Sedimentary Geology, 54 (1987) 93-112

93

Elsevier Science Publishers B.V., Amsterdam - Printed in The Netherlands

CELESTITE REPLACEMENTS OF EVAPORITES IN THE SALINA GROUP

ERNEST H. CARLSON

Department of Geology, Kent State University, Kent, OH 44242 (U.S.A.) (Received August 20, 1986; accepted for publication February 27, 1987)

ABSTRACT Carlson, E.H., 1987. Celestite replacements of evaporites in the Salina Group. Sediment. Geol., 54: 93-112. Replacements of evaporites by celestite were discovered recently at three sites in northwestern Ohio. These replacements are more durable than the original evaporites and provide new paleoenvironmental data for the upper Silurian rocks of the region. The occurrences are situated along the western margin of the Ohio (Cayugan) Basin and appear in the Greenfield Dolomite and in undifferentiated Salina dolostones. The replacements include: lenticular and prismatic crystals of gypsum, nodules of anhydrite and laminar evaporites. The lenticular crystals contain inclusions of carbonate and anhydrite, and are believed to have altered to anhydrite prior to replacement. The prismatic crystals are exceptionally well-preserved, with euhedral, deeply embayed outlines and internally zoned growth bands containing large numbers of inclusions of dolostone and anhydrite. Optical data for the latter crystals indicate that they are oriented replacements of gypsum, and suggest that the original gypsum was unchanged prior to replacement. The nodular and laminar occurrences display features such as chicken-wire and enterolithic structures, and were comprised of anlaydrite prior to replacement. Replacement postdates the dolomitization and cementation of the Salina sediments enclosing the evaporites, but occurred prior to deep burial. The rocks hosting the replacements, therefore, did not provide the strontium. The strontium may have been released from dolomitization of the underlying Lockport (Niagaran) beds or from dissolution of subaerially exposed Salina gypsum prior to the middle Devonian.

INTRODUCTION

In a replacement, as a general rule, a more soluble mineral will always be replaced by a less soluble mineral (Edwards, 1954). Since the solubility products for anhydrite, gypsum and celestite are 10 -4"6, 10 -4.8 and 10-6.5, respectively (Truesdell and Jones, 1974; Ball et al., 1980), celestite should replace both anhydrite and gypsum. Replacements of anhydrite and gypsum by celestite have not been reported previously in northwestern Ohio, but have been described elsewhere by West (1960, 1964, 1965, 1973), West et al. (1968), Frazier (1975), Nickless et al. (1975, 1976), Wood and Shaw (1976) and Wells et al. (1983). These workers and others generally 0037-0738/87/$03.50

© 1987 Elsevier Science Publishers B.V.

94

r N

Fig. 1. Map showing the location of the three quarry sites and the general geologic features of northwestern Ohio (rock contacts modified after Janssens, 1977). Inset shows the location of the study area in relation to the Ohio (Cayugan) Basin (after Ailing and Briggs, 1961). agree that inclusions of anhydrite or gypsum in celestite are indicative of replacement. One purpose of this paper is to describe some early diagenetic structures and textures of anhydrite and gypsum that have been preserved by replacement in rocks of the Salina Group. These celestite replacements are significant because they provide new paleoenvironmental and petrologic data in the area. Secondary goals are to evaluate the timing of replacement and possible sources of strontium. In this regard most workers believe that the celestite appearing in the evaporitic sediments of an intertidal environment is primary or early diagenetic (see, for example, Longman and Mench, 1978; Olaussen, 1981). The inset in Fig. 1 shows the location of the study area relative to the Cayugan paleogeography. The area is situated on the western fringe of the Ohio Basin, a restricted basin that is characterized by beds of halite in the central part, gypsum and anhydrite on the flanks and carbonate rocks on the surrounding shelf. The Cayugan rocks thicken eastward toward the center of this basin (Ailing and Briggs, 1961; Ultieg, 1964). The bedrock surface in the area of investigation consists of Niagaran and Cayugan strata that can be observed in several rock quarries (Figs. l and 2). The

95

WEST

FINDLAY FLANK

ARCH EAST F L A N K Devonian

Devonign

Boss Islands Dolomite

G F

Undifferentiated Q.

Salina

0

E

Q.

o

C c-

e--

B

Tymochtee Dolomite Greenfield

ILockport

Dolomife

(at)

Dolomite

A

O3

Lockpor t Dolomite

Fig. 2. Stratigraphic nomenclature for Silurian rocks, northwestern Ohio (modified after Janssens, 1977).

