Calcite precipitation and dissolution of biogenic ... - Wiley Online Library

Calcite precipitation and dissolution of biogenic aragonite in shallow Ordovician calcite seas TIMOTHY J. PALMER AND MARK A. WILSON Palmer, T.J. & Wilson, M.A. 2004 12 01: Calcite precipitation and dissolution of biogenic aragonite in shallow Ordovician calcite seas. Lethaia, Vol. 37, pp. 417±427. Oslo. ISSN 0024-1164. The Ordovician was a time of extensive and pervasive low-magnesium calcite (LMC) precipitation on shallow marine sea ¯oors. The evidence comes from ®eld study (extensive hardgrounds and other early cementation fabrics in shallow-water carbonate sequences) and petrography (large volumes of marine calcite cement in grainstones). Contemporaneous sea-¯oor events, particularly relationships with boring and encrusting organisms and reworking in sequences of intraformational conglomerates, con®rm the early timing of such LMC cementation, and also of widespread associated aragonite dissolution. Local evidence points to the dissolved aragonite as a signi®cant source of the calcite cement. This scenario, and the fabrics that provide the evidence for it, are likely to be pointers to other times in the stratigraphic record when LMC was the predominant shallow marine precipitate (Calcite Sea times). The combination of rapid calcite precipitation and aragonite dissolution at a time early in the Phanerozoic when many major invertebrate groups were becoming established may have acted as an in¯uence on the evolution of both their skeletal mineralogy and their ecology. & Aragonite dissolution, calcite precipitation, calcite seas, early marine diagenesis, hardgrounds, intraformational conglomerate, marine cements. Timothy J. Palmer [[email protected]], The Palaeontological Association, c/o Institute of Geography and Earth Sciences, University of Wales, Aberystwyth, Ceredigion SY23 3DB, Wales, UK; Mark A. Wilson, Department of Geology, College of Wooster, Wooster, Ohio 44691, USA; 17th April 2003, revised 23rd July 2003.

One of the most signi®cant shifts of orthodoxy among carbonate sedimentologists in recent years has been the gradual acceptance that many ancient seas did not behave like Recent ones in respect to the inorganic precipitation of carbonate minerals (see Stanley & Hardie 1998). Recent shallow carbonate-precipitating seas produce fabrics such as ooliths and submarine cements that are composed of the more soluble forms of CaCO3 such as aragonite and high-magnesium calcite (HMC). At the present day, only with passage into deeper, colder marine waters does calcite with less magnesium precipitate; only in the deep sea does lowmagnesium calcite (LMC) predominate among marine cements in nodules and crusts (Schlager & James 1978). In contrast, ooliths and submarine cements from shallow water facies in ancient rocks are commonly (unless dolomitized) composed of LMC. Following many years in which uniformitarian expectations led workers to assume this mineralogy was of diagenetic origin (e.g. Shearman et al. 1970), they are now recognized as being primary (Sandberg 1975, 1983; Wilkinson et al. 1982, 1985). A widely held view is that periods of earth history during which LMC was the dominant shallow marine precipitate (`Calcite Sea' times) have alternated with periods (`Aragonite Sea'

times) in which calcite only precipitated in deeper water, and shallower marine carbonate precipitates were dominated by aragonite and HMC, as at present (Wilkinson et al. 1985; Wilkinson & Given 1986). Calcite seas particularly predominated in the early Palaeozoic and the later Mesozoic. Since thermodynamic considerations predict that calcite, the less soluble polymorph of calcium carbonate, should always precipitate in preference to aragonite, aragonite seas represent times of inhibition of calcite precipitation, at least in shallow water (Morse & Mackenzie 1990, p. 707). In contrast, calcite seas represent times when styles of carbonate dissolution and precipitation typical of the Recent deep sea operated in shallow seas as well (Bates & Brand 1990). The theoretical reasons why seawater has been calcite-inhibiting during some periods of geological history have been widely discussed (Sandberg 1975, 1983, 1985; Wilkinson et al. 1985; Wilkinson & Given 1986; Burton & Walter 1987; Wilkinson & Walker 1989; Bates & Brand 1990; Opdyke & Wilkinson 1993; Stanley & Hardie 1998). Elevated Mg:Ca ratios inhibiting calcite crystal growth, and enhanced CO2 levels, in turn correlated with periods of more rapid sea-¯oor spreading and reverse weathering, warmer climate, and higher sea-levels DOI 10.1080/00241160410002135 # 2004 Taylor & Francis

