periostracum at the shell margin. These show characteristic extinction crosses in polarized light. Traced away from the periostracum margin, the growing spherulites push their organic membranes aside, until they lock as a series of carbonate blocks separated by passively formed interprismatic walls (PI. 2, fig. 5 ) . Beyond this point growth can only proceed inwards, i.e. away from the periostracum, and thus a prism forms. Not all of the spherulites present at the periostracal surface grow into prisms, for many disappear as a result of geometric selection. (b) Compositeprisms. This structure is well-developed in the Nuculacea, Lucinacea, and most Tellinacea. It is always aragonitic and always forms the outer shell layer. The structure consists of large, first order prisms (sensu Berggild), which lie horizontally, with their long axes subparallel to the periostracum and radiating from the umbo. Each of these prisms is separated from its neighbours by a thin, interprismatic organic matrix wall, although this is far less conspicuous than its counterpart in simple prismatic structure. At its origin, each prism is small but grows in size as the shell grows. Branching is not usually seen. In some bivalve groups, the inner surface of each first order prism projects as a flange, representing the trace of a marginal denticle, as in the Nuculacea (Pl. 2, fig. 7;Fig. 4). Each first order prism is built up of smaller, elongate, needle-like second order prisms (sensu Berggild), each of which radiates outwards from the axis of the first order prism and is surrounded by an organic matrix sheath. Growth lines within each first order prism are strongly convex ventrally. The development of this shell structure type is clearly comparable to the spherulitic

506


growth pattern seen in simple prisms. The prisms, instead of growing normal to the periostracum, grow parallel to it. The early stages of their development are hidden by early shell growth. (4) Crossed-lamellar structure This shell structure type is well-developed in many heterodonts, the Arcacea, Limopsacea and some Pectinacea and Limacea. It also occurs in some gastropods, scaphopods and chitons, and has been discussed by several authors, e.g. Boggild (1930) first-order larnels laths making up second-order lamels

\>

Fig. 5 . (a) Block diagram of crossed-lamellar structure, showing relationships of first order lamels, second order lamels and laths building the structure. ( b ) Pattern of the first order lamels in a radial section of the umbonal region of Trisidos tortuosa. x 280. (c) First order lamel pattern of the inner surface of the crossed-lamellar layer of Glycimeris glycimeris. x 280.

and MacClintock (1963, 1967). In extant bivalves crossed-lamellar structure is invariably aragonitic. It is built up of lenticular first order lamels (Bsggild, 1930), which interdigitate with their neighbours, and extend normal to the shell interior, throughout the whole shell layer (see Fig. 5). The first order lamels are built up of second order lamels that are inclined towards the shell interior, but are normal to the sides of the first order lamels. In adjacent first


507

order lamels the second order lamels are inclined in opposite directions but only two directions of inclination are present within the structure (Fig. 5 ; P1. 3, fig. I). Each second order lame1 is in turn built of laterally joined laths with each lath surrounded by organic matrix (Pl. 3, fig. 2). This shell structure type is strikingly coloured in peels and thin sections, adjacent lamels being straw or red-brown in colour. Crossed-lamellar structure superticially resembles polysynthetic twinning as developed in some inorganic systems (see, for instance, Figs. 128-131 of KlassenNeklyudova, 1964). However, the crystallographic orientations of alternate first order lamels are not consistent with such a relationship.

( 5 ) Complex crossed-lamellar structure We use the term complex crossed-lamellar structure in the original sense of Beggild (1930),and it is thus equivalent to the term ‘complex’ of Oberling (1964), which is not to be confused with the ‘complex’ structure of Baggild (1930)! This shell structure type is invariably aragonitic, and is present in the Arcacea, Limopsacea, many heterodonts and a few members of several other groups. It is built up of the same basic elements as crossed-lamellar structure, i.e. laterally contiguous laths which form second order lamels, and similarly beautiful colour patterns are seen in peels or sections. Second order lamels, but not first order lamels, are arranged in blocks, (Pl. 2, fig. 4). These blocks are not regular, and any section shows two opposed directions of lath inclination, separated by areas of granular appearance, corresponding to sections across laths. Since this pattern is seen in all sections, at least four directions of inclination of laths must be present, and perhaps more. There is considerable variation in the appearance of this structure, depending on the size of the blocks. Indeed, these may be large and distinct in the inner parts of a shell layer, but may pass into a complex of wisps of interdigitating second order lamels (Pl. 2, fig. 4). In the Limopsacea and some other bivalves there is a distinct columnar arrangement of the laths and MacClintock has interpreted similar structures in gastropods as a series of hollow cones, each built of radiating needles. He has also suggested that complex crossed-lamellar structure represents a form of spherulitic growth. (6) Homogeneous structure This is an unfortunate term, used by Baggild (1930) and other workers for a shell structure in which no obvious fabric is visible at optical level (e.g. P1. I , fig. 2), but which shows parallel extinction over wide areas of shell, with crystallographic c axes approximately normal to shell and layer surfaces. The structure is aragonitic. ‘ Homogeneous’ is a term of convenience, for electron-microscopy reveals that some homogeneous structures are simply very finely divided examples of crossedlamellar or complex-crossed lamellar structure. Indeed, these structures often grade into homogeneous structure (Pl. 4, fig. I). Some bivalves show a rather distinctive homogeneous structure, e.g. Arctica islandica. Here the whole shell is built up of small, irregular aragonite granules, each enclosed in an organic membrane. These granules are arranged in regular sheets and columns (Pl. I , fig. 2).

508

WILLIAM JAMES KENNEDY AND OTHERS (7) Myostracal layers

We have adopted this term, introduced by Oberling (1964), for the distinctive form of shell material laid down under areas of muscle attachment. It is partly equivalent to the term ‘helle Schicht’ of many authors, the ‘hypostracum’ of Thiele (1893) and Lowenstam (1964), and the ‘pellucid layer’ of Japanese workers (Kobayashi, 1964). The most important myostraca are the pallial, adductor and pedal myostraca. We do not know how bivalves are attached to their shells, although it has been suggested (Hubendick, 1958) that this might be by means of microvilli. It is clear, however, that muscle fibres are not directly attached to the shell.

inner shell layer rnyostracal pillars

tostracum

Fig. 6. (a) Block diagram of adductor myostracum. The outer shell layer is shown stippled. (b) Block diagram of myostracal pillars.

Two extreme conditions of myostracal development can be recognized. The first extreme is the occurrence of a thick sheet of carbonate within the shell (Pl. 3, fig. 4). This is invariably aragonitic, and has a characteristic prismatic structure. In the other extreme a myostracal trace can be followed through the shell, but no definite sheet of prisms can be located optically. Every intermediate exists, and a thick myostracum may pass laterally into a mere line. In most cases, a distinct prism layer is developed in association with pallial attachment in the umbonal area. When well developed, myostracal layers have a characteristic appearance :colourless, or pale grey, against the surrounding shell layers in sections, and transparent, against opaque shell layers on surfaces. The prisms building myostraca lie with their long axes normal to the layer surface, and crystallographic c axes in the same direction. Prism


509

outlines are highly irregular (Pl. I , fig. 3 ; Fig. 6 a), with re-entrant angles, great variation in size, and no development of thick interprismatic walls, although each prism is surrounded by organic matrix (Pl. 4, fig. 5 ) . In general, the pallial myostracum separates shell layers of differing structural type, and there is an apparent disconformity of growth lines at that point. This highly distinctive myostracal structure is not confined to well authenticated areas of muscle attachment. In the Chamacea, Carditacea and some Arcacea, Limopsacea and Myacea there are columns of myostracal-type prisms extending up from the pallial myostracum through the inner shell layer, to the shell interior (Pl. 2, fig. 8). On the shell interior they outcrop as oval or rounded bosses, These seem to represent sites of prolonged muscle attachment, previously unrecorded in bivalves, and we now know that they correspond to papillae on the outer mantle surface. In addition, discontinuous sheets and lenses of prisms occur in the inner layer of many bivalves (Pl. 2, fig. 6). These might be interpreted as indications of periodic muscle attachment, but we have no direct evidence of this. One interesting feature we have observed in bivalves where the pallial or adductor myostraca run through a single shell layer (as in the Spondylidae), is that there is continuity of the structural elements across quite thick sheets of prisms (Pl. 2, fig. 10).

