Insights into the growth morphology of calcite cement

The classic work on the morphology of limestone calcite cements done in the 1960s is extended here by utilising growth zones to reconstruct the growth of cement crystals. Only cement composed of fitted polyhedral monocrystals that form by passive crystallisation of calcite on the walls of liquid‐filled, static pores and fissures is considered. Cement can either be initiated by (1) nucleation, when new crystals start but are not attached to their substrate, or (2) seeding, when new crystals are seamlessly connected to and influenced by substrate crystals. After seeding, epitaxial cement growth starts with many sub‐crystals that coalesce distally, followed by layered mantle growth. Junctions between three intercrystalline boundaries in cement aggregates with one interfacial angel = 180° are of two types: the first, enfacial junctions are caused by a pause in the growth of one crystal and the second is caused by movement of all boundaries due to dissolution of adjacent calcite. Growth zone offsetting at some intercrystalline boundaries is caused by dissolution of calcite at boundaries when permeability values are low. The same width to height ratio of mature aggregate crystals is predicted from the shape of the crystal's growth surfaces; dogtooth calcite forms columnar and nail‐head calcite forms tabular‐shaped crystals. Seeding on different sized crystals causes variations in epitaxial growth rate with faster growth on large crystals resulting in a disorganised cement fabric; the variation in epitaxial growth rate is perpetuated into mantle growth. Echinoderm syntaxial crystals dominate many pore cements due to the large size of their seed ossicles, at the same time, syntaxial crystals form on relatively tiny seeds. Texturally mature crystal aggregates with isopachous fabric are initiated from three different substrate to cement arrangements. Calcite cement zones preserve their original positions allowing the investigation of cement's growth and chemical history.


| INTRODUCTION
The thin section investigation of limestone cements started with Bathurst (1958Bathurst ( , 1975, Folk (1965) and Friedman (1964); their work is still widely quoted in textbooks on carbonate petrography (Flügel, 2004;James & Jones, 2015;Moore & Wade, 2013;Scholle & Ulmer-Scholle, 2003). This review is restricted to fitted polyhedral calcite monocrystals (Jones, 2017) that form sparry aggregates by passive precipitation of calcite crystals on the walls of liquid-filled, static pores and fissures. This type of cement is volumetrically important in shallow marine limestones, other cements, such as fibrous reef cements and cements precipitated in pores partially filled with air, are excluded.
Calcite cement crystals are characterised by both their intercrystalline boundaries, and internal zones revealed by staining, cathodoluminescence (CL) and charge contrast (CC) imaging. The CC is caused by areas of differing conductivity allowing subtle compositional variations and microstructural features to be seen on the scanning electron microscope (Buckman et al., 2016). Staining (Dickson, 1965(Dickson, , 1966(Dickson, , 2003Evamy, 1963) and CL (Sippel, 1965(Sippel, , 1968Sippel & Glover, 1965) were available to Bathurst and Folk but surprisingly were not included in their works. Zones in calcite cement were divided by Reeder (1991) into those that are coincident with growth surfaces (e.g. growth, oscillatory zones) and those that are not (e.g. sector and intrasectoral zones). Simple growth zones can be used to reconstruct the step-by-step growth of cement crystals and are extensively used in this review; some calcite cements with complex zonal types (Hendry & Marshall, 1991;Paquette et al., 1993;Paquette & Reeder, 1990) are not discussed.
The purpose of this paper is to describe and discuss some of the key factors affecting the shape and distribution of calcite cements in limestones and illustrate how they can be determined. Some of those key factors are: (1) style of cement initiation at pore walls (nucleation versus seeding), (2) early epitaxial cement growth changing to mantle growth, (3) boundaries between cement crystals, (4) crystallographic controls on crystal shapes and competition forming crystal aggregates, (5) recognition of breaks in crystal growth, (6) effects on variable cement growth rates and shape as cement crystals evolve from epitaxial to mantle growth, and (7) ultimate cement fabric within pores. Combining techniques shown with ideas presented will allow readers to better understand calcite cement growth histories in their limestones.

| CEMENT CRYSTAL INITIATION AND CONNECTION TO SUBSTRATE
Cements can start by either nucleation or seeding. Nucleation is here used for the birth of new crystals that occur on, but not connected to, a substrate; the two are separated by intercrystalline boundaries. Seeding is the process that seamlessly connects cement to substrate crystals; any substrate crystal can act as a seed when at least some of its crystal planes are continued into the cement. Some early marine cements are catalysed microbially leading to the precipitation of amorphous calcium carbonate (ACC) (Diaz et al., 2022); however, the cement discussed here is thought to precipitate entirely as crystalline calcite. The choice between nucleation and seeding is, primarily, controlled by fluid saturation; experimental precipitation of calcite shows seeding occurs at lower saturation states than nucleation (Tang et al., 2000). However, precipitation in the natural environment involves many factors, such as surface reactions of the precipitated phases, precipitation within cells rather than in ambient waters, foreign ions causing inhibition; these factors are still incompletely understood (Boon et al., 2020;Morse et al., 2007) and beyond the scope of this paper.

| Nucleation
Changes in cement fabric after nucleation are limited by the size of the pore relative to the spacing of nuclei on its wall. Following dense nucleation on the walls of a large pore, the morphology of the cement aggregate changes through three well-documented growth stages: isolated, competitive and parallel stages (Bathurst, 1975;Buckley, 1951;Dickson, 1983Dickson, , 1993Grigor'ev, 1965;Nollet et al., 2005;Rodriguez-Navarro & García-Ruiz, 2000;Schmidegg, 1928). If nuclei are widely spaced in small pores, cement fabric may not progress beyond the isolated stage.
All stages of morphological development are found in cement that grew from one wall of a fissure in Devonian limestone, exposed near Brixham, UK (Dickson, 2019;Hardman et al., 2020;Richter, 1966). The isolated and competitive stages are completed close to the wall followed by splays of elongate crystals that are terminated by Triassic sand filling the fissure's central space (Figure 1). The Devonian grey and purple wackestone forming the fissure wall contains stretched recrystallised fossils; the fissure wall is composed of calcite crystals a few to tens of microns in size. The initial cement crystals are all a few microns in size including those that cover the 30 μm wall crystals from which they are separated by intercrystalline boundaries.
Cement in the pores of a Permian dolomitised patch reef from Aberford, UK (Dickson, 1993(Dickson, , 2019Smith, 1981) show no fabric maturation. Transparent blocky millimetre-sized calcite crystals cover the pore walls composed of euhedral dolomite crystals (30 μm to 100 μm) and fractures through the dolomitised reef rock ( Figure 2). None of the cement crystals are in optical continuity with the dolomite wall rock in the plane of the thin section; this calcite cement did not seed on dolomite. Each calcite cement crystal has one nucleation point on the pore wall, the distance between these points is large relative to the size of the pores. Consequently, fabric maturation did not take place as the short distance between nucleation points and the pore's centre meant all the available pore space was filled before competition began. The growth of individual Aberford crystals is described in Section 3.2.