evaporites occur in the Salina Group, which is subdivided into the Greenfield Dolomite, Tymochtee Dolomite and undifferentiated Salina dolostones in exposures near the crest of the Findlay Arch. The maximum thicknesses of these units in the vicinity of the study area are 30, 37 and 130 m, respectively (Janssens, 1977). In the subsurface further east the rocks of the Salina Group are divided into units A through G. Replaced evaporites were observed at the following quarry sites (Fig. 1): the quarry of the Maumee Stone Company, southeast of Portage in section 7, Portage Township, central Wood County; the National Lime and Stone Company quarry, east of Carey in sections 9, 15 and 16, Crawford Township, northwestern Wyandot County; and the Charles H. McCarthy quarry, southeast of Upper Sandusky in section 15, Pitt Township, southeastern Wyandot County. Several specimens from each site were examined in thin section. Some of the thin sections were stained with alizarine red S and others were observed under cathodoluminescence as an aid in distinguishing the carbonates. The identity of the celestite was confirmed by X-ray diffraction studies. The evaporites investigated in this paper appeared as isolated crystals of gypsum, nodules of anhydrite, and laminar evaporites prior to replacement. These occurrences are similar to those of modern evaporites in the coastal sabkhas of the Persian Gulf (Shearman, 1978), and the structures and textures of the Cayugan replacements resemble the equivalent features of their modem analogs as well. ISOLATED CRYSTALS OF GYPSUM

The Greenfield Dolomite, Tymochtee Dolomite and undifferentiated Salina dolostones locally display a gash-like texture in widely scattered areas of western Ohio. Summerson (1966) demonstrated that this texture is due to crystals of

96 gypsum, or more commonly, molds of gypsum crystals. These crystals were replaced by celestite near Portage and Upper Sandusky, providing an opportunity to examine the original crystal morphology. The replaced crystals, which are distinguished by lenticular and prismatic habits, are described below.

Lenticular crystals Lens-shaped crystals of celestite were observed in the lower part of the Greenfield Dolomite near Portage, Wood County (Fig. 1). The celestite occupies slit-like openings in a fine-grained buff dolostone that is exposed along the northern part of the east face on the upper level. The crystals, which appear in a zone several centimeters thick, are oriented randomly in the rock and frequently form irregular intergrowths of two or three individuals (Figs. 3 and 4). Individual crystals are lenticular, with sections parallel to the direction of flattening being polygonal and those perpendicular to it being biconvex (Figs. 3 and 4). The crystals range from 2 to 4 m m across and have rough surfaces and ragged edges. The celestite is gray, cloudy, and dull, contrasting with the colorless, vitreous type that occurs at the other sites. Microscopically, the crystals contain inclusions of anhydrite, dolomite and dolostone. The inclusions of dolomite and dolostone are round grains that range between 25 and 50 # m in diameter, and their abundance and relatively even distribution are apparently responsible for the cloudiness of the celestite (Fig. 4).

Fig. 3. Group of lenticular crystals of celestite that were etched differentiallyfrom the dolostone matrix in 10% acetic acid. The scale is marked in millimeters. Near Portage.

97

Fig. 4. Photomicrographin cross-polarLTedlight of some lenticular crystals Of celestite that are at the extinct position, displaying different alignments in the rock. Individual crystals contain numerous inclusions of ardaydriteand carbonate. Near Portage. The inclusions of anhydrite, which are optically aligned in individual crystals, are anhedral and equant and average about 60/~m in diameter. The habit of the crystals suggests that they were originally gypsum, while the inclusions in them indicate that they consisted of anhydrite at the time of replacement. Crystals from the Portage quarry and modem individuals of gypsum from Laguna Madre, Texas, and the Persian Gulf (Masson, 1955; Kinsman, 1966; Shearman, 1966), have lenticular morphology and contain inclusions of rock matrix. Conversely, straight crystal outlines and rectangular re-entrants, which are characteristic of anhydrite, are absent (Shearman, 1966; Lucia, 1972), while inclusions of rock matrix, which are normally missing from anhydrite (Kinsman, 1966), are present. On the other hand, the celestite clearly replaced anhydrite as evidenced by the relatively large, corroded inclusions of the latter in the former. These data from the Portage site, thus, suggest that the crystals formed as gypsum, but had altered to anhydrite before strontium was introduced.

Prismatic crystals The prismatic crystals of celestite occur in the Tymochtee Dolomite just below the contact with undifferentiated Salina rocks, near Upper Sandusky, Wyandot County (Fig. 1). The crystals are embedded within a contorted layer of dark gray, shaly dolostone, which appears at the base of much thicker unit of light brownishgray, microcrystalline dolostone (Fig. 5). The celestite-bearing horizon is about 4 cm

98

Fig. 5. Cut specimen of Tymochtee Dolomite with a contorted layer of densely packed, prismatic crystals of celestite. The layer is 1 cm thick at its widest point. Near Upper Sandusky.

Fig. 6. Photomicrograph in cross-polarized light of a prismatic crystal of celestite at the extinct position. The left side of the photo shows a deeply embayed segment of the crystal. Note the concentric bands of inclusions. Near Upper Sandusky.