(8) Layer relationships In describing the mineralogy and structure of bivalves, we speak of outer, middle and inner shell layers. These are strictly morphological distinctions ; we reject Oberling’s (1964) nomenclature as unsound. The contacts between shell layers show many features of interest, as in every case there is a transition zone, if not at optical level, at electronmicroscope level. At the contact between the nacreous and prismatic layers of Margaritifera and Anodonta there is for example a region where nacre tablets are developed amongst the innermost parts of prisms. In Neotrigonia there is a similar transition. At the contact of the prismatic and foliated layers of Ostrea there is a zone rich in organic matrix and with small hexagonal or rounded crystals (Watabe & Wilbur, 1961). At optical level, there is often a relationship between the building blocks of one layer and another. Thus lamels in crossed-lamellar layers often seem to have influenced the growth and siting of lath blocks in adjacent foliated or complex-crossed lamellar layers (Pl. 2, fig. 10).Clearly, pre-existing layers have a strong influence on the form of superimposed layers. (9) Other calcified elements In addition to the calcified valves, in many bivalves especially the Pteriomorphia, calcification also occurs in the inner layer of the ligament. This mineralization is invariably in the form of aragonite. The aragonite occurs as minute bundles of fine needles (Pl. 4, fig. 3) with their long axes growing normal to the surface of the mantle isthmus, which is responsible for the secretion of the ligament. The prodissoconch of bivalves is also calcified, and appears to have a structure different from that of the remainder of the shell. Because of the difficulty of identifying larval bivalves, it is only possible to study

510


spat of commercially used forms, but this has already given interesting information. Thus the prodissoconch of oysters, which have a largely calcitic shell, is aragonitic (Stenzel, 1964). Prodissoconch I of Pinctada martensii is formed from dahllite (Watabe, 1956; Wada, 1961). IV. THE DISTRIBUTION OF ARAGONITE, CALCITE AND THE SHELL STRUCTURE TYPES IN RECENT BIVALVES

In this description of the distribution of shell structure types and the shell mineralogy of bivalves, we use the classification of Newel1 (1965). As will be seen below, there is considerable uniformity of these features at superfamily level within Recent Bivalvia. Work in progress suggests that this is also broadly true for Tertiary and Mesozoic bivalves, and that it will prove an additional key to bivalve phylogeny. Subclass PALAEOTAXODONTA Order NUCULOIDA Superfamily Nuculacea These are entirely aragonitic, with an outer, composite prismatic layer, a middle layer of lenticular nacre and an inner layer of sheet nacre. The inner layer is bounded by the trace of the pallial line. Superfamily Nuculanacea These are entirely aragonitic, with a two-layered shell. There is an outer and an inner homogeneous layer, separated by the trace of the pallial line. In some fossil forms (i.e. Nuculana (Ryderiu)graphica Tate (Lias, Britain)) there is an inner, nacreous layer. Subclass CRYPTODONTA Order SOLEMYOIDA Superfamily Solemyacea Again entirely aragonitic, but with an outer shell layer of simple prisms, separated by the trace of the pallial line from an inner homogeneous layer. Subclass PTERIOMORPHIA Order ARCOIDA Superfamily Arcacea Also entirely aragonitic, with a two-layered shell consisting of an outer, crossedlamellar layer, and an inner, complex-crossed lamellar layer. These are separated by the pallial myostracum. Myostracal pillars are present in the inner layer of some species. Superfamily Limopsacea The shell composition and structure is identical with that of the Arcacea.


511

Order MYTILOIDA Superfamily Mytilacea This superfamily has a complex and variable mineralogy. Some forms are entirely aragonitic, in others both aragonite and calcite are present. It is on this group that the most detailed investigations in shell mineralogy have been carried out, and these are discussed below. In Mytilus sensu stricto two conditions are seen. Some species have an entirely aragonitic two-layered nacreous shell, the layers being separated by the pallial myostracum. These species are tropical (Hudson, 1968). I n Mytilus edulis (Linnaeus) there are three shell layers, an outer, finely prismatic calcite layer, and within this, two aragonitic nacre layers, separated by the pallial myostracum. In Mytilus culiforniunus Conrad there is a further, innermost area of finely prismatic calcite. Species of Modiolus show a similar differentiation into two-layered aragonitic shells and three-layered bi-mineralic shells. Bruchiodontes and Septifer have two-layered, nacreous, aragonitic shells. Lithophuga and Botula have a third, outer calcitic layer. In Stuveliu the shell is wholly aragonitic, with an outer, nacreous layer and an inner, complex crossed-lamellar layer. In many species, sheets of myostracal prisms are present in the inner shell layer. The ligament is aragonitic. Superfamily Pinnacea These show a complex mineralogy. There is an outer layer of simple calcitic prisms, and, within this, two aragonitic layers of sheet nacre. The ligament is aragonitic. Order PTERIODIA Superfamily Pteriacea This superfamily has a similar complex mineralogy to the Pinnacea: an outer layer of simple calcitic prisms, and within this, two aragonitic nacre layers, separated by the trace of the pallial myostracum. The ligament is aragonitic. Superfamily Pectinacea Family Pectinidae Another family showing a complex mineralogy. Most of the shell is built of an outer, calcitic foliated layer. Some forms possess a middle aragonitic crossed-lamellar layer, whilst all have a coarsely foliated calcitic inner layer. There are, of course, aragonitic myostraca and the ligament is aragonitic. Family Spondylidae Here, there is an outer, calcitic foliated layer, with both middle and inner crossedlamellar layers. These are separated by the trace of the pallial myostracum. The ligament is, again, aragonitic. Family Plicatulidae These have the same shell-structure and mineralogy as the Spondylidae.

5 12


Superfamily Anomiacea Most of the shell of this group is built of an outer, calcitic foliated layer. Within this there is an aragonitic complex crossed-lamellar layer, restricted to the area around the muscle scars. Ligament and byssus are both calcified and both aragonitic. Superfamily Limacea These show a complex mineralogy. There is an outer, calcitic foliated layer and a crossed-lamellar layer within this. Below the hinge, in a few species, there is a restricted area of complex crossed-lamellar structure on the inner side of this. Suborder OSTREINA Superfamily Ostreacea Oysters are almost wholly calcitic, with an outer prismatic layer, which is very thin, most of the shell consisting of foliated structure. Aragonite is confined to the prodissoconch, ligament and myostraca (Stenzel, 1962, 1963, 1964). Subclass PALAEOHETERODONTA Order ACTINODONTA Superfamily Unionacea These are entirely aragonitic. I n the Unionidae and Margaritiferidae there are three shell layers: an outer, simple prismatic layer, a middle layer of lenticular nacre and an inner layer of sheet nacre bounded by a trace of the pallial myostracum. In the cemented Etheriidae there are only two layers, both of sheet nacre, and separated by the trace of the pallial myostracum. Order TRIGONOIDA Superfamily Trigonacea The Trigonacea have an entirely aragonitic three-layered shell: an outer layer of simple prisms, a middle layer of lenticular nacre and an inner layer of sheet nacre, bounded by the trace of the pallial line. Subclass HETERODONTA ( ?) Order HIPPURITOIDA Superfamily Chamacea Nearly all members of this group have wholly aragonitic two-layered shells. There is usually an outer, crossed-lamellar layer and an inner, complex crossed-lamellar layer with myostracal pillars and sheets of myostracal-type prisms, which is bounded by the trace of the pallial myostracum. Two species, Chama pellucida Broderip and C. exogyra Conrad, develop an outer calcitic layer with a finely prismatic structure (Lowenstam, 1954a, b, 1963, 1964). Calcite is developed in some extinct Hippuritoidea, i.e. the rudists (Kennedy & Taylor, I 968).