| Seeding
The nature of the contact between cement and seed crystals can vary. A perfect match occurs when calcite cement seeds on an older calcite cement. When cementation follows a pause in growth and the surface of the seed crystal ages (Abdullah & Gmira, 2014), cement often starts growing from multiple sites on each substrate crystal and with growth, the sub-crystals amalgamate into a single euhedral crystal. An example of such seeding is provided by the scalenohedral {2134} crystal from Llanharry mine (Gayer & Criddle, 1969) that was covered by a thin goethite precipitate allowing aligned isolated rhombohedral {1014} crystals to seed on the scalenohedron ( Figure 3A). The Miller-Bravais notation (Phillps, 2011) is used for calcite crystallographic forms throughout this review. Where the rhombs touch, they amalgamate into a single crystal mass; continued growth of these rhombs is anticipated to produce a single six-faced rhomb entombing the brown F I G U R E 1 Scanned rock slab cut normal to the fissure's wall, Devonian limestone, Shoalstone Point, Brixham, UK. Devonian limestone at base of figure. Calcite cement layer from one side of the fissure has yellow proximal and red distal parts. Isolated and competitive stages occur close to fissure wall. The centre of the fissure (top of figure) is composed of red Triassic sandstone.
F I G U R E 2 Thin section of blocky calcite cement filling pores in a dolomitised Permian patch reef, Cadeby Formation, Aberford, UK. The bryozoan reef framework has a fringe of euhedral dolomite crystals and is fractured to upper right. Calcite cement crystals are transparent. Pores in the dolomitised reef framework are filled with red epoxy.
F I G U R E 3 Scalenohedral crystals covered by brown goethite with seeded rhombohedral crystals. (A) Face to left edge covered with randomly oriented (nucleated) transparent rhombs; other faces have scattered oriented seeded rhombs. Rhombs on right face amalgamated into crystal mass. (B) Orange calcite rhombs seeded on the edges and corners of a brown scalenohedron. scalenohedron. The brown scalenohedron, shown on Figure 3B, has younger orange-coloured rhombs seeded only on the high surface energy edges and corners of the older crystal, a similar example is shown by Sunagawa (2007, fig. 7.10).
Sediment grains are often rounded by abrasion and biologically produced grains often have complex shapes, the surfaces of such grains rarely match the equilibrium euhedral shape of cement crystals. The space between the substrate and euhedral surface is occupied by cement that starts at many points on the substrates surface, the subcrystals so produced amalgamate with growth to form a single euhedral crystal. This transitional stage of cement growth between the rounded surface of a detrital quartz monocrystalline grain and its first euhedral surface has been called non-euhedral by Lander et al. (2008) and epitaxial by Wendler et al. (2016). The epitaxial and later euhedral growth stages are shown in a crinoid overgrowth from the Ronaldsway Formation, Isle of Man (Dickson et al., 1987) that grew into an exceptionally large pore ( Figure 4). The stereom (Smith, 1980) of the Ronaldsway crinoids and their earliest cement were composed of Mg calcite (11 mole% MgCO 3 ; Dickson, 2004) that is now a mixture of calcite with dolomite microcrystals Walker et al., 1989). Transformation of the Mg calcite has obscured the textural details of the original cement to crinoid contact.
Fortunately, the contact between calcite cement and an echinoid plate is well preserved in a Lower Cretaceous example from the Lekhwair Formation, Abu Dhabi. The external shape of the echinoid's labyrinthic stereom (Smith, 1980) is preserved ( Figure 5A). The skeleton, intergranular and intraskeletal cements all formed at different times but are part of one crystal as they all show the same green birefringence colour ( Figure 5B). The calcite cement is divided into four consecutive CL stages marking growth surfaces in the cement ( Figure 5C,D). The first stage is composed of orange luminescent triangles that were initiated at many points on the sides of the ossicle normal to its c-axis. The second stage has fewer curved, spire-shaped growths with their long axes parallel to the echinoid's c-axis. Some spires grew away from the skeleton's outer surface into intergranular pore space and some grew centripetally into intraskeletal pores ( Figure 5D).
The curved growth surfaces found in the early epitaxial cement seeded to echinoderm ossicles is due to the nature of its contact with the crystalline calcite cement. Modern echinoderm skeletons are secreted as amorphous calcium carbonate (ACC) that transforms into Mg calcite mesocrystals during the animal's lifetime (Donovan, 1991). These mesocrystals consist of similarly oriented Mg calcite nanocrystals surrounded by surface layers of ACC and/or organic macromolecules (Dubois, 2014;Seto et al., 2018). The complex structure of echinoderm crystals ensures seeded cement is initially full of defects inherited from the skeleton that are progressively eliminated throughout the epitaxial stage, as shown by the gradual simplification of the cement's growth surfaces. Echinoderm skeletons are not all the same, the magnesium content of modern echinoderms varies within skeletons and latitudinally (Weber, 1969(Weber, , 1973 and is also thought to change throughout the Phanerozoic (Dickson, 2002). Magnesium calcite stabilises within the sedimentary environments usually to calcite and dolomite microcrystals or sometimes just calcite (Dickson, 2003) cement can seed on these by-products at any stage of this transformation. Consequently, echinoderms and their overgrowths display many different textural relationships.
The complex biomineralisation involving ACC employed by echinoderms also applies to other biogenic crystals (Checa, 2018;Clark, 2020). These also can act as seeds, but, F I G U R E 4 Thin -section CL image of syntaxial crystals seeded on crinoid ossicles that grew into an exceptionally large pore in this skeletal grainstone, Lower Carboniferous Ronaldsway Formation, Skillicore, Isle of Man. Crinoids have bright CL spots that are remnants of its stereom structure. The crinoids and early Mg calcite cement are now composed of calcite with dolomite microcrystals (not seen at this magnification). Bright orange CL cement with irregular zigzag growth zones is characteristic of epitaxial growth. Late orange CL zones have flat well-defined growth zones. together with their overgrowths, they are so much smaller than echinoderm crystals that they are difficult to detect and often go unnoticed using the petrological microscope.
Two morphologically different epitaxial stages can be identified depending on the compatibility of the cement with its seed crystal. When the match is good, cement seeds at multiple points on an older cement's surface and the resulting sub-crystals have plane faces ( Figure 6A). When the match poor the initial sub-crystals have curved vicinal faces ( Figure 6B).
The terms syntaxial and epitaxial have different meanings in different disciplines; for instance, syntaxial is used by structural geologists for a type of vein (Bons et al., 2012) and different types of epitaxial growth are distinguished at the atomic level in crystal growth literature (Bachmann, 2001). Bathurst (1975) stated syntaxial is synonymous with epitaxial, however, a useful distinction was made by Lander et al. (2008) between non-euhedral and euhedral growth, non-euhedral growth was called epitaxial by Lander et al. (2008), and these terms are followed here ( Figure 6). Euhedral layers are commonly called concentric, but this does not apply when changes of crystallographic form occur, when all layers mantle the whole preceding layer, mantle growth is preferred here.