99

Fig. 7. Photomicrographin cross-polarizedlight of a prismaticcrystalof celestite.The view showsa long, parallelogram-shapedsection of a monocliniccrystal. Near Upper Sandusky.

thick and was observed on the north face of the quarry about 6 m below the top of the wall. The crystals are remarkably well-preserved, forming elongate, euhedral prisms that are typically about 4 mm long and 2 mm across and are often deeply embayed (Fig. 6). These prisms are apparently monoclinic, as they yield long cross-sections that are parallelogram-shaped and short sections that are diamondshaped (Fig. 7). The crystals, which are comprised of colorless and transparent celestite, are randomly aligned in the rock, densely packed and sometimes broken or bent (Fig. 8). Irregular intergrowths or twinned crystals are generally absent. Individual prisms display numerous inclusions of anhydrite, dolomite and dolostone in thin section. The anhydrite inclusions are unoriented, equant to cigar-shaped and unusually small, typically ranging between 5 and 10 g m in length. The dolomite and dolostone inclusions are round, sand-like grains, averaging about 50/~m in size. The anhydrite and carbonate are both concentrated in a series of concentric bands about 200 gm thick which parallel the crystal faces (Figs. 6 and 9). Similar growth patterns of anhydrite and carbonate inclusions occur in the gypsum crystals of modern coastal sabkhas and are believed to originate from the periodic influx of less saline water (Illing et al., 1965; Butler, 1970; Schreiber, 1978). The crystal morphology, the types of inclusions and the well-preserved features suggest that the crystals were originally comprised of gypsum and that they had not altered prior to replacement. Similar prisms of gypsum with embayed outlines and bands of inclusions occur in the Persian Gulf, as noted above. In addition, the celestite lacks the relatively large (e.g., 50 g m or more) and sometimes optically

100

Fig. 8. Photomicrograph in cross-polarized light of a prismatic crystal of celestite. The crystal is bent and displays undulatory extinction. Near Upper Sandusky.

ili~i~¸ ,,

Fig. 9. Photomicrograph in cross-polarized light of a prismatic crystal of celestite at the extinct position. The crystal contains concentric bands of relatively large, round carbonate inclusions and tiny anhydrite inclusions parallel to the crystal faces. Near Upper Sandusky.

101 aligned inclusions of anhydrite that are recorded at the Carey and Portage sites and elsewhere (West et al., 1968; Frazier, 1975; Wood and Shaw, 1976; Wells et al., 1983). It is probable, therefore, that celestite replaced the gypsum before alteration to anhydrite could occur. In this regard the optical orientation of the celestite agrees with the outlines of gypsum crystals in the rock, and evidence that celestite forms oriented replacements of gypsum is presented in a later section of this paper. Although relics of gypsum were not observed in the replacements, they have not been reported in the literature, either, suggesting that replacements of gypsum by celestite are always complete. NODULAR AND LAMINAROCCURRENCES Replaced evaporites also are found in nodular and laminar form. Two occurrences of these types are known in the area and both appear in the Greenfield Dolomite. Because nodular or laminar evaporites have not been reported previously in that formation, these deposits have a special paleoenvironmental significance. Primary textures in these replaced evaporites are largely destroyed, but some of the original structures are fairly well preserved.

Nodules of anhydrite Replacements of nodular evaporites in the Greenfield Dolomite occur near Carey in Wyandot County (Fig. 1). These nodules appear in a depositional-type breccia of tan dolostone at the base of the formation, which is separated disconformably from the underlying Lockport Dolomite (Niagaran) by green shale. The nodule-bearing bed is exposed 3 m below the top of the east face and extends the length of the east wall, a distance of 800 m, but does not appear elsewhere in the quarry due to an eastward dip. The nodular structure is preserved as molds of nodules which average about 3 cm across and appear through a thickness of about 1.5 m. Isolated molds are ellipsoidal in shape and slightly flattened parallel to bedding, while closely spaced, coalescing voids have knobby outlines. Clasts in the host bed adjacent to the molds appear to wrap around them (Fig. 10). In unweathered rock the rugs are occupied by interpenetrating networks of celestite crystals (Fig. 10). Individual crystals are euhedral with uncorroded surfaces, displaying the characteristic tabular habit of celestite. They are about 3 m m long and 0.75 mm thick, colorless to pale bluish-white and transparent to translucent. Most of the plates are not attached to the walls and do not radiate inward from them, indicating that crystal growth originated from within the nodules. Microscopically, the celestite crystals are distinguished by cores that are rich in anhydrite inclusions and rims that are inclusion-free. The anhydrite inclusions are corroded, rectangular- or lath-shaped, typically 50-100 #m in length and optically unaligned. No inclusions of carbonate rock were noted in the celestite. In weathered rock the

102

Fig. 10. Cut specimen of the basal Greenfield breccia showing a network of celestite crystals in a nodular mold. The clasts drape around the walls of the mold. Near Carey.

celestite has been removed and the walls of the molds are lined with small, fibrous balls of white strontianite. The presence of relatively large, corroded and unaligned inclusions of anhydrite suggests that the nodules were comprised of anhydrite at the time of replacement. The celestite in the molds is not pseudomorphous and other primary textures of the evaporites have not been preserved. However, the Carey replacements are free of included rock matrix, resembling in this regard the nodules of early diagenetic anhydrite that occur in m o d e m evaporitic sediments (Kinsman, 1966; Shearman, 1966; Butler, 1970). L a m i n a r occurrences