513

Order VENEROIDA Suborder LUCININA Superfamily Lucinacea This superfamily has an entirely aragonitic three-layered shell, consisting of an outer composite prismatic layer, a middle crossed-lamellar layer and an inner complex crossed-lamellar layer bounded by the trace of the pallial line. Myostracal prisms are prominent in the inner layer. Superfamily Leptonacea These have entirely aragonitic two-layered shells, composed of an outer, crossedlamellar layer and an inner, complex crossed-lamellar layer, which are separated by the trace of the pallial myostracum. Superfamily Cyamiacea These are wholly aragonitic and have a structure comparable to the Leptonacea. Superfamily Carditacea These have entirely aragonitic two-layered shells formed of an outer crossedlamellar layer and a complex crossed-lamellar inner layer with myostracal pillars and layers of myostracal-type prisms. Superfamily Crassatellacea Another superfamily with an entirely aragonitic two-layered shell. There is an outer crossed-lamellar layer and an inner layer built up of myostracal-type prisms with small areas of complex crossed-lamellar, or homogeneous structure. The two layers are separated by the trace of the pallial myostracum. Superfamily Cardiacea These have an entirely aragonitic two-layered shell. There is an outer, crossedlamellar layer and an inner complex crossed-lamellar layer, bounded by the trace of the pallial myostracum. Superfamily Tridacnacea The Tridacnacea have an entirely aragonitic two-layered shell, with an outer crossed-lamellar layer and an inner complex crossed-lamellar layer, bounded by the trace of the pallial myostracum. Sheets of myostracal-type prisms are present in the inner layer of some species. Superfamily Mactracea An entirely aragonitic group with a two-layered shell consisting of an outer, crossedlamellar layer and an inner complex crossed-lamellar layer bounded by the trace of the pallial myostracum. The lamels in the outer layer are rather fine, and parts of the layer appear homogeneous. Bands of myostracal-type prisms are present in the inner layer. 33

Biol Rev. 4 4

5 ‘4


Superfamily Solenacea These have an entirely aragonitic two-layered shell. There is an outer, crossedlamellar layer and an inner homogeneous layer, bounded by the trace of the pallial myostracum. Electron microscopy shows that the ‘homogeneous’ layer seen at optical level is in fact built of very fine complex crossed-lamellar structure. Superfamily Tellinacea In most species the entirely aragonitic shell is three layered, with an outer, composite prismatic layer, a middle crossed-lamellar layer and an inner, complex crossedlamellar to homogeneous layer bounded by the trace of the pallial myostracum. A few species have two layered shells with an outer crossed-lamellar layer and an inner complex crossed-lamellar layer. Suborder ARCTICINA Superfamily Arcticacea A group with entirely aragonitic two-layered shells. In Arctica islandica (Linnaeus) both layers are homogeneous whereas in Trapezium, Velorita and Coralliophaga there is an outer crossed-lamellar layer and an inner complex crossed-lamellar layer. Superfamily Dreissenacea Yet another superfamily having an entirely aragonitic two-layered shell with an outer, crossed-lamellar layer and an inner complex crossed-lamellar layer bounded by the trace of the pallial myostracum. Superfamily Glossacea These possess an entirely aragonitic two-layered shell, which in Glossus humanus (Linnaeus) consists of an outer homogeneous layer and an inner, complex crossedlamellar layer, with bands of myostracal-type prisms. However, in Meiocardia lamarckii (Reeve) the outer layer has a crossed-lamellar structure. Superfamily Corbiculacea These have an entirely aragonitic two-layered shell, with an outer, crossed-lamellar layer and an inner complex crossed-lamellar layer bounded by the trace of the pallial myostracum. Superfamily Veneracea The outer layer of these entirely aragonitic two-layered shells appears rather complex, due to changes in structure resulting from the form of the margin. There are areas of crossed-lamellar and homogeneous structures, and an outer zone of crossed lamels with divergent feathery form. The inner layer, bounded by the trace of the pallial myostracum, has a complex crossed-lamellar, or homogeneous structure.

Bivalve shell mineralogy Order MYOIDA Suborder MYINA Superfamily Myacea These entirely aragonitic two-layered shells have an outer crossed-lamellar layer and an inner complex crossed-lamellar layer bounded by the trace of the pallial myostracum. Sheets of myostracal-type prisms are well developed in the inner layer of all species and in addition myostracal pillars are present in some species of Corbulu. Superfamily Gastrochaenacea These have wholly aragonitic two-layered shells : an outer crossed-lamellar layer and an inner, complex crossed-lamellar layer, bounded by the trace of the pallial line. Superfamily Hiatellacea Wholly aragonitic two-layered shells with both layers having a homogeneous structure. Suborder PHOLADINA Superfamily Pholadacea Family Pholadidae The Pholadidae have wholly aragonitic two-layered shells, consisting of an outer crossed-lamellar layer and an inner complex crossed-lamellar layer bounded by the trace of the pallial myostracum. Layers of myostracal-type prisms are present in the inner layer. Subclass ANOMALODESMATA Order PHOLADOMYINA Superfamily Pholadomyacea This superfamily has a three-layered entirely aragonitic shell with an outer prismatic layer, and the middle and inner nacreous layers separated by the trace of the pallial myostracum. Superfamily Pandoracea These have an entirely aragonitic three-layered shell. There is an outer, prismatic layer, a middle layer of lenticular nacre and an inner layer of sheet nacre bounded by the trace of the pallial myostracum. Suborder CLAVAGELLINA Superfamily Clavagellacea In this group both the shell and the tube are aragonitic. The shell appears to be homogeneous, as does the tube, which shows an obscure concentric structure.

33-2

5'6


Order POROMYOIDA Superfamily Poromyacea These have a wholly aragonitic, two layered shell of homogeneous structure. V. CONTROLS ON SHELL MINERALOGY

Introduction From the previous section it is clear that the distribution of aragonite and calcite, and of each shell structure type, is not random, but is very constant within bivalve superfamilies. Calcite is confined to the superfamilies Mytilacea, Pinnacea, Pteriacea, Pectinacea, Anomiacea, Limacea and Ostreacea of the Pteriomorphia, and to two species of Chamacea and the extinct rudists, of the subclass Heterodonta. (I)

(2) Biological

controls The prime control of shell mineralogy is clearly genetic and similar structure/ mineralogy relationships can be traced back to at least the Mesozoic/Palaeozoic boundary. We can assume from this that bivalves established their shell mineralogy many millions of years ago. On present knowledge the chief exception, the rudists, were derived from the wholly aragonitic pachydont Megalodon, calcite probably not appearing until the Upper Jurassic. The recent exceptions, Chama pellucida and C . exogyra, are a more short lived peculiarity, arising in the late Tertiary. All Tertiary species of Chama, and the few known Cretaceous forms we have seen are wholly aragonitic. T o investigate the reasons for strong biological control on shell mineralogy, it is necessary to consider the methods by which the animal could achieve this. From our brief introductory account of shell formation, these can be listed as follows: ( a ) via the mantle cells beneath the various shell layers; (b) via the nature of the extrapallial fluid; (c) via the nature of the periostracum; (d) Via the organic matrix of the various shell layers. All of these are closely interconnected, in that the mantle cells are responsible for the nature of the extrapallial fluid, from which the periostracum and organic matrix are deposited. It is therefore apparent that any conclusions drawn here concern the final stages of calcification. The initiation of these stages lies within the cell, and beyond our present knowledge. (a) The mantle cells. The work of Trueman (1949),Yonge (1953),Beedham (1958a), Dunachie (1963)and Kawaguti & Ikemoto (1962)has shown that cells in various parts of the outer mantle vary in shape, structure and histochemical properties. Each cell changes in respect of these features as it moves around the outer mantle fold, lying below the various shell layers in turn. These are the periostracum, outer shell layer, middle shell layer (perhaps), pallid myostracum and inner shell layer. This plasticity of function is not just a property of cells at the mantle margins. Where a shell is damaged or parasitized, this will often act as a stimulus for regeneration