BOUNDARIES
Crystals in cement aggregates are separated from one another by intercrystalline boundaries. Three common intercrystalline boundaries, compromise, enfacial and substrate boundaries, form as cement crystals grow together. Substrate boundaries mark hiatuses in growth between nucleated crystals and their substrates. Enfacial junctions occur at the intersection of compromise and hiatus boundaries in single cement generations. Intercrystalline boundaries can be modified after their initial formation.

| Compromise boundaries
Compromise or impingement boundaries occur between crystals that simultaneously grow together, when adjacent F I G U R E 5 Thin-section images of echinoid plate and surrounding epitaxial cement, Lower Cretaceous Lekhwair Formation, Abu Dhabi. (A) Plane light image, echinoid stereom marked by black pores a few microns in size, clear areas are calcite cement. (B) Crossed polars image, stereom and cement are single crystal with the same green birefringence colour, intergranular cement has faint vertical birefringence stripes. (C) CL image of same area, echinoid stereom has speckled appearance due to pores luminescing yellow in a nonluminescent background, calcite cement divided into four consecutive stages on its CL zones. (D) Simplified sketch traced from Figure 5C; the stereom is coloured green, four cement growth stages are numbered, and the direction of cement growth shown by white arrows. crystals have polyhedral growth surfaces and impinge at a constant rate and angle, the boundary is planar. Their formation has been reconstructed graphically by Bathurst (1975); Buckley (1951); Dickson (1983Dickson ( , 1993; Grigor'ev (1965); Nollet et al. (2005); Rodriguez-Navarro and García-Ruiz (2000) and Schmidegg (1928). The six crystals shown on Figure 7A are separated by intercrystalline boundaries; the stained version of the same image ( Figure 7B) shows that the same flat growth surfaces (ferroan zones) in adjacent crystals meet at these boundaries confirming they are compromise boundaries. These intercrystalline boundaries should be planar, their irregularity was explained by Bathurst (1975) as due to the low angle of the boundary's intersection with the thin section's surface; such boundaries become planar when rotated into a vertical position using a universal stage.

| Substrate boundaries
The central calcite crystal shown on Figure 8A has a substrate boundary with its pore wall composed of euhedral dolomite crystals. This crystal's growth is reconstructed using its micron-sized yellow CL growth zones (numbered on Figure 8B). The crystal entered the thin section at a point shown by the blue asterisk on Figure 8B; its nucleation point is not present in the plane of this section. As the crystal expanded it met the substrate at the white asterisk from where it spread laterally to form the substrate boundary. The time hiatus between the cement and its dolomite substrate is several million years, according to Lee and Harwood (1989).

| Enfacial junctions
An enfacial junction is where three intercrystalline boundaries meet and where one of the three interfacial angles is 180° in a crystal aggregate. They form when one crystal stops growing forming a hiatus boundary and F I G U R E 6 Diagrammatic representation of the two stages of seeded growth in calcite cement crystals connected to different substrates. The crystals are sliced N/S parallel to their fastest growth direction. (A) An equant calcite cement rhomb {1014} seeded on the face of an acute rhomb {0118}; their lattices match, their contact is perfect. The new cement is seeded at two points on the old crystal's surface. The two sub-crystals formed have growth surfaces that are parallel to the faces of the final euhedral crystal. (B) An acute calcite cement rhomb {0118} growing on the surface of an echinoderm's skeleton (its porous structure is not shown); the contact surface in the crinoid is shown as a dotted line as its match with the calcite cement is imperfect. The new cement is seeded at four points; its sub-crystals have curved growth surfaces that morph towards the first euhedral surface.

F I G U R E 7
Thin-section images of a cement aggregate composed of six calcite crystals, Alston, U.K. (A) Plane light image, each crystal surrounded by compromise boundaries; cleavage lines are shown. (B) Stained version (alizarin red s and potassium ferricyanide stain) of part A. The intensity of alizarin red stain is correlated with calcite's crystallographic orientation and the potassium ferricyanide stain with ferrous iron concentration (Dickson, 2003). the two other crystals grow against the halted surface (Bathurst, 1964). The salient features of enfacial junctions are shown on Figure 9. Bathurst (1975) proposed a high percentage of enfacial junctions amongst triple junctions (30%-73%) as a criterion for the recognition of cement. However, enfacial junctions are absent from continuously growing cements (Dickson, 1983) and Bathurst overestimated their abundance as he failed to recognise a different type of triple junction that does not involve a break in growth (see Section 3.4).
Several enfacial junctions are present in the cement that fills pore 1 in the thin section from Abercriban quarry shown on Figure 10A (Searl, 1988). The cement has multiple CL zones that allow the establishment of seven growth stages ( Figure 10B). The enfacial junctions occur at the junction between growth stages D and E that marks a change from acute to equant in the cement crystal's habit ( Figure 10C,D). The younger stages (E to G) have a discordant contact with stage D marking the hiatus in growth. One of these enfacial junctions is enlarged as Figure 11A, its CL image ( Figure 11B) shows crystal 1 has a hiatus intercrystalline boundary that lies parallel to its growth zones. Crystals 2 and 3 stopped growing at the end of stage D but growth resumed with stages E and F at points on the D surface forming sub-crystals that spread diachronously across but join seamlessly with the D surface ( Figure 11C). This type of seeding is shown diagrammatically on Figure 6A. The hiatus surface at the end of stage D can be followed around the whole of pore 1 (pink line on Figure 10D). Younger calcite crystals either seeded seamlessly with stage D crystals or are separated from it by a hiatus intercrystalline boundary; all the enfacail junctions in this cement occur against this hiatus boundary. The patchy seeding behaviour of cement on this hiatus surface at the end of stage D may be due to its ageing (Abdullah & Gmira, 2014).