A bed of laminar evaporites that has been replaced with celestite occurs near Portage in Wood County (Fig. 1). The deposit is composed of alternating layers of celestite and dolostone, about 1.5 m thick, and located along the upper level of the south face. The lower contact is marked by prominent sulfur stains near the base of the Greenfield Dolomite (Fig. 11), which reportedly has a conformable contact with the underlying Lockport Dolomite at the site (Janssens, 1977). The bed can be traced along the east side of the south wall for a distance of about 200 m before it dips beneath the floor further west. Along the east face the same unit is a celestite-free, banded dolostone with knobby, sheet-like pores that are aligned

103

Fig. 11. View of the replaced bed of laminar evaporites in the basal Greenfield Dolomite on the upper level of the south face of the Maumee Stone Company quarry near Portage. The bed is recessed and is marked by the upper zone of light-colored, sulfur-rich seepage that is emerging from a bedding plane at its base.

parallel to the bedding. These channel-type openings, which apparently originated by the dissolution of the celestite, are partially occupied by crystals of post-replacement calcite. The laminar occurrence on the south face is comprised of an exceedingly dense rock that contains about 25% celestite by volume (Fig. 12). The celestite is found in thin layers of closely spaced or coalescing nodules and in contorted laminae, with nodules being the dominant form. Banded, tan, microcrystalline dolostones are intercalated with the celestite, with layers of both types ranging from 1 to 10 mm in thickness. Banding in the dolostone wraps around the thickened, bulbous masses of celestite in the layers which are flattened in the plane of bedding. The celestite is colorless and transparent and occurs in long, anhedral plates that are optically continuous along bedding for lengths of up to 8 cm. Pre-replacement features, such as chicken-wire and enterolithic structures, are sometimes preserved in the nodular layers and laminae (Fig. 12). Numerous relics of anhydrite indicate that the laminar evaporites consisted of anhydrite at the time of replacement. The anhydrite inclusions are unaligned optically, corroded and rectangular- to lath-shaped, averaging about 50 gm in length. In the nodular masses, the anhydrite is locally concentrated in parallel rows of inclusions that form grid-like networks. The edges of these growth patterns, which are up to 3 mm long and 1 mm across, resemble the outlines of crystals of

104

Fig. 12. Cut specimen of the basal Greenfield rock with alternating layers of dolostone (light gray) and celestite (dark gray with locally white cleavage traces). The black and white divisions on the scale are one centimeter long. Near Portage.

anhydrite because of their lath-like shape and the straightness of the rows of inclusions. Although it is not clear how individual inclusions of anhydrite are related to these larger outlines, the data suggest that the laminar evaporites were composed of anhydrite prior to replacement. However, the mineralogy of the original evaporites is not known. Dolostone inclusions also are quite common, being anhedral, delicately shaped and extremely variable in size. This carbonate is concentrated locally in bridge-like patterns of inclusions up to several millimeters in length. DISCUSSION

Oriented replacement Data obtained in the rocks near Upper Sandusky are believed to show that celestite forms an oriented replacement of gypsum. Oriented replacement is exhibited in the following textures: (1) outlines of euhedral gypsum crystals with different alignments in the rock; (2) rotated fragments of broken crystals; and (3) bent crystals. Changes in the optical orientation of celestite correspond to variations in the crystallographic orientation of gypsum in these rocks, while the bent crystals display undulatory extinction (Fig. 8). The exact optical alignment of celestite

105 relative to gypsum is not known, since relics of gypsum have not been recognized in the replacements. At the Upper Sandusky site, however, the celestite is always length-slow and shows extinction that is slightly inclined to the long edges of the monoclinic prisms. Conversely, there is no evidence to indicate that replacements of anhydrite by celestite are oriented. At the Portage site, for example, where celestite has replaced anhydrite, adjacent crystals of celestite display a common position of extinction in the rock even though the crystal outlines are randomly aligned (Fig. 4). Oriented replacement involves the replacement of one substance by another whose crystal structures and/or chemical compositions show a mutual correspondence (Dent Glasser et al., 1963). The general structural and chemical similarities of gypsum, anhydrite and celestite are self-evident, all being mineral species of the sulfate group. Although the details of the individual crystal structures are different, all three are characterized by sheets parallel to (010} that contain the sulfur of the sulfate groups and either the calcium of gypsum and anhydrite or the strontium of celestite (Bragg et al., 1965). The restriction of oriented replacement to the gypsum-celestite pair is probably due to the volume relationships in the lattices. The loss of water in gypsum would provide extra room and, therefore, would allow the larger ions of strontium to replace calcium. In the anhydrite lattice, on the other hand, replacement of calcium with strontium would require more space than is available, and oriented replacements would be discouraged from forming.