5 I7

in the whole outer mantle of the animal, and a complex sequence of periostracum, outer shell layer and inner shell layer may be repeated in miniature, and preserved in the body of the shell (Pl. I , fig. 5). However, because of lack of continuity between mantle and shell over wide areas, it is difficult to see or envisage a direct cellular control. (b) The extrapallialfluid. The extrapallial fluid of most bivalves can be regarded as lying within two distinct divisions of the extrapallial cavity. One lies within the pallial line and the other lies between this and the periostracal groove. Since this division usually corresponds to the division between shell layers, there is the obvious possibility that slight differences which might develop in these regions could control the nature of shell layers. To date no distinction has been made between these two regions in investigations of the fluid. It must be noted that layers of different mineralogy are bathed by the same area of fluid (usually that outside the pallial line). The extrapallial fluid contains a similar range of cations and anions to those present in the blood of the animal concerned, i.e. Naf, Kf, Ca2+,Mg2+,Mn4+,C1-, SO:- and PO:-, (de Waele, 1930; from Anodonta cygnea). The presence of mucopolysaccharides in the blood and organic matrix of the shell suggests, not surprisingly, that these are also present in the extrapallial fluid. More important is the work of Kobayashi (1964), who has established that where the bivalve shell is almost wholly calcitic (in Aepuipecten, Chlumys and Crassostrea species) there is a single protein fraction in the fluid, whilst in shells which are wholly aragonitic (Busycon, Elliptio, Menenaria and Crassatella) or contain substantial amounts of aragonite and calcite (Modiolus sp.) there are three protein fractions. There is thus a tenuous connexion between the protein composition of extrapallial fluid and crystal type. (c) The periostracum. This is an extremely complicated tanned protein layer, laid down in the periostracal groove (Kessel, 1944), with a complex layered structure (Kawaguti & Ikemoto, 1962; Dunachie, 1963; Beedham & Owen, 1965). It is the substrate upon which calcification begins, and in species with substantial amounts of aragonite and calcite in the shell, this initial calcification is always in the form of calcite, and the periostracum is thick. Degens et al. (1967) give data for the amino acid constituents of the periostracum of eleven bivalve species. Unfortunately, only one of these, Mytilus edulis, has an outer calcitic shell layer, so that we cannot yet assess whether there is a difference in the composition of the periostracum upon which calcite and aragonite are deposited. ( d ) The organic matrix. As the calcified shell of the bivalve forms, two distinct phases are laid down simultaneously from the extrapallial fluid, initially on to the periostracum, and subsequently on to pre-existing shell carbonate and protein. As already noted, in the second type of deposition, epitaxis of carbonate on carbonate, and probably of protein on protein is very important. Beyond this, simple inorganic processes, with slight biological modification are responsible for many shell fabrics, whilst similar processes seem to be acting during initial deposition upon the periostracum.

5’8


OTHERS

The organic matrix is a complex series of proteins, intimately related to the carbonate, occurring within and around the smallest building blocks of each shell fabric. Frequently it behaves passively, being pushed aside by carbonate masses, e.g. by simple prisms (p. 505). The electron microscope studies of GrCgoire (1957, 1961a, b, 1962, 1967), Watabe (1965)and our own work have revealed much of the fine structure of organic matrix, especially in nacre. The protein here is in the form of fenestrate sheets of knobbly cords, with a distinctive structure in the various molluscan classes. Watabe’s work has revealed the fine-structure of these cords, showing that they in turn are built of many minute helically coiled fibres. More recently, Travis, Francois, Bonar & Glimcher (1967) have reported and figured collagen fibrils from the organic matrix of bivalves, although in only minor amounts. The significance of this remains to be explained. It is apparent that crystallites of calcium carbonate building the bivalve shell are within, around and upon organic matrix, and many workers have suggested that the matrix has a prime role in calcification, influencing crystal type and structure of shell layers (Roche, Ranson & Eysseric-Lafon, 1951; Ranson, 1952; Watabe & Wilbur, 1960; Wilbur & Watabe, 1963; Degens et al. 1967; Wada, 1961, etc.). Several workers have reported parallelism between the structure of the organic matrix and that of the carbonate (e.g. Watabe, 1965), and these observations have led to what can be regarded as a general theory of calcification in molluscs (Hare, 1963)~ supported by Degens et al. (1967), Towe & Cifelli (1967) and Travis et al. (1967),who have extended it to other invertebrates, comparable to that acting in bone tissues (Glimcher, 1959, 1960; Eastoe, 1968, with references). The view is that mineralization is basically epitaxial, with the proteinaceous organic matrix acting as a ‘ template’ for calcification (Degens, 1965). In particular, Hare (1963) has suggested that certain side chains in the protein matrix may concentrate Ca2+ and COi- at specific positions and thus provide the critical pre-calcification situation. Depending on the siting of such nucleation points, it would thus be possible to explain the development of either calcite or aragonite on a particular matrix, and to explain observed parallelism between carbonate and organic matrix structure. Hare considers that aspartic and glutamic acid side chains attract Ca2+ ions, whilst basic side chains attract C0:- and HCO; ions. As an alterative view, Wada (1964a, b) has suggested that certain acid mucopolysaccharides, bound in with the conchiolin protein by specific amino acid side chains, may offer appropriate sites. Considerable work is still needed before the template pattern established for bone can be accepted for bivalve calcification. It has, however, been established that there are differences in the protein make-up of organic matrices of aragonitic and calcitic shell layers (see analyses by Roche et al. 1951;Beedham, 19586; Tanaka, Hatano & Itasaka, 1960; Hare, 1963; Hare & Abelson, 1964, 1965; Stegemann, 1961, 1963; Degens et al. 1967; Bricteux-GrCgoire, Florkin & GrCgoire, 1968), so that a template mechanism would definitely seem possible. Striking evidence of the role of organic matrix in controlling shell mineralogy comes


5 19

from the work of Watabe & Wilbur (1960), Wilbur & Watabe (1963)and Wilbur (1964). These workers experimented with organic matrix obtained by decalcifying bivalve shell layers, using this as a basis for precipitation. Their experiments were carried out in vitro, in aqueous solutions of calcium bicarbonate, and in vivo by inserting the organic matrix between the shell and mantle of another bivalve. It was demonstrated that the organic matrix from Elliptio (aragonitic) and from the aragonitic layers of Atrina and Pinctada induced aragonite precipitation in vitro, and that matrix from the calcite layer of Pinctada induced calcite precipitation under similar conditions. Further, decalcified organic matrix from aragonitic shells induced aragonite deposition in 25 yo of the experiments when it was placed between the shell and mantle of Crassostrea, which is largely calcitic. (e) Discussion. The role of periostracum and/or organic matrix in initiating calcification, and thus controlling the deposition of either aragonite or calcite, cannot be doubted, but any influence beyond this seems doubtful in view of the obvious similarities between shell fabrics and inorganic crystal growth forms. It still remains to be explained how a bivalve can lay down two sorts of calcium carbonate simultaneously in the same area of extrapallial fluid. We believe that the periostracum plays an important role, and replacement of calcite deposition by aragonite deposition may, in part, be a reflexion of the waning influence of this part of the shell. Mention must be made here of a fundamentally different concept of calcification, which has recently been propounded by Digby (1965, 1968a, b). Digby has suggested that in crabs, in vertebrate bone and in molluscs calcification is possible as a result of the presence of a tanned protein layer. This acts as a semi-conducting membrane; calcium salts are deposited on the inner surface, which acts as a cathodic region. Whilst Digby’s observations are not readily explained away, it is difficult to conceive a way in which they could produce the observed variation in shell mineralogy, shell structure type or the observed variation in distribution of both these features.