| Substrate boundaries
Substrate boundaries occur after nucleation as the growing crystal extends across the substrate eventually meeting growth from other adjacent nucleated crystals. Seeded cement crystals have no boundary with their substrate. The central calcite crystal shown on Figure 8A has a substrate boundary with its pore wall that is composed of euhedral dolomite crystals. This crystal's growth is reconstructed using its micron-sized yellow CL growth zones as time markers (numbered on Figure 8B). The crystal entered F I G U R E 9 Sketch of enfacial junction at red asterisk. The boundary of the old crystal, a hiatus boundary is coloured red and is met by a compromise boundary, coloured blue, that lies between two younger crystals. Black lines are growth surfaces, the red arrow shows the older crystal's growth direction, and the green arrows show the newer crystal's growth directions.
the thin section at a point shown by the blue asterisk on Figure 8B; its nucleation point is outside the plane of this section. As the crystal expanded it met the substrate at the white asterisk from where it spread laterally to form the substrate boundary. The time hiatus between this cement and its dolomite substrate is several million years according to Lee and Harwood (1989).

| Boundary modification
Two further types of boundaries exist that show offsetting of growth zones in crystals that once met at the intercrystalline boundaries. The first type involves minor offsetting shown on Figure 12A where the offsetting by a few microns (in two dimensions) can be reversed by opening the boundary as shown on Figure 12B. These minor offset boundaries affect randomly scattered boundaries in some cement aggregates. The offsetting is unlikely to be an original feature as this would involve randomly scattered irregular surfaces in a growth surface that otherwise was constructed only of euhedral crystal surfaces. Zone offsetting is also unlikely to be caused by movement along the boundary, as the two crystals shown on Figure 12A fit tightly together along an irregular boundary. The most likely explanation for offsetting is that calcite was dissolved along the boundary; dissolution did not penetrate the adjacent calcite crystals as their zones are sharply defined up to the boundary.
The second type of offset boundary occurs only at pore centres where groups of crystals have variable offsetting of their zones against crystal boundaries. An example from pore 2 in the Abercriban thin section ( Figure 13B) involves stage D cement composed of non-luminescent calcite with seven hairline yellow luminescent zones. The number of hairline zones present in adjacent crystals is variable, but the zonal successions have no missing zones. In plane light, these triple boundaries and enfacial junctions have one interfacial angle = 180°; they can only be distinguished when the arrangement of the crystals internal growth zones are revealed. The salient features of pore centre crystals are shown on Figure 14. The orientation of pore centre crystals is variable as that is fixed at the pore's margin and only those with their fastest growth direction oriented normal to the pore wall reach the pore's centre. Consequently, those that reach the centre are haphazardly oriented being derived from different parts of the pore wall. The dissolution behaviour of anisotropic calcite varies with its orientation (Meng et al., 2013;Smith et al., 2013) so pore centre crystals that are haphazardly oriented are variably dissolved accounting for adjacent crystals having variably offset zones.
The cause of severe offsetting at crystal boundaries is linked to their position at pore centres and not the age of the cement. In the thin section from Abercriban Quarry ( Figure 10) two pores exist, cement stage D occurs in both pore 1 and pore 2. The cement at the centre of pore 2 has seven severely offset hairline zones at crystal boundaries while in pore 1, stage D cement occurs halfway through its filling and has only two poorly developed hairline zones that show no offsetting at crystal boundaries.

| SHAPE OF AGGREGATE CRYSTALS
The shape of fitted crystals, no space between crystals, in cement aggregates can be simplified as cuboids: equant, acicular, columnar, bladed, tabular and platy (Dickson, 2019). These three-dimensional bodies, however, must be assessed in two dimensions for thin section studies. Many carbonate petrographers and textbooks (Scholle & Ulmer-Scholle, 2003, p.305) follow Folk's, 1965 scheme for thin sections (1965) and use just three shape categories -equant, bladed and fibrous -which is inadequate for the range of calcite crystal shapes known to exist (Dickson, 1983(Dickson, , 2019Kendall & Broughton, 1978).
Fitted crystals in cement aggregates are anhedral; crystallographic terms should not be used to describe their shape. Choquette and James (1987) and James and Jones (2015) use bladed-prismatic to describe the shape of cement aggregates that they qualify as 'elongate scalenohedral crystals having prismatic terminations'. However, acute scalenohedral crystals produce columnar not bladed crystals and prisms are open forms that cannot form terminations. Crystallographic terms are appropriate only for the cement's last growth surface that is preserved in incompletely filled pores. Acute euhedral growth surfaces on crystal aggregates are sometimes referred to as dogtooth calcite (Braithwaite & Montaggioni, 2009). Such calcite is then often unnecessarily qualified as scalenohedral (for example, Christ et al., 2012Christ et al., , 2015. Dogtooth calcite is commonly terminated by acute scalenohedra, but it can be terminated by acute rhombohedra and obtuse scalenohedra do not form dogtooth calcite. The shape of individual crystals in cement aggregates during isolated and competitive stages of growth are variable. Only after competition is over do the crystals all have similar shapes; they lengthen and have constant width to height ratios. The shape of these texturally mature crystals having past the competitive stage of fabric development is often given a single, collective name and their width to height ratios can be predicted if the crystallographic form of their growth zones can be determined. Three simple six-faced rhombohedra are used to demonstrate this prediction ( Figure 15).