Timing of replacement The timing of the replacement, one of the most critical factors in evaluating possible sources of strontium, cannot be bracketed closely for the Salina evaporites. Evidence relating to the question of timing is somewhat diverse because of the varied character of the replacements in the area. In the discussion below, this evidence has been arranged under three phases of the diagenetic history as follows: dolomitization of the enclosing sediment, compaction and cementation of the host sediment, and depth of the evaporites below the surface.

Dolomitization The celestite replacements postdate the dolomitization of the sediment that encloses them. The evaporites shouldered aside this soft carbonate mud during a very early stage of diagenesis. Displacive growth of the evaporites is evidenced in the following features: (1) near Upper Sandusky by concentric growth zones of inclusions of dolostone sand in the replaced gypsum crystals; (2) near Carey by dolostone clasts that drape over the nodular molds of replaced anhydrite; and (3) near Portage by banded dolostone that wraps around bulbous masses of the replaced laminar evaporites. Inclusions of dolostone are present in all of the evaporites that were later replaced except for those near Carey, Dolomitization of the enclosing sediment, therefore, occurred penecontemporaneously with evaporite formation and prior to replacement.

106

Compaction and cementation In the nodular occurrences near Carey, replacement of anhydrite occurred after the enclosing sediment was lithified and at the same time that the evaporites were undergoing dissolution. Cementation of the sediment must have preceded dissolution and replacement of the evaporites, because the walls of the molds are devoid of collapse structures (Fig. 10). Concurrent dissolution and replacement during a post-lithification stage could presumably account for the open intergrowths of celestite and the patterns of inclusions in individual crystals occupying the nodular molds. Inclusions of anhydrite, which are concentrated in the cores of the crystals but are absent from the rims, indicate that the cores formed by replacement and the rims by continued growth in open space. Apparently, the initial replacement of some of the anhydrite was followed closely by dissolution of the rest of it, permitting the formation of euhedral overgrowths of inclusion-free celestite on the cores. Chowns and Elkins (1974) and Frazier (1975) observed similar patterns of inclusions in replacements of nodular evaporites. The bent and broken crystals of celestite near Upper Sandusky (Fig. 8) may have resulted from gypsum crystals that were deformed during compaction, and later replaced. Because celestite forms oriented replacements of gypsum, the evidence of this deformation would be preserved in the celestite. Alternatively, the celestite could have originated as a very early replacement of undeformed gypsum crystals, with compaction and the resulting deformation postdating replacement. This writer prefers the former interpretation, because bent and broken crystals of evaporites occur in the sediments of recent coastal sabkhas (Shearman, 1978), while replacement-type celestite apparently is not found in those deposits (see discussion on p. 108). The occurrence of bent celestite of replacement origin in ancient evaporitic sediments has been recorded elsewhere by West (1960, 1964) and Nickless et al. (1975). Although the last group of workers believe that the bending is indicative of pre-compaction replacement, the deformed crystals they observed may well be oriented replacements and as such would be subject to other interpretations, i.e., post-compaction replacement. Depth below the surface Differences in depth of burial at the time of replacement may explain why gypsum was replaced near Upper Sandusky and anhydrite was replaced at the Carey and Portage sites. According to Murray (1964), the replacement of gypsum by anhydrite due to burial diagenesis occurs at depths of a few hundred meters below the surface. Because the rocks near Upper Sandusky are higher stratigraphically (uppermost Tymochtee Dolomite) than those at the Carey and Portage localities (basal Greenfield Dolomite), the gypsum at the former site would have been shallower than the anhydrite at the latter localities. The thickness of the overlying Silurian section at Upper Sandusky is not known because those rocks have been removed by erosion, but surface and subsurface data in the vicinity (Janssens, 1977)

107 suggest that an additional 100-150 m of upper Silurian rocks were once present at the site. The occurrence of replaced gypsum near Upper Sandusky, therefore, indicates that replacement took place during a relatively early stage of burial diagenesis and possibly prior to the middle Devonian, when the overburden was still relatively thin. During a period of subaerial exposure, shallow evaporites not replaced previously and not sheltered by shale would have been dissolved and removed quickly by downward circulating meteoric waters. The unconformity separating upper Silurian and middle Devonian rocks is the oldest, regionally developed erosion surface that is associated with karst-like features in Cayugan rocks. This erosion surface has also beveled Cayugan beds to lower stratigraphic levels southward along the crest of the Findlay Arch (Summerson and Swann, 1970; Janssens, 1977). Solutional cavities occur at depths of 15-21 m below the top of the truncated Silurian section, but the effects of solution are believed to have extended to much deeper levels (Summerson and Swann, 1970; Summerson, 1970). Although depths to the Greenfield and Tymochtee evaporites from that unconformity are not known, unreplaced evaporites could have been removed by erosion at that time. Replacement, therefore, may possibly be bracketed in time between the dolomitization of the earliest Greenfield (upper Silurian) sediments and subaerial erosion in the post-Silurian and pre-middle Devonian interval. While dissolution and replacement of the nodules at the Carey site occurred simultaneously, as noted earlier, replacement of the laminar evaporites near Portage took place without the concomitant effects of dissolution. The features documenting the dominance of replacement over dissolution at the latter site include: (1) the well-preserved structures in the replacements; (2) the widespread distribution of inclusions in the celestite; (3) the anhedral shape of the celestite crystals; and (4) the absence of porosity in the laminated rock. The extent that replacement of the evaporites was accompanied by dissolution at the Carey and Portage sites may have been related to the depth of the Greenfield horizon at the time of replacement. Because of the beveled erosion surface, Greenfield rocks near Carey would have been closer to the surface, and the shallower evaporites there may have been subjected to higher flow velocities than the deeper ones at the Portage site. Sources of strontium