(3) Environmental controls Boggild (1930) made the earliest contribution to our knowledge of environmental controls on shell mineralogy when he noted that of all the molluscs examined by him only one freshwater form (a gastropod) had calcite in the shell. Two decades later Lowenstam (19544 b) investigated these matters in detail, and from this and subsequent work (Lowenstam, 1964; Dodd, 1963, 1964, 1966a; Eisma, 1966) it is now possible to discuss variation in shell mineralogy in terms of two factors, temperature and salinity. (a) Temperature. Lowenstam (1954.~2,b, 1964) established that there is an inverse relationship between the percentage of calcite in the shell and the mean temperature of the environment inhabited by the bivalve. Thus where a bivalve lays down both aragonite and calcite in its shell those individuals inhabiting cooler waters will have more calcite in their shells than those inhabiting warmer waters. In addition, where species within a genus may either have wholly aragonitic, or aragonite and calcite shells, the latter are cooler water forms. As well as demonstrating these facts in bivalves,

520


Lowenstam indicated that the same factor operated in gastropods, cirripedes, octocorals, serpulids and bryozoans. Lowenstam’s data must be interpreted with some caution. We know that in bivalves depositing both aragonite and calcite, the two polymorphs are restricted to definite shell layers, and that the proportions of these layers vary from species to species within a genus, and with the size and age of the individual. We would therefore question Lowenstam’s data for Modiolus (‘ Volsella’) sincethis relates to three or four species, with the weight of the shells varying from 2.63 to 24-49g., and the aragonite content varying by 18%. Similarly, data for Pinna, Pteria, Anomia and Pecten can be questioned where differences in aragonite: calcite ratios are of the order of 2 yo and where weights differ by factors of up to ten. Thus, an assemblage of Pedalion from a single locality varied from 76 to 42% aragonite, with weights between 8.99 and 0.040g. It is also difficult to assess data for ‘Pecten’ and ‘Pinna’,as there are great variations in the extent of aragonite layers in these groups, in particular the sympatric genera of Pinnacea, Atrina, Streptopinna and Pinna. However, these are small criticisms of an outstanding work. Having established that the aragonite :calcite ratio is temperature controlled, it follows that this ratio will change throughout the life of an individual bivalve inhabiting an area with marked seasonal variations in temperature. It will also follow that the aragonite: calcite ratio of a whole shell will vary with its age and size. Any quantitative temperature data derived from these facts will therefore relate to some mean temperature. It is also clear that a shell which begins growth in the spring will have a higher aragonite content at any given size when compared with a shell which began growth in the autumn. These differences will tend to diminish as shells increase in size. Thus a collection of shells from a given locality will show variation in aragonite:calcite ratios, depending on their size and time of spawning; the shells which most satisfactorily represent the mean temperature of the environment will be the largest. It is also necessary for the shell to be secreted all the year round. If a species ceases to deposit carbonate below a particular temperature, any measurement of temperature based on shell mineralogy will be too high. It seems, however, that secretion in bivalves is continuous throughout the year, only varying in extent. I n spite of these inherent difficulties, Dodd (1963, 1964)has studied two mytilid species, Mytilus californianusand M . edulis edulis and diegoensis,and has produced interesting results. Mytilus californianusis an interesting species in that there are two calcitic layers within the shell (Fig. 7). Thus there is an outer, finely prismatic calcite layer, a nacreous layer between this and the pallial myostracum, a nacreous layer within the pallial line and, finally, a finely prismatic calcitic inner layer below the umbones. This species spawns and deposits shell carbonate throughout the year, and shows temperature effects in a number of ways. Therefore if young individuals are collected throughout the year it is to be expected that those that began growth in the spring will have a higher aragonite content than those beginning growth in the autumn. This is, indeed, true (Fig, 8), but only for individuals more than 15 mm. long. The effect


521

becomes more pronounced as individuals increase in size, i.e. as the influence of the non-temperature dependant early stages diminishes (Dodd, 1963). It is also to be expected that measurement on shells of varying lengths will produce a distinct oscillatory curve, as can be seen from Fig. 9 (Dodd, 1963, Fig. 4), although, again, it is not seen in small individuals. Beyond this, data from localities with differing temperatures must be studied, outer prismatic

ner nacreous

outer prismatic /calcite layer

inner calcite layer

aragonite layer

(b) Fig. 7. Diagrammatic radial sections through the shells of ( a ) Mytilus edulis and ( b ) Mytilus californianus to show the disposition of shell layers and their mineralogy.

Fig. 8. Seasonal variation in per cent aragonite in specimens of various size ranges of Mytilus californianus from Corona del Mar, California (from Dodd, 1963, Fig. I ) . -, 11-14mm; -.- , 17-zzmm; ......, z5-z9mm.

522


either by taking the mean value of the aragonite: calcite ratio of many shells from each locality, or by taking the ratio of a single large shell. From Fig. 10 (from Dodd, 1963) it is clear that there is a strong correlation between temperature and aragonite content. Indeed, a least-squares line from Dodd's data is:

yo aragonite

("C) -6.1. This is a correlation between the aragonite: calcite ratio and the mean locality temperature, not the mean temperature of shell deposition. 110

-

100

-

90

-

80

-

70

-

60

-

A

= 2-55temperature

Y

-5 M

- 50 40

-

30

-

20

-

10

25

f I I I

I

I

30

35

40

aragonite

(%)

Fig. 9. Growth series of Mytilus califo~niunusfrom Corona del Mar, California, showing variation in per cent aragonite with shell length. Specimens collected in early autumn (from Dodd, 1963, Fig. 4).

Investigation of the structure of M . californianus reveals how this variation is produced (Dodd, 1964, Fig. 5). The outer calcitic layer thickness increases regularly with increasing size, but the inner, calcitic layer shows periods of maximum deposition


523

alternating with periods of minimal deposition. Thus tongues of calcite and aragonite interdigitate, corresponding to summer and winter deposition. How the animal achieves this, however, remains something of a problem. In Mytilus edulis subspecies, Dodd found no such well defined mineralogical variation, other than with size. This seems to be due to the fact that there is only an outer calcitic layer (Fig. 8) which increases regularly in thickness with increasing size and age. Yet, Lowenstam (19544 b) found distinct temperature controls on mineralogy in East Coast of America M . edulis.

Fig. 10. Variation of per cent aragonite in Mytilus edulis with mean annual temperature of collecting locality. Only values for station means and largest individuals used (from Dodd, 1963, Fig. 5). 0 , Mean for locality; largest individual. Correction :-the equation of the least squares line should read: yo aragonite = 2.55 T ("C)- 6 . 1 .

+,

(b) Salinity. Lowenstam (1954)noted that salinity appeared to affect the mineralogy of Mytilus. In particular, he noted that species from the Baltic had anomalously high aragonite contents, too high to be explained by temperature alone. Subsequently Dodd (1963) investigated this in specimens of Mytilus edulis edulis from the Juan de Fuca Strait and Hood Canal regions of Washington. At localities in these areas, which he considered to have uniform temperatures, the salinities varied from 32.4 to 18.60%~. On the basis of analyses, he demonstrated an inverse relationship between salinity and aragonite content (Fig. I I ) . Dodd also mentioned similar (unpublished) findings from Mytilus edulis diegoensis populations collected from San Francisco (1963),which confirmed Lowenstam's work.


524

OTHERS

Three years later Eisma (1966) produced data based on European Mytilus edulis that flatly contradicted these results. He took specimens from Dutch coastal waters and the Zuider Zee, and found that there was no relationship between salinity and aragonite content (Fig. 12). 25

al

.-uC

15

10

20

25 salinity

Fig.

g

50

I I.

Variations in per cent aragonite with salinity at collecting time in Mytilus edulis edulis, from Washington (from Dodd, 1963, Fig. 8).

-

.-g 4 0 0

c)

M

. . .

.

e

30

30

(yoo)

I

Fig.

12.

I

I

I

I

I

I

Relation between aragonite content and avcrage salinity for Dutch Mytilus edulis shells (from Eisma, 1966, Fig. 2).

Such results are puzzling, and Eisma sought to explain them in two ways. ( I ) By considering that there was specific difference between his shells and those from the Baltic and Washington, and that they behaved differently with respect to salinity. Dodd has shown that there is a difference in variations of shell mineralogy in M.