| The link between growth zones and the shape of mature crystals
Three differently shaped calcite rhombohedra with their fastest growth directions oriented at right angles to the substrate are shown on Figure 15. Their growth rates normal to every face are all set at one, so their growth vectors can be compared.
The greatest growth vectors for the acute rhomb {101 1} is twice that of any other growth vector (Figure 15), giving this shape a competitive advantage that results in a narrow competitive stage. A horizontal cross-section through the centre of the acute rhomb has the shape of a regular hexagon and through its tip an equilateral triangle. Lengthening acute crystals with the same width and height produce columnar crystals with straight extinction that are length fast (Dickson, 1978(Dickson, , 1983. The greatest growth vectors for the equant rhomb {101 4} lie in its acute corners and are only slightly longer than vectors in its obtuse corners (Figure 15). Competition between differently oriented equant rhombs is weak resulting in a wide competitive stage. A horizontal crosssection through the centre of the equant rhomb has the shape of a slightly distorted hexagon and through its tip an isosceles triangle. Lengthening equant crystals with approximately the same width and height produce approximately columnar shapes with oblique extinction, varying 28° about the vertical (Figure 15) that are approximately length slow.
The greatest growth vectors for the obtuse rhomb {0118} is almost three times longer than any other vector (Figure 15) resulting in a narrow competitive stage. A horizontal cross-section through the centre of an obtuse rhomb has the shape of an irregular flattened hexagon and through its tip an isosceles triangle. Lengthening obtuse crystals sliced N-S parallel to its greatest growth vectors will be three times as wide as an E-W slice parallel to its greatest growth vectors. These crystals have tabular shapes with oblique extinction, varying 14° about the vertical that are approximately length slow.

| A columnar aggregate
The cement from the Brixham fissure (Figure 1) shows narrow isolated and competitive stages, followed by wide splays of elongate crystals ( Figure 16A). The texturally mature crystals are columnar with equal width to height ( Figure 16B). A prominent non-ferroan/ferroan contact on Figure 16A, and multiple CL zones on the longitudinal section on Figure 16C, show an asymmetric arrangement of its apical zone's that are characteristic of a N-S slice parallel to the c-axis of an acute positive rhomb. A slightly tilted basal section through the Brixham calcite has CL zones approximately in the shape of an isosceles triangle ( Figure 16D). These features are characteristic of columnar-shaped aggregate F I G U R E 1 4 Sketch of triple junction at the meeting of three crystals 1, 2 and 3 with one interfacial angle = 180° present at pore centres. Numbered black lines are growth zones. Yellow arrows show direction of growth. Intercrystalline boundaries are thicker black lines with yellow fringes.

F I G U R E 1 5
Orthographic projection of three calcite rhombohedra placed with their greatest growth vectors oriented vertically; growth rate normal to all crystal faces is set at one. Acute rhomb {1011}, red and black striped arrows mark the fastest growth directions that is coincident with the c-axis. Purple arrows are growth vectors in obtuse corners. Equant rhomb {1014}, red arrow marks six fastest growth vectors. Black arrow marks c-axis. Obtuse rhomb {0118}, red arrow marks six fastest growth vectors. Black arrow marks c-axis. Data from Dickson, 1983. crystals with growth surfaces formed from acute crystallographic forms.

| A tabular aggregate
Calcite cement collected from an incompletely cemented fracture in Cretaceous Chalk from Selwicks Bay, Flamborough, UK (Faÿ-Gomord et al., 2018;Roberts et al., 2020;Sagi et al., 2016) shows euhedral crystals formed of obtuse rhombohedra {0118} on its final growth surface ( Figure 17A). A thin section cut normal to this fracture wall ( Figure 17B) shows the cement crystals are elongate normal to the wall. The apical angles of two central crystals are shown on Figure 17B. The wide crystal has an apical angle of 147° corresponding to a slice through an obtuse rhomb parallel to its greatest growth vectors and normal to its c-axis. The narrow crystal has an apical angle of 51° corresponding to a slice through an obtuse rhomb parallel to both its greatest growth vectors and its c-axis. No growth zones were detected within these crystals; it is assumed that they grew entirely from obtuse rhombohedra accounting for the tabular shape of the crystals.

| BREAKS IN GROWTH
Different episodes or generations of cement are commonly separated by obvious lines of substrate boundaries. Pauses in precipitation within cement episodes are marked by discordant growth zones and enfacial junctions. Part or all previously deposited cement can be removed by dissolution as described by Braithwaite and Heath (1989), Dorobek (1987), Horbury and Adams (1989), Meyers (1978), Pedone et al. (1994) and Walkden and Berry (1984).
The terminal part of crystals filling the Brixham fissure ( Figure 1) show both obvious and subtle evidence of dissolution. Obvious cavities distributed along the last few growth zones in the host crystal are emphasised in plane light because they contain quartz sand and fine sediment; when they were open, Triassic sediment that filled the centre of the fissure, penetrated these cavities ( Figure 18A). These cavities were cemented by later calcite that grew in optical continuity with the host calcite and formed the intergranular cement to the overlying sand; they all extinguish together ( Figure 18B). However, dissolution penetrated deeper into the Brixham crystals F I G U R E 1 6 Thin-section images of same sample as used for Figure 1, Brixham cement, U.K. (A) Section stained, proximal calcite pink, distal calcite purple (ferroan calcite). Zonal contact between two calcites and one thin ferroan zone present. Dotted line is position of Figure  16B cut parallel to substrate. (B) Thinsection parallel to substrate, crystals have approximately the same height and width. (C) CL image from centre of Figure 16A, multiple zones invisible by staining. Apical angles measured against growth vector -white arrow. (D) CL image of section cut parallel to substrate; calcite crystals c-axes approximately normal to surface. Narrow dark sector zones along contact between main rhombohedra faces. creating a network of narrower cavities that are invisible in plane light ( Figure 18A). This network of cavities luminesces bright orange; they cut across the fine orange luminescing zones present in the host calcite ( Figure 18C).
The continuity of epitaxial growth zones can be misleading as they change around individual seed crystals. This is shown in cement overgrowths on two adjacent crinoid ossicles from a Lower Canyon (Upper Carboniferous) grainstone from Andrews, Texas, USA (Saller et al., 1999). The cement overgrowths on both crinoids 1 and 2 ( Figure 19) are divided into four stages by their CL zones although only the dominantly non-luminescent stage 1 is important here (Figure 19B,C). Around crinoid 1, three hairline orange luminescent zones are present in stage 1 cement; the first (1a) is irregular and close to the surface of crinoid 1, the second (1b) and third (1c) have the shape of plane-faced, approximate hexagons ( Figure 19B). On crinoid 2, stage 1 cement is wider than on crinoid 1 and zones 1a, 1b and 1c are irregular and more complex; 1a has triangular shapes and 1b and 1c have double luminescent zones. On crinoid 1, epitaxial growth was complete before the second hairline zone (1b) was precipitated as it has a euhedral shape characteristic of mantle growth. Whereas in crinoid 2, all three zones 1a, 1b and 1c have irregular zigzag shapes characteristics of epitaxial growth. The different shapes of the same zones on these two crinoids are due to epitaxial growth lasting longer on larger seed crinoid.