The strontium in the replacements could have been derived from a number of potential sources. Isotopic ratios of strontium, commonly used to establish the source of that element, have not been determined for replacement-type celestite. For cavity-type occurrences of celestite in nearby Lockport and Detroit River Group (middle Devonian) beds, however, 87Sr//86Sr ratios suggest that strontium was derived locally instead of from oil-field (basinal) brines or deep-seated fluids (Kessen et al., 1981). Because replacement- and cavity-type celestite are spatially

108 associated in the study area, it can probably be assumed safely that both kinds of celestite were derived from locally obtained supplies of strontium. Local sources of strontium that have been proposed for replacement-type celestite elsewhere include: (1) marine-derived brines (Frazier, 1975); (2) conversion of aragonite to calcite (Nickless et al., 1975); and (3) dolomitization of limestone (Wood and Shaw, 1976). Another local supply could be derived from the dissolution of subaerially exposed gypsum (Carlson, 1983, 1986). To be viable, however, the mechanisms of any of the above supplies must be able to account for two, unique features of the Salina replacements. These characteristics are: the large quantities of celestite that have been concentrated at the localities, e.g., Carey and Portage; and the close spatial association of the replacement-type celestite with celestite-filled cavities and fractures at the sites and the well-defined, northerly trend that these occurrences collectively form in the region (Carlson, 1983). In the light of these restrictions, the possible importance of the four sources of strontium above, is discussed below. The sediments of modern coastal sabkhas commonly exhibit a gypsum-celestite association, due to reactions with marine-derived brines (Evans and Shearman, 1964; Kinsman, 1966; Shearman, 1966). This evaporitic celestite occurs as small crystals that display typical celestite morphology and comprises somewhat less than one percent of the volume of the sediment. Early diagenetic exchanges between brines and the sediments of coastal sabkhas apparently do not liberate large quantities of strontium and do not favor the formation of replacements, since replacement-type celestite has not been recorded in that environment. Several formations in the study area could have released strontium from the conversions of aragonite to calcite during limestone diagenesis or calcite to dolomite during dolomitization. Regarding the time of dolomitization, most workers believe that fine-grained dolostones are formed during early diagenesis, while sparry dolostones are thought to originate at a later stage (see, for example, Jodry, 1969). The Greenfield and Tymochtee Dolomites, which are dolostones of the former type, were not the source beds because they were dolomitized before the evaporites that they enclose were replaced. The Lockport Dolomite, on the other hand, belongs to the latter group of dolostones and could have provided strontium for the replacements in spite of its stratigraphic position below them. Favoring this origin are the large amounts of strontium that would have been liberated from the Lockport sediments and the close proximity of the Lockport rocks to the replacements. The lack of a regional correlation between the distributions of the Lockport Dolomite and the replaced evaporites, however, would argue against this source, because the Lockport is dolomitized over a very broad area (Berry and Boucot, 1970) and the replaced evaporites are confined to a comparatively narrow belt. In that respect, the distribution of the replaced evaporites cannot be correlated with any other formation of carbonate rocks either. Dissolution of Salina gypsum could have provided a local source of strontium.

109 This strontium is believed to occupy the calcium sites in the gypsum lattice, and is enriched ten times over the content in the enclosing dolostones (Carlson, 1983; 1986). Removal of gypsum, however, would not have occurred on a large scale until the Salina beds were tilted and truncated in the interval between the upper Silurian and middle Devonian (Summerson and Swann, 1970; Janssens, 1977). Because of its relatively high solubility, this gypsum would have been removed selectively by meteoric waters, yielding downward circulating, strontium-rich fluids. The following characteristics of the Salina gypsum would favor it as a source of strontium: (1) the stratigraphic position of the eroded evaporites above the replacements; and (2) the close spatial correlation between the evaporites removed from the crest of the Findlay Arch and the replacement-type celestite. The transfer process is visualized as follows. Dissolution of gypsum in rocks at and near the surface would yield downward circulating waters that are saturated or nearly saturated with respect to Ca 2+ and SO~-, and enriched in Sr 2÷. Celestite, however, is less soluble than either gypsum or anhydrite and the resulting concentrations of Sr 2÷ and SO42- might exceed the solubility product of celestite. Therefore, if saturated waters came in contact with more gypsum or anhydrite at greater depth, replacement of these minerals by celestite could occur. In nearly saturated waters dissolution of gypsum or anhydrite, and partial replacement of them by celestite might occur simultaneously. The actual composition of the waters would be more complex than the above mechanism would suggest, since the waters would also be saturated with respect to the major and trace constituents of the other phases along the flow path, such as dolomite, calcite and the days. The strontium partitioned in those minerals would add to that held in gypsum in determining the total strontium content of the waters.