525

edulis edulis and M . edulis diegoensis, but both show (or are said to show) salinitycontrolled variation. (2) By considering that observed increases in aragonite content in brackish water Mytilus are due not to salinity, but to an accompanying increase in temperature, not detected by Lowenstam and Dodd. In particular Lowenstam’s findings can be explained by the fact that in some parts of the Baltic very high summer temperatures of local occurrence are found. This could then be the reason, for only two of eight Baltic samples studied by Lowenstam show a strong negative salinity/aragonite % correlation. 35

30

h

.=uc

25

0

2 m L

20

I

20

Fig.

13.

I

I

25 salinity

30

(yoo)

Variation of per cent aragonite with salinity for specimens of Mytilus edulis diegoensis from the San Francisco, California, area (from Dodd, 1966, Fig. 3).

Referring to Dodd’s material, Eisma questions the presumed temperature uniformity in the region studied. Instead, he suggests that the results are due to high temperature in areas of lower salinity, as these protected bays are not subject to mixing with cooler oceanic waters. His Fig. 3 strongly supports this point. As Dodd (1966, in reply to Eisma) notes, the answer to this paradox lies in the study of experimentally grown animals reared under known temperature/salinity conditions. In support of his contentions, he produces data for M . edulis diegoensis (Fig. 13) which show an inverse salinity relationship.

526


These differences in results may be due to subspecific differences between the populations studied by various workers. Only experimental data can solve the problem. As well as the biological controls, there is thus a strong environmental effect on shell mineralogy. There is a definite negative correlation between temperature and the calcite content of shells. There is also a possible negative correlation between aragonite content and salinity. It remains to reconcile these two facts. T h e only clues we have at present are from the work of Degens et al. (1967), who have demonstrated that there is constancy of amino acid composition of organic matrix between bivalves of the same species in the same environment (see also Degens & Spencer, 1966), but that there are differences between bivalves from different environments. It is unfortunate that the only species studied to date, Anadara and Mulina, are both aragonitic, but these show the following variations. In Anudura threonine, glutamic acid, proline, leucine, isoleucine, valine and phenylalanine are relatively concentrated with respect to glycine, serine, lysine, histidine and arginine in those specimens which are from warmer waters. In Mulina glycine, isoleucine, leucine and valine are relatively concentrated with re, p ect to threonine, glutamic acid and proline in specimens from warm waters. Environmental changes also affect the composition of the periostracum (Degens et al. 1967, table 11). When we know of comparable changes in bimineralic species it will be possible to assess the meaning of these variations, and perhaps to understand the mineralogical variation more fully. Q

VI. SUMMARY

Bivalves lay down two forms of calcium carbonate in their shells, aragonite and calcite. Shells may be wholly aragonitic, or may contain both aragonite and calcite, in separate monomineralic layers. Shells are built up of several layers of distinct aggregations of calcium carbonate crystals. These aggregations are referred to as shell structures and their general features are described. Aragonite occurs as prismatic, nacreous, crossed-lamellar, complex crossed-lamellar and homogeneous structures. Calcite occurs as prismatic or foliated structures. Myostracal layers (calcium carbonate laid down below sites of muscle attachment) are always aragonitic. T h e ligament and byssus when calcified are also invariably aragonitic. A summary of the occurrence of calcite and aragonite and the associated shell structures is given. Calcite is found only in the outer layer of superfamilies belonging to the subclass Pteriomorphia with the exception of two species only from the Heterodont superfamily Chamacea. Generally within a superfamily shell structure and mineralogy are very constant. In all superfamilies these combinations have existed for many millions of years. It is therefore demonstrated that the prime control on shell mineralogy is genetic. Possible controls on mineralogy by the mantle cells, nature of the extrapallial fluid, nature of the periostracum and the organic matrix of the shell layers are discussed. It is known that environmental factors may modify the basic mineralogy/shell structure pattern within a superfamily. Thus there is an inverse relationship between


527

the percentage of calcite in the shell and the mean temperature of the environment inhabited by the bivalve. A critical examination of published data shows that the evidence is convincing only in the superfamily Mytilacea. The species Mytilus californianus, which shows the greatest temperature effects, is peculiar amongst the Mytilacea in having an inner calcite layer as well as an outer one. Conflicting evidence for an inverse relationship between salinity and aragonite content is reviewed. The differences of opinion cannot be resolved without experimental work. We are grateful to the following for much useful discussion, and encouragement in many ways: D r J. R. Baker, Dr G. E. Beedham, Dr B. C. M. Butler, D r A. Hallam, Dr J. D. Hudson, Dr R. P. S. Jefferies, Mr J. Macrae, Dr W. S. McKerrow, Mr N. J. Morris, M r C. P. Palmer, M r N. Tebble, D r E. R. Trueman and Professor A. Williams. Our best thanks are to Mr R. Cleevley for critically reviewing the manuscript. The following have rendered us considerable technical assistance : the staff of the electromicroscopy unit of the British Museum (Natural History), under the direction of Mr B. Martin; the technical staff of the Department of Geology, King’s College, London and of the Department of Geology and Mineralogy, Oxford; Mrs J. M. Hall, and Mr G. Burton. VII. REFERENCES BATHURST, R. G. C. (1966). Boring algae, micrite envelopes and lithification of molluscan biosparites. Geol. J . 5. 15-32, BEEDHAM, G. E. ( 1 9 5 8 ~ )Observations . on the mantle of the Lamellibranchia. Q. Jl microsc. Sci. 99, I 8 I -97. BEEDHAM, G. E. (19586). Observations on the non-calcareous component of the shell of the Lamellibranchia. Q. Jl microsc. Sci. 99, 341-57. BEEDHAM, G. E. & OWEN,G. (1965). The mantle and shell of Solemya parkinsoni (Protobranchia: Bivalvia). Proc. zool. Sac. Lond. 145, 405-30. H. (1968). An electron microscope study of the formation of the perioBEVELANDER, G. & NAKAHARA, stracum in Macrocallista maculata. Calc. Tiss. Res. I, 55-67. 0. B. (1930). The shell structure of the mollusks. K . danske Vidensk. Selsk. SKY.2, 232-325. BOGGILD, BRICTEUX-GRBGOIRE, S., FLORKIN, M. & GRBGOIRE, C. (1968). Prism conchiolin of modern or fossil molluscan shells. An example of palaeization. Comp. Biochem. Physiol. 24, 567-72. W. B. (1844). On the microscopic structure of shells. Rep. B Y . Ass. Adwmt Sci. 1-24. CARPENTER, CARPENTER, W. B. (1848). Report on the microscopic structure of shells, Pt. 2 . Rep. B Y . Ass. Adwmt Sci. (from 1847) London, 93-134. PIS. 1-20. DEGENS, E. T. (1965). Geochemistry of Sediments: a Brief Survey. ixf34z pp. New Jersey. DEGENS,E. T. & SPENCER, D. W. (1966). Data file on amino acid distribution in calcified and uncalcified tissues of shell forming organisms. Tech. rep. Woods Hole oceanogr. Znstn ref. no. 66-77. E. T., SPENCER, D. W. & PARKER, R. H. (1967). Palaeobiochemistry of molluscan shell proteins. DEGENS, Comp. Biochem. Physiol. 20, 553-79. DIGBY, P. S. B. (1965). Semi-conduction and electrode processes in biological material. In crustacea and certain soft-bodied forms. Proc. R . Sac. B 161,504-25. DIGBY,P. S. B. ( 1 9 6 8 ~ ) .Mobility and crystalline form of the lime in the cuticle of the shore crab, Carcinus maenas. J . Zool. 154, 273-86. DIGBY,P. S. B. (19686). The mechanism of calicification in the molluscan shell. Symp. zool. Sac. Lond. 22, 93-107. DODD,J. R. (1963). Palaeoecological implications of shell mineralogy in two pelecypod species. J . Geol. ~~

71,1 - 1 1 . DODD,J. R. (1964). Environmentally controlled variation in the shell structure of a pelecypod species. J . Palaeont. 38, 1 0 6 5 7 1 . . influence of salinity on mollusk shell mineralogy. J . Geol. 74, 85-9. DODD,J. R. ( 1 9 6 6 ~ )The DODD,J. R. (19666). Processes of conversion of aragonite to calcite with examples from the Cretaceous of Texas. J. sedim. petrol. 36, 733-41. J. F. (1963). The periostracum of Mytilus edulis. Trans. Roy. Sac. Edinb. 65, 383-411. DUNACHIE, EASTOE, J. E. (1968). Chemical aspects of the matrix concept in calcified tissue organisation. Calc. Tiss. Res. 2, 1-19.