| RELATIVE GROWTH RATES DURING EPITAXIAL AND MANTLE STAGES
Epitaxial growth rate in quartz overgrowths was found by Lander et al. (2008) to be at its fastest immediately after seeding, it then slows until the first plane euhedral surface forms. They also found that during epitaxial growth, its rate slowed faster in cement seeded on smaller crystals and was completed before that seeded on larger crystals. Euhedral mantle growth rate varies with crystallographic form.

| Epitaxial growth
The Pwll-y-Cwm Oolite from Baltic quarry, south Wales (Searl, 1988) demonstrates how the cement's morphology is controlled by varying epitaxial growth rates due to its sediment being composed of millimetre-sized crinoids, bivalve crystals ten's of microns across, and micrite a few microns in size that coat some crinoids and form the cortex of some ooids ( Figure 20A). The area of epitaxial growth seeded on each substrate crystal can be predicted from the length of the seed crystal's surface exposed on the pore walls and the shape of the cement's first euhedral surface. The Baltic quarry cement is divided into stages using its CL zones and the cement's first acute euhedral surfaces is seen in the nonluminescent zone 2 cement seeded on a bivalve shell ( Figure 20B). The predicted areas of epitaxial cement are shown on Figure 20C. Stage 1 cement seeds only on the exposed surface of the crinoid in the upper pore, stage 2 continues the crinoid's growth and seeds on all other crystals exposed in the pore walls. The crinoid's overgrowth fills most of the upper pore, it meets cement growth from a micritic ooid at the top of this pore, they meet at the same stage of growth and are separated by compromise boundaries (Figure 20D). The zones in the crinoid's cement are much wider than the same zones in the micrite's cement because the former is seeded on a larger crystal that produces a greater volume of epitaxial cement than on the smaller micrite.
The cement fabric in the lower left and right pores is different from the upper pore because only bivalve and micrite crystals are present in their pore walls. Epitaxial cement seeded on the micrite is predicted to extend only 6 μm and that on the bivalve crystals 48 μm into the pores ( Figure 20C). These cement overgrowths, however, extend 20 μm and 60 μm into the pores indicating the change to mantle growth had occurred as indicated by the euhedral shape of zones 3 and 4 and their terminal surfaces. The centre of the two lower pores is filled with single large crystals that have the same extinction and twin plane orientation as the crinoid and its overgrowth. A hypothetical stage 5 cement grew on the crinoid's syntaxial overgrowth, out of the thin section's plane and when reaching stage 6 entered and filled the centre of the lower pores. Details of the bivalve and micrite overgrowths in the lower left pore are shown on Figure 21. Each bivalve crystal and its cement have the same birefringence colour and twin planes, adjacent crystals are different ( Figure 21B). The dominantly non-luminescent stage 2 cement seeded on bivalve crystals has discontinuous spire-shaped internal growth zones and a zigzag terminal surface, both features characteristic of epitaxial growth ( Figure 21A). The distribution of components is shown on Figure 21C.
The Baltic Quarry cement does not show the usual isolated, competitive and parallel stages of morphological development as its cement distribution is controlled by variation in the epitaxial growth rates of cement that grows on differently sized seeds.

| Mantle Growth
Following the formation of different sized epitaxial crystals, mantle growth continues to grow faster on larger crystals. This is shown by a skeletal peloidal packstone from the Isle of Man; its cement has five growth stages defined by staining and numbered for two crystals on Figure 22. The two numbered crystals grew towards the centre of a large pore, its roof and floor formed of disarticulated brachiopod shells. One crystal, seeded on the chonetid roof grew down to meet the second crystal growing up and seeded on an echinoderm plate on the pore's floor ( Figure 22). The first non-ferroan zone grew epitaxially forming a larger crystal on the echinoderm than on the smaller brachiopod seed. The terminal surfaces of zones 2, 3 and 4 are euhedral indicating mantle growth had taken over. The non-ferroan zone 4 is well defined and is much wider on the echinoderm syntaxial than on the chonetid syntaxial overgrowth showing mantle growth is faster on larger, than smaller, existing crystals. The euhedral shape of zone 4 in the echinoderm syntaxial is composed of segments (crystallographic forms) of different thickness showing the rate of mantle growth is also controlled by crystallography.
Calcite crystals are often formed from more than one crystallographic form; a common combination is the open primary prism {1010} and the closed secondary equant rhombohedron {1014}. These forms can form crystals of different overall shape or habit depending on which form dominates ( Figure 23A,B). The relative growth rate between these forms can change as shown in the cross section through a crystal with eight growth increments on Figure 23C. The growth rate of the rhomb remains constant while that of the prism increases after the fourth increment. Continued growth of this crystal would result in the disappearance of the prism. This elimination of the faster growing forms has led to the suggestion that the last form to appear at the centre of a pore will be the slowest growing form (Sunagawa, 2007). This geometrical control, however, is overridden at any time by a change in the precipitation driving force and reaction kinetics, both functions of the fluid's supersaturation (Prywer, 2002). The change in growth rate shown at the fourth increment on Figure 23C is accompanied by a change in the orientation and magnitude of the crystal's greatest growth vectors; these control the morphological development of cement aggregates.