Geologic significance of replaced evaporites Replacements of gypsum or anhydrite by celestite last longer than the initial evaporites, due to the comparatively low solubility of celestite, and provide useful paleoenvironmental data that might otherwise be missing. The discovery of replaced evaporitic bodies in the basal Greenfield Dolomite, for example, probably indicates that similar occurrences of evaporites were widespread in that formation. Unfortunately, some of this celestite has been removed, leaving nodular molds and channel-type porosity in the place of the former evaporites. The replacements could provide a record of the original mineralogy of the evaporites and the mineralogical changes that might have occurred up to the time of replacement. The morphology of the replacements, for example, would be preserved from and indicative of the original evaporites. The presence of oriented replacements, on the other hand, would be determined by the mineralogy of the evaporites at the time of replacement. The replacements frequently contain inclusions that are genetically important. Inclusions of anhydrite and their patterns, for instance,

110 would be distinctive of the evaporites at the time of replacement, since replacement would have sheltered them from further alteration. Conversely, the carbonate inclusions would be unchanged from and diagnostic of the original evaporites. Under favorable conditions, therefore, the mineralogical character of the evaporites could be documented at both the time of formation and the time of replacement. The replacements in the study area are located near the western margin of the Ohio (Cayugan) Basin and two of the three sites are situated near the base of the Salina section. These findings are similar to those of West (1973) and Nickless et al. (1975) who had earlier recorded celestite replacements elsewhere at the margins of evaporitic sequences. They also noted that the replacements were closely associated with limestones at those boundary locations, and they postulated that the strontium for the celestite was liberated by the diagenesis of the bordering limestones. An alternative mechanism, however, could account for the replacements situated near the bottom of an evaporitic sequence, and would involve the dissolution of subaerially exposed beds of gypsum at the top of the section and the transfer of the released strontium downward, as proposed in this paper. ACKNOWLEDGEMENTS The writer gratefully acknowledges the assistance of the following quarry personnel: John Bearss and Max Goldacker of the Maumee Stone Company; George Wisti and Jim Kinsler of the National Lime and Stone Company; and Charles H. McCarthy of the McCarthy Quarry Company. The thin sections were made by Joe Francis and Tedd Ronning, the photographs of the hand specimens were taken by Professor Rodney Feldmann and Dale Tshudy and the drafting was done by Karen Taylor. Dr. Neil Wells assisted with the operation of the Luminoscope. This is Contribution No. 327, Dept. of Geology, Kent State University. REFERENCES Ailing, H.L. and Briggs, L.I., 1961. Stratigraphy of Upper Silurian Cayuganevaporites. Bull. Am. Assoc. Pet. Geol., 45: 515-547. Ball, J.W. Nordstrom, D.K. and Jenne, E.A., 1980. Additional and revised thermochemical data and computer code for WATEQ2--a computerized chemical model for trace and major element speciation and mineral equilibria of natural waters. U.S. Geol. Surv., Water Resour. Invest., 78-116: 1-109. Berry, W.B.N. and Boucot,A.J., 1970. Correlation of the North American Silurian rocks. Geol. Soc. Am., Spec. Pap., 102: 1-289. Bragg, L., Claringbull, G.F. and Taylor, W.H., 1965. Crystal Structures of Minerals. Cornell Univ. Press, Ithica, N.Y., 409 pp. Butler, G.P., 1970. Holocene gypsum and anhydrite of the Abu Dhabi sabkha, Trucial Coast: an alternative explanation of origin. In: J.L Ran and L.F. Dellwig (Editors), Third Symposiumon Salt. North. Ohio Geol. Soc., 1: 120-152. Carlson, E.H., 1983. The occurrence of Mississippi Valley-typemineralization in northwestern Ohio. In: G. Kisvarsanyi, S.K. Grant, W.P. Pratt and J.W. Koenig (Editors), Int. Conf. on Mississippi Valley Type Lead-Zinc Deposits. Univ. of Missouri, Rolla, Mo., pp. 424-435.