528


EISMA,D. (1966). The influence of salinity on mollusk shell mineralogy: a discussion. J . Geol. 74, 89-94. GLIMCHER, M. J. (1959). Molecular biology of mineralised tissues with particular reference to bone. Rev. mod. Phys. 31, 359-93. M. J. (1960). Specificity of the molecular structure of organic matrices in mineralisation. GLIMCHER, In Calcification in Biological Systems, ed. R. F. Sognnaes. Am. Ass. Advmt Sci. 64,421-87. D. D. (1958). Dendritic form of crystals produced in antiskeletal GORODETSKY, A. F. & SARATOVKIN, growth. Repts 1st conf. on crystal growth, U.S.S.R. 151-8. G ~ G O I RC. E ,(1957). Topography of the organic components in mother-of-pearl. J. biophys. biochem. Cytol. 3, 797-808. GR~GOIRE, C. ( 1 9 6 1 ~ ) Structure . of the conchiolin cases of the prisms in Mytilus edulis Linne. J . biophys. biochem. Cytol. 9, 395-400. GRBGOIRE, C. (1961b). Sur la structure submicroscopique de la conchioline associke aux prismes de coquilles de mollusques. Bull. Inst. r . Sci. nut. Belg. 37, 1-34. GRBGOIRE, C. (1962). On submicroscopic structure of the Nautilus shell. Bull. Inst. r . Sci. nut. Belg. 38, 1 7 1 . G R ~ G O I RC.E (1967). , Sur la structure des matrices organiques des coquilles des mollusques. Biol. Rew. 42,653-88. HALL,A. & KENNEDY, W. J. (1967). Aragonite in fossils. Proc. R . SOC.B 168, 377-412. HARE,P. E. (1963). Amino acids in the proteins from aragonite and calcite in the shells of Mytilus californianus. Science, N . Y . 139, 216-7. HARE, P. E. & ABELSON, P. H. (1964). Proteins in mollusk shells. Rep. Dir. geophys. Lab. Curnegie Instn 63, 267-70. HARE, P. E. & ABELSON, P. H. (1965). Amino acid composition of some calcified proteins. Rep. Dir. geophys. Lab. Carnegie Instn 64, 223-32. HUBENDICK, B. (1958). On the molluscan adhesive epithelium. Ark. Zool. (2nd ser.) 2, 31-6. HUDSON, J. D. (1968). The microstructure and mineralogy of the shell of a Jurassic mytilid (Bivalvia). Pulaeontology. 11, 163-82. S. & IKEMOTO, N. (1962). Electron microscopy on the mantle of a bivalve, Fubulinu nitudula. KAWAGUTI, Biol. J . Okayama Univ. 8, 21-30. KENNEDY, W. J. & HALL, A. (1967). The influence of organic matter on the preservation of aragonite in fossils. Proc. geol. SOC.1643, 253-5. W. J. & TAYLOR, J. D. (1968). Aragonite in rudists. Proc. geol. SOC.645, 323-31. KENNEDY, KESSEL, E. (1944). Uber Periostracum-Bildung. Z. Morph dkol. Tiere. 40, 348-60. KLASSEN-NEKLYUDOVA, M. V. (1964). Mechanical twinning of crystals. (Engl. transl. Consultants Bureau, New York) xiv+212 pp. KOBAYASHI, I. (1964). Microscopical observations on the shell structure of Bivalvia. Part I . Burbntiu obtusoides (Nyst). Sci. Rep. Tokyo Kyoiku Daig. sect. C 8, 295-31 I . KOBAYASHI, S. (1964). Studies on shell formation. A study of the proteins of the extrapallial fluid in some molluscan species. Biol. Bull. mar. biol. Lab. Woods Hole 126, 414-22. LOWENSTAM, H. A. ( 1 9 5 4 ~ )Environmental . relations of modification compositions of certain carbonatcsecreting marine invertebrates. Proc. natn. Acad. Sci. U . S . A . 40, 39-48. LOWENSTAM, H. A. (19546). Factors affecting the aragonite: calcite ratios in carbonate secreting marine organisms. J. Geol. 62, 284-322. LOWENSTAM, H. A. (1963). Biologic problems relating to the composition and diagenesis of sediments. In The Earth Sciences: Problems and Progress in Current Research, pp. 138-95. University of Chicago Press, Chicago. LOWENSTAM, H. A. (1964). Coexisting calcites and aragonites from skeletal carbonates of marine organisms and their strontium and magnesium contents. In Recent Researches in the Fields of Hydrosphere Atmosphere and Nuclear Geochemistry, pp. 373-404. Marnzen, Tokyo. MACCLINTOCK, C. (1963). Reclassificationof the gastropod Proscutum Fischer based on muscle scars and shell structure. J. Palaeont. 37, 141-56. MACCLINTOCK, C. (1967). Shell structure of patelloid and bellerophontoid gastropods (Mollusca). Bull. Peabody Mus. not. Hist. 22, 1-140. NEWELL, N. D. (1965). Classification of the Bivalvia. Am. Mus. Nowit. 2206, 1-25. OBERLING, J. J. (1964). Observations on some structural features of the pelecypod shell. Mitt. naturj. Ges. Bern. 20, 1-63. RANSON, G. (1952). Les huitres et le calcaire. Calcaire et substratum organique chez les mollusques et quelques autres invertbbres marins. C. r . hedb. Skunc. Acud. Sci. Paris, 234, 1485-7. M. (1951). Sur la composition des scleroproteines des ROCHE,J., RANSON,G. & EYSSERIC-LAFON, coquilles des mollusques. (conchiolines). C. r. Sianc SOC.Biol. 145, 1474-7. W. J. (1923). Bau und Bildung der Perlmuttermasse. 2002. Jb. Anat. Abt. 45, 1-48. SCHMIDT,