| GROWTH FABRICS
The categorisation of cement into growth fabrics such as syntaxial, isopachous, and so on (Flügel, 2004;James & Jones, 2015;Scholle & Ulmer-Scholle, 2003), encourages the view that each cement fabric precipitates as a distinct episode separate from other cements. This could be assumed for the ooid grainstone from the F I G U R E 1 9 Thin section of Lower Carboniferous skeletal peloidal grainstone, Lower Canyon, Andrews, Texas, USA. (A) Plane transmitted light image, crinoids 1 and 2 turbid grey, micritic grains recrystallised, and primary intergranular and secondary intragranular pores filled with plastic, coloured brown. (B) CL image of crinoid 1. Basal section through round crinoid ossicle. Four numbered stages defined by luminescence colours. Stage 1 non-luminescent calcite has three hairline yellow luminescent zones (1a, 1b and 1c). (C) CL image of crinoid 2. Syntaxial cement seeded to upper surface of crinoid almost fills intergranular pore. Early orange CL growth 1a has surface with many triangles. 1b and 1c are double zones of irregular shape.
Lekhwair Formation, Abu Dhabi whose echinoid overgrowths dominate its pores due to their large crystal size. However, relatively small crystals also occur on micrite ooids ( Figure 24). As the large echinoid overgrowth grew quickly covering the small crystals on a nearby ooid and took longer to cover those on ooids further away. The timing of the echinoid syntaxial growth means the syntaxial crystals grew longer producing larger crystals than those on ooids close to the echinoid ( Figure 24A). This relationship indicates the echinoid and micrite cements grew at the same time.
The term isopachous is applied to elongate crystals of equal length often forming fringes 20 μm to 100 μm thick around intragranular-and intergranular pores (Christ et al., 2015;Hird & Tucker, 1988;Marshall & Ashton, 1980;Tucker & Wright, 1990). Isopachous fabric can arise in different ways depending on how its crystals are initiated (Figure 25). If the cement nucleates on equally spaced, randomly oriented seed crystals, the cement passes through the isolated and competitive growth stages before becoming isopachous ( Figure 25A,B). If the cement seeds on equally spaced, parallel-aligned substrate crystals, it adopts the substrates' structure becoming isopachous immediately without the isolated and competitive stages ( Figure 25C). This last option commonly occurs when cement seeds to bivalves with calcite layers (Figure 21), brachiopods (Figure 22), ostracods (Folk, 1962), and to ooids. The last case is shown by cement surrounding a well-preserved ooid from a Smackover grainstone, Arkansas, USA (Heydari & Moore, 1994: figure 26). The ooid's cortex is composed of groups of elongate, radially oriented crystals interspersed with clusters of equant crystals; (Figure 26A). Cement seeds onto the ooid's cortex and continues its structure to form a radial isopachous fringe. The orientation of the substrate and cement crystals is revealed in more detail using electron backscatter diffraction (EBSD). The radial orientation F I G U R E 2 0 Thin section of skeletal ooid grainstone, Carboniferous, Pwlly-Cwm Oolite, Baltic Quarry, Wales. (A) Plane transmitted light image of intergranular pores, 1, 2 and 3 between crinoid ossicle, ooids and bivalve shell layer rhombohedron. (B) CL image, same area as Part A. Crinoid surface partly covered by micrite. Intergranular cement with orange and non-luminescent zones divided into numbered stages 1 to 4 and 6. (C) Areas of epitaxial growth calculated from the length of the crinoid, bivalve and micrite crystals exposed in the pore wall and the shape of the cement's first Wulff surface sliced vertically E-W through the {0118} rhombohedron. (D) Sketch traced from Part B. Crinoid ossicle coloured pink with its syntaxial cement coloured yellow, its growth direction in pore 1 shown by arrows. Second crinoid (not present in this image) has syntaxial overgrowth that occurs at top right of image uncoloured.
of all the crystals in three segments marked by black lines on Figure 26B is clear when plotted stereographically ( Figure 26D,E,F). The crystals c-axes are radially oriented, and the azimuth of their a-axes are plotted showing little dispersion along a great circle, which applies to both the elongate and equant crystals despite their difference in shape.
The Baltic quarry grainstone (Figures 20 and 21) contains a crinoid syntaxial cement and isopachous cement within a few millimetres of each other. They have growth zones that indicate they grew at the same time. This option of synchronous growth for crinoid syntaxial and isopachous cements was not considered by Wood (2017, 2020) when modelling the evolution of a rock's porosity and permeability.

QUESTIONS AND RESEARCH DIRECTIONS IN CALCITE CEMENT GROWTH
To study the most complete record of cementation in a rock its largest pore should be sought. This can be illustrated by the skeletal peloidal grainstone from Abercriban quarry (Figure 10), which has pores that vary from a few micrometres (between peloids) to a few centimetres (beneath bivalve shells). The growth zoning shows the largest pore preserves the most cement stages, however, even younger stages might occur in larger pores that were not sampled by this thin section. F I G U R E 2 2 Stained thin section of Lower Carboniferous grainstone from the Skillicore Formation, Isle of Man. Sediment composed of peloids and micrite geopetally distributed within a gastropod shell and a productid brachiopod shell. Calcite cement divided into five non-ferroan and ferroan growth zones on one crystal growing down from a chonetid shell and up from an echinoderm plate into the shelter cavity.
The carbonate for cement growth within a pore is commonly supplied from outside the pore, and transport of dissolved carbonate to the pore's interior becomes progressively restricted as cementation proceeds. At first while transmissivity is high, the pore walls become lined with cement crystals but with continued cement growth pore throats become closed and permeability is reduced until a threshold is reached (van Der Land et al., 2013). After this threshold, the central pore can still be connected along sheet pores in the intercrystalline boundaries of previously deposited cement. The fluid in these nanometre-sized boundary pores can be undersaturated with respect to calcite, which dissolves, while the fluid in the large residual pore becomes supersaturated and calcite precipitates. Such pore-size-controlled-solubility has been called on to effect late-stage cementation in sediments by Ague (2009), Emmanuel andBerkowitz (2007), Emmanuel et al. (2010), Putnis (2015), Putnis et al. (1995), and Putnis and Mauthe (2001), and the preferential cementation of macropores in microporous limestones by Ehrenberg and Walderhaug (2015). Sometime after the permeability threshold is reached, the intercrystalline boundaries may be closed and the fluid-filled, crystal-lined residual pore is inactive, such pores (vugs) are found today in sediment deposited many millions of years ago that are now buried several kilometres below the earth's surface.
Morphological evidence for the dissolution along intercrystalline boundaries between cement crystals comes from the offsetting of zones described in Section 3.4. Rocks with partially cemented pores that have not reached the permeability threshold should have no offset boundaries. The extreme offsetting that occurs in pore centre cements (pore 2, Figure 10) is associated with the last most-restricted cementation in a pore. Fitted pore centre crystals with severely modified boundaries are problematic, the volume of calcite lost needs to be accommodated in the surrounding material, but none exists; further research is needed on this.
Growth zones, some only 1 μm wide, in calcite cement crystals match known crystallographic forms that are interpreted as being preserved from the time of their precipitation. The distribution particularly of Fe 2+ and Mn 2+ used to identify these zones and, by implication, some other, but not all (Cherniak, 1998) components, have retained their original positions providing ideal material to investigate changes in the chemical history of cementation. Samples can be obtained assuming the youngest cement lies at pore's centres. But, as shown in this paper, the distribution of calcite zones is complex, and any sampling should be undertaken after the distribution of these zones has been carefully mapped.
The identity of crystallographic form/habit is rarely reported in cement studies, yet it has the potential to provide information about the precipitation environment that cannot be found in any other way. Crystallographic form/habit can be seen in three-dimensions using the SEM on aggregate cement crystal's last growth surface in incompletely filled pores ( Figure 27A) and on zoned thin sections, such as that through the calcite cement crystal from the Aberford quarry ( Figure 27B). Midway through the growth of the Aberford crystals, three prominent sector zones appear; one prominent growth zone in these sectors is extrapolated and their interfacial angles measured. These correspond to a slightly tilted section through an acute scalenohedron {2134} (Dickson, 2019). To identify crystallographic form the crystal's interfacial angles are matched with published data (Palache et al., 1951), which is facilitated if the section is oriented in the same orientation as the published data or they can be corrected stereographically (Dickson, 1983). Milodowski et al. (1998Milodowski et al. ( , 2005Milodowski et al. ( , 2018) examined the connection between the crystallographic habit of calcite cement crystals and its ambient fluid in fractures at depths of +100 m to −1500 m ordnance datum, in the Sellafield area, UK. They found nail-head (c-axis flattened) crystals occur at the shallowest depths in fresh groundwater, equant crystals occur at intermediate depths in fresh to brackish groundwater, and dogtooth (c-axis elongated) crystals occur at deeper intervals in brackish to saline groundwater. Studies linking crystallographic habit with their coeval waters are rare, further studies are needed to build an inventory so the full potential of determining crystallographic properties of calcite cement can be achieved.