111 Carlson, E.H., 1986. Localization of S r - F - Z n - P b mineralization in Lockport mound structures, northwestern Ohio. Geol. Soc. Am., North-Central Section, Field Trip Guidebook, Kent State Univ., Kent, Ohio, 25 pp. Chowns, T.M. and Elkins, J.E., 1974. The origin of quartz geodes and cauliflower cherts through the silicification of anhydrite nodules. J. Sediment. Petrol., 44: 885-903. Dent Glasser, L.S., Glasser, F.P. and Taylor, H.F.W., 1963. The role of oriented transformations in mineralogy. Min. Soc. Am., Spec. Pap., 1: 200-203. Edwards, A.B., 1954. Textures of the Ore Minerals and their Significance. Australas. Inst. Min. Metall., 242 pp. Evans, G. and Shearman, D.J., 1964. Recent celestine from the sediments of the Trucial Coast of the Persian Gulf. Nature, 202: 385-386. Frazier, W.J., 1975. Celestite in the Mississippian Pennington Formation, central Tennessee. Southeast. Geol., 16: 241-248. Illing, L.V., Wells, A.J. and Taylor, J.C.M., 1965. Penecontemporary dolomite in the Persian Gulf. Soc. Econ. Paleontol. Mineral., Spec. Publ., 13: 89-111. Janssens, A., 1977. Silurian rocks in the subsurface of northwestern Ohio. Ohio Geol. Surv., Rep. Invest., 100: 1-96. Jodry, R.L, 1969. Growth and dolomitization of Silurian reefs, St. Clair County, Michigan. Bull. Am. Assoc. Pet. Geol., 53: 957-981. Kessen, K.M., Woodruff, M.S. and Grant, N.K., 1981. Gangue mineral STSr/86Sr ratios and the origin of Mississippi Valley-type mineralization. Econ. Geol., 76: 913-920. Kinsman, D.J.J., 1966. Gypsum and anhydrite of Recent age, Trucial Coast, Persian Gulf. In: J.L. Rau (Editor), Second Symp. on Salt. North. Ohio Geol. Soc., 1: 302-326. Longman, M.W. and Mench, P.A., 1978. Diagenesis of Cretaceous limestones in the Edwards aquifer system of south-central Texas: a scanning electron microscope study. Sediment. Geol., 21: 241-276. Lucia, F.J., 1972. Recognition of evaporite-carbonate shoreline sedimentation. Soc. Econ. Paleontol. Mineral., Spec. Publ., 16: 160-191. Masson, P.H., 1955. An occurrence of gypsum in southwest Texas. J. Sediment. Petrol., 25: 72-77. Murray, R.C., 1964. Origin and diagenesis of gypsum and anhydrite. J. Sediment. Petrol., 34: 512-523. Nickless, E.F.P., Booth, S.J. and Mosley, P.N., 1975. Celestite deposits of the Bristol area. Trans. Inst. Min. Metall., 84: B62-B64. Nicldess, E.F.P., Booth, S.J. and Mosley, P.N., 1976. The celestite resources of the area north-east of Bristol. Miner. Assess. Rep., 25: 1-83. Olaussen, S., 1981. Formation of celestite in the Wenlock, Oslo Region Norway--evidence for evaporitic depositional environments. J. Sediment. Petrol., 51: 37-46. Schreiber, B.C., 1978. Environments of subaqueous gypsum deposition. Soc. Econ. Paleontol. Mineral., Short Course, 4: 43-73. Shearman, D.J., 1966. Origin of marine evaporites by diagenesis. Trans. Inst. Min. Metall., 75: B208-B215. Shearman, D.J., 1978. Evaporites of coastal sabkhas. Soc. Econ. Paleontol. Mineral., Short Course, 4: 6-29. Summerson, C.H., 1966. Crystal molds in dolomite: their origin and environmental interpretation. J. Sediment. Petrol., 36: 221-224. Summerson, C.H., 1970. Patterns of Devonian sand on the North American craton and their interpretation: reply. Geol. Soc. Am. Bull., 81: 2521-2522. Summerson, C.H. and Swann, D.H., 1970. Patterns of Devonian sand on the North American craton and their interpretation. Geol. Soc. Am. Bull., 81: 469-490. Truesdell, A.H. and Jones, B.F., 1974. WATEQ, a computer program for calculating chemical equilibria of natural waters. J. Res. U.S. Geol. Surv., 2: 233-248.

112 Ultieg, J.R., 1964. Upper Niagaran and Cayugan stratigraphy of northeastern Ohio and adjacent areas. Ohio Geol. Surv., Rep. Invest., 51: 1-48. Wells, L.D., Jones, D.K. and Schaub, W.J., 1983. Celestite from Salem, Indiana. Min. Rec., 14: 7-12. West, I.M., 1960. On the occurrence of celestine in the Caps and Broken Beds at Durlston Head, Dorset. Proc. Geol. Assoc., London, 71: 391-401. West, I.M., 1964. Evaporite diagenesis in the Lower Purbeck Beds of Dorset. Proc. Yorkshire Geol. Soc., 34: 315-331. West, I.M., 1965. Macrocell structure and enterolithic veins in British Purbeck gypsum and anhydrite. Proc. Yorkshire Geol. Soc., 35: 47-58. West, I.M., 1973. Vanished evaporites--significance of strontium minerals. J. Sediment. Petrol., 43: 278-279. West, I.M., Brandon, A. and Smith, M., 1968. A tidal flat evaporitic facies in the Vis6an of Ireland. J. Sediment. Petrol., 38: 1079-1093. Wood, M.W. and Shaw, H.F., 1976. The geochemistry of celestites from the Yate area near Bristol (U.K.). Chem. Geol., 17: 179-193.