Bivalve shell mineraiogy

529

SCHMIDT, W. J. (1924). Die Bausteine des Tierkorpers inpolarisiertem Licht. 528 pp. Bonn. SCHMIDT, W. J. (1925). Bau und Bildung der Prismen in den Muschelschalen. Mikrokosmos, 18,49-54. SORBY, H. C. (1879). The structure and origin of limestones. Q.Jlgeol. SOC.Lond. 35,56-95. STEGEMANN, H. (1961). Uber die Skleroprotide der Mollusken. Naturwissei2schaften, 48, 5 0 1 . STEGEMANN, H. (1963). Proteine (Conchagene) und Chitin im Stutsgewebe von Tintenjischen. HoppeSeyler’s 2. physiol. Chem., 331, 269-79. STENZEL, H. B. (1962). Aragonite in the resilium of oysters. Science, N . Y . 136, I 121-2. STENZEL, H. B. (1963). Aragonite and calcite as constituents of adult oyster shells. Science, N . Y. 14, 23 2-3. STENZEL, H. B. (1964). Oysters: composition of the larval shell. Science, N . Y. 45, 155-6. TANAKA, S., HATANO, H. & ITASAKA, 0. (1960). Biochemical studies on pearl. IX. Amino acid composition of conchiolin in pearl and shell. Bull. chem. soc. Japan 33, 543-5. W. J. & HALL,A. (1969). The shell structure and mineralogy of the Bivalvia. TAYLOR, J. D., KENNEDY, Introduction; Nuculacea-Trigonacea. Bull. Br. Mus. tmt. Hist. Zool. Supplement 3 , 125 pp. THIELE, J. (1893). Beitrage zur Kenntnis der Mollusken. 11. Uber die Molluskenschale. 2. wiss. Zool. 55, 220-5 I. TOWE, K. M. & CIFELLI,R. (1967). Wall ultrastructure in the calcareous foraminifera: crystallographic aspects and a model for calcification. J. Palaeont. 41, 742-62. D. F., FRANCOIS, C. J., BONAR, L. C. & GLIMCHER, M. J. (1967). Comparative studies of the TRAVIS, organic matrices of invertebrate mineralised tissues. J. Ultrast. Res. 18, 519-50. TRUEMAN, E. R. (1949). The ligament of Tellina tenuis. Proc. zool. SOC.Lond. 119,717-42. TSUJII,T., SHARP,D. G. & WILBUR, K. M. (1968). Studies on shell formation. 1’11. The sub-microscopic structure of the shell of the oyster Crassostrea virginica. J. biophys. biochem. Cytol. 4, 275-9. WADA,K. (1961). Crustal growth of molluscan shells. Bull. natn. Pearl. Res. Kashikojima 7,703-828. WADA,K. ( 1 9 6 4 ~ )Studies . on mineralisation of the calcified tissue in molluscs. VII. Histological and histochemical studies of organic matrices in shells. Bull. natn. Pearl. Res. Lab., Kashikojima 9, 1079-86 (one table). WADA,K. (19646). Studies on the mineralisation of the calcified tissue in molluscs. X. Histochemical determination of the nature of acid mucopolysaccharide in organic crystals. Bull. Jap. SOC.scient. Fish. 30,993-8 (four figs.). WAELE, A. DE. (1930). Le sang d’dnodonta cygnea et la formation de la coquille. Mem. Acad. roy. Belg. Cl. Sci. 10 (fasc. 3), 1-51. WATABE, N. (1956). Dahllite identified as a constituent of Prodissoconch I of Pinctada martensi (Dunker)., Science, N . Y . 124,630. WATABE, N. (1965). Studies on shell formation. XI. Crystal matrix relationships in the inner layers of mollusk shells. J. Ultrastruct. Res. 12,351-70. WATABE, N. & WADA,K. (1956). On the shell structures of the Japanese pearl oyster Pinctada martensii (Dunker) Prismatic layer. I. Rep. Fac. Fish. prefect. Univ. Mie 2, 227-32. WATABE, E. N. & WILBUR, K. M. (1960). Influence of the organic matrix on crystal type in molluscs. Nature, Lond. 188, 334. WATABE, E. N. & WILBUR, K. M. (1961). Studies on shell formation. IX. An electron microscope study of crystal layer formation in the oyster. J . biophys. biochem. Cytol. 9,761-72. WILBUR, K. M. & WATABE, N. (1963). Experimental studies on calcification in molluscs and the alga Coccolithus huxleyi. Ann. N . Y . Acad. Sci. 109, 82-1 12. WILBUR, K. M. (1964). Shell formation and regeneration. In Physiology of Mollusca, ed. K. M. Wilbur and C. M. Yonge. Vol. I , 243-82. YONCE,C. M. (1953). Form and habit in Pinna carnea Gmelin. Phil. Trans. R . SOC.B,237,335-74. EXPLANATION O F PLATES PLATEI Fig. I . Scanning electron-micrograph of a fractured radial section of the sheet nacre layer of Neotrigonia pectinata. x 2800. Fig. 2. Radial thin section of the outer homogeneous layer of Arctica islandica. x 32. Fig. 3. Acetate peel of a planar section of the adductor myostracum of Chamu lazarus showing the irregular outlines of the myostracal prisms. x 100. Fig. 4. Scanning electron-micrograph of a fractured section of the lenticular nacre of Neotrigonia pectinata. x 1400. Fig. 5 . Thin section of the inner nacreous layer of Margaritifera margaritifera showing regeneration bands of periostracum and an associated prism layer. x 80. 34 Biol. Rev. 44

5 30


Fig. 6. Electron-micrograph of the inner surface of the sheet nacre layer of Neotrigonia dubia. x 4000. Two-stage Formvar replica, Au/Pd shadowed. Fig. 7. Electron-micrograph of the inner surface of the lenticular nacreous layer of Anodonta cygnea. x 4000.

PLATE2 Fig. I . Radial section of the periostracum, outer prismatic layer and middle nacreous layer of Murgaritifern marguritifera. x 80. Fig. 2. Acetate peel of a radial section of the crossed lamellar layer of Triduma squamosa. x 20. Fig. 3. Acetate peel of the outer calcitic, prismatic layer of Pinctada murgaritifera cut slightly oblique to the long axis of the prisms. Note the interprismatic conchiolin walls and growth bands. x 125. Fig. 4. Acetate peel of a radial section of the inner complex crossed-lamellar layer of Anadara antiquota. x 80. Fig. 5 . Inner margin of the periostracum of Anodonta cygneu showing initial calcification spherules. x

125.

Fig. 6. Acetate peel of radial section of the inner complex crossed-lamellar layer of Solenotellina biradiato. x 80. Fig. 7. Acetate peel of a transverse section of Nucula plucentina showing the outer composite prismatic layer (above) and the middle lenticular nacre layer. x 80. Fig. 8. Acetate peel of a radial section of the inner complex crossed-lamellar layer of Crossatella radiutu showing myostracal pillars. x 80. Fig. 9. Radial section (acetate peel) of the outer foliated layer of Spondylus calcifm. x 2 5 . Fig. 10. Radial section of Spondylus calcijer showing the foliated layer (bottom), middle crossedlamellar layer, pallial myostracum, inner crossed-lamellar layer and adductor myostracum. x 20.

PLATE3 Fig. I. Scanning electron-micrograph of a fractured section of the crossed-lamellar layer of Barbatia helblingi, showing parts of seven f i s t order lamels each built of parallel laths. x 1200. Fig. 2. Scanning electron-micrograph of a fractured section of the crossed-lamellar layer of Scutarcopagia scobinata showing the parallel component laths of a single first order lamel and enclosing organic matrix membranes. x 2000. Fig. 3. Electron-micrograph of the inner surface of the foliated layer of Ostrea hyotis. x 4000. Twostage Formvar replica Au/Pd, shadowed. Fig. 4. Scanning electron-micrograph of a radial fractured section of Unio welwichi showing prismatic pallial myostracum underlain by sheet nacre. x 2700. Fig. 5. Scanning electron-micrograph of a fractured section of the foliated layer of Ostrea edulis. Fig. 6 . Scanning electron-micrograph of the inner surface of the inner homogeneous layer of Thracia phuseolina. x 6500. PLATE 4 Fig. I. Acetate peel of a radial section of the outer shell layer of Gufrarium pectinaturn showing the transition from crossed-lamellar to homogeneous structure as typically developed in the Veneracea. x 80.

Fig. 2. Scanning electron-micrograph of the inner surface of the inner complex crossed-lamellar layer of Caryocorbula amethystinu. The whole surface consists of outcropping laths with several different attitudes. x I 160. Fig. 3. Aragonite fibres exposed on the fibrous layer of the resilium of Chlamys senatoriu, scanning electron-micrograph. x 330. Fig. 4. Scanning electron-micrograph of a radial section of inner complex crossed-lamellar layer of Pholas dactylus. x 600. Fig. 5 . Electron-micrograph of the inner surface of the adductor myostracum of Ostrea iridescem. Dark lines are organic membranes enclosing each myostracal prism. x 3000. Fig. 6. Scanning electron-micrograph of a fractured section of the prismatic layer of Unio welwichi. x 1400.

Biological Reviews, L b L . 44,No. 4

Plate

I

Biologicul Reviews, J,’o/. 44,No. 4

Plate 2

Biological Reviews, Lbl. 44, K O .4

Plate 3

Biological Rerims, Vol. 44,No. 4