| CONCLUSIONS
Calcite cement morphology and development is diverse due to the many different factors affecting inception and growth. After nucleation, the fastest growing crystals survive to form an aggregate of similarly shaped elongate crystals providing the shape of their growth surfaces remains constant. This simple fabric maturation from a plane substrate has been categorised as isolated, competitive, and parallel stages by Buckley (1951), and shown in two-dimensional diagrams by Dickson (1983Dickson ( , 1993, Grigor'ev (1965), Nollet et al. (2005), and Rodriguez-Navarro and García-Ruiz (2000). The same fabric development occurs after seeding if cement is connected to a substrate composed of the same-sized, randomly oriented crystals. However, when the substrate is composed of the same sized, parallel oriented crystals, the cement has a well-organised noncompetitive fabric and when composed of different sized, randomly oriented crystals the cement has a disorganised fabric. Irrespective of the cement to substrate connection, if the shape of the crystals growth surfaces change, so does the direction and rate of the cement's fabric maturation.
Cement that seeds at sites on a substrate crystal forms many sub-crystals that with growth, amalgamate and simplify towards a single euhedral surface. The cement between the substrates surface and its first euhedral surface is the epitaxial stage which is followed by euhedral layers forming the mantle stage of cement growth. The shape of epitaxial growth surfaces changes with the type of substrate crystal. When cement seeds on cement, the sub-crystals have planar faces parallel to the first euhedral surface, but when cement seeds on complex biogenic crystals, often the most common seed in carbonate sediments, the sub-crystals initially have curved vicinal faces that change shape towards the first euhedral surface.
Growth zones have been extensively used here to reconstruct the step-by-step growth of cement crystals. They have been correlated across thin sections and are regarded as synchronous over such small distances.
Enfacial junctions are found in single cement generations when a break in growth occurs causing an interfacial angle of 180°. Enfacial junctions between cement crystals are triple junctions where a compromise boundary, separating two younger crystals, intersects a paused older crystal's boundary. Breaks in the growth that affect single cement generations are infrequent. Enfacial junctions may be distinguished from a second type of triple junction where both have one interfacial angle of 180°, using the crystal's growth zones. Enfacial junctions have one crystal with missing growth zones whereas in the second type of triple junction the zones are complete.
The position of intercrystalline boundaries is usually regarded as permanent but offsetting of adjacent growth zones in adjacent crystals shows they have moved due to dissolution of calcite along the boundary Minor offsetting occurs randomly distributed in some cements and severe offsetting is restricted to pore centre crystals. Further research on offset boundaries is needed.
Shape terms are collectively applied to texturally mature cement aggregates where all crystals have a similar shape. On reaching textural maturity, crystals have the same width to height ratios but lengthen into available space. The length to width ratio can be predicted from the crystallographic form of the crystal's growth zones. Two common crystallographic forms adopted by calcite crystals are the acute scalenohedron {2134} and the obtuse rhombohedron {0118} that give rise to columnar-and tabular-shaped crystals respectively.
Epitaxial growth on large seed crystals lasts longer and slows less quickly than on small seeds, consequently, large seeds have epitaxial growths of greater volume. This variation dominates the fabric of cement when it grows on a wide range of seed crystal sizes. Sediments containing echinoderms, whose crystals are usually an order of magnitude larger than any other seeds, have epitaxial growths that are so dominant that other epitaxial growths may go unrecognised.
The identification of crystallographic form in carbonate literature has generally been imprecise, yet the shape of specific forms determines the position of the crystal's growth vectors that in turn, control fabric maturation. Crystallographic form is also linked to the precipitation environment, features of which can be determined by the crystal's form/habit. This review covers the morphological development of only one type of cement found in carbonate rocks over the few square millimetres of a thin section. This cement, however, is a treasure house of empirical information often implanted episodically but it preserves this information in ancient examples over hundreds of millions of years over distances of many square kilometres.

ACKNOWLEDGEMENTS
I thank Peter Swart for his encouragement and Jay Gregg for suggestions that transformed my initial ramblings and to Art Saller who pointed out some incomprehensible sections in the revised manuscript and made them readable. Christopher Jeans provided the cement sample from fractures in the Cretaceous Chalk at Selwicks Bay, Flamborough, Yorkshire, UK. Thanks are also due to Iris F I G U R E 2 7 (A) Secondary electron image of cement's surface in an incompletely filled pore. Crystal's euhedral surfaces dominated by the equant rhombohedron, assumed to be the primary form {1014}, in combination with other forms including secondary equant rhombohedron {0114}. (B) Charge contrast image of thin section surface through an Aberford blocky calcite cement crystal. Three isolated sector zoned blocks occur halfway through zonal sequence. One prominent zone within these sectors extrapolated as black line to show alternate sharp and blunt interfacial angles characteristic of scalenohedra. The interfacial angles indicate this basal section is tilted slightly down towards the north.