418

Timothy J. Palmer & Mark A. Wilson

(Fischer 1984; Mackenzie & Pigott 1981; Burton & Walter 1987; Morse & Mackenzie 1990), are two favoured mechanisms, though some workers think that several processes were at work (Burton 1993). Fluid inclusion studies provide direct evidence for low Mg:Ca in Ordovician seas compared with the Recent (Lowenstein et al. 2001). Our concern, however, is not so much with the causes of direct LMC precipitation in calcite seas, but with the early diagenetic consequences and the characteristic fabrics which may be used to indicate calcite sea times in the stratigraphic record. We have extensively studied shallow-water carbonates of Ordovician age, especially fabrics which were associated with sea-¯oor cementation and the genesis of crusts, cobbles (large intraclasts), and hardgrounds. These traditional products of early diagenesis show widespread evidence of sea-¯oor cements, as would be expected. There is also, however, considerable evidence that such cementation was operating much more widely in Ordovician shallow seas. Several recent studies indicate that a marine cement generation was part of the normal cementation history of limestones that did not form discernable hardgrounds or cobble layers on the sea-¯oor. Precipitation of LMC on the sea-¯oor was often accompanied by sea-¯oor dissolution of biogenic aragonite, and both processes resulted in characteristic porosity and cement fabrics. Such fabrics may serve as markers of calcite sea conditions when encountered in other sequences. Our observations in North America have been made in the Cambro-Ordovician carbonate sequences in Utah and Nevada (Pogonip Group of Hintze 1973; Ross et al. 1989, ranging up to Arenig in age; see discussion in Wilson et al. 1989, 1992); in the extensive shallowwater bioclastic limestones of Cincinnatian age (Caradoc and Ashgill) in the type region around Cincinnati, Ohio; and, to a lesser extent, in the Champlainian (Llandeilo and Caradoc) of Iowa (Palmer & Palmer 1977; Palmer 1978) and southern Ontario (Brett & Brook®eld 1984). We have also studied the Arenig shallow-water carbonate sequence of the Volkhov region (the type Volkhovian), 110 km east of St. Petersburg, Russia (Rozhnov & Palmer 1996). The early diagenetic features discussed in this paper are likely to be indicative of all calcite seas, not just Ordovician ones. Many of the fabrics discussed here are also present in limestones of Cambrian, Devonian, Cretaceous, and Jurassic age.

Cementation on Ordovician sea ¯oors Lithi®cation of carbonate sediment on shallow sea ¯oors by synsedimentary carbonate precipitation was

LETHAIA 37 (2004)

Fig. 1. Early marine calcite cement growing syntaxially on echinoderm and as blades on shell debris in Ordovician hardground. Kanosh Formation, Utah. Peel. Width of view = 4.0 mm.

a common phenomenon in the Ordovician. The most readily recognizable results of this process were hardgrounds, which are readily recognized in the ®eld through their specialized boring and encrusting fauna (Palmer 1982; Wilson & Palmer 1992). The early cements in Ordovician hardgrounds, unless dolomitized, retain their original LMC mineralogy (Wilkinson et al. 1982, 1985). In many, the marine precipitate forms the ®rst generation of cement within the primary porosity, growing syntaxially on the bioclasts, most of which were derived from calcitic taxa such as echinoderms, trilobites, ostracods, and brachiopods (Fig. 1). Cathodoluminescence studies show no signs of recrystallization in these early cements, and the calcite is often cloudy with inclusions. Some such inclusions have been shown to contain contemporaneous seawater (Johnson & Goldstein 1993, 1999). Early marine cement fringes in grainstones may characteristically be overlain by internal sediment that ®ltered down from the hardground surface (Aissaoui & Purser 1983; Reid et al. 1990; Nelson & James 1999; see Fig. 2). For a long time it has been recognized that the Ordovician was a time when hardgrounds formed very readily and were very abundant. They have been noted in Ordovician shelf-carbonate sequences in many parts of the world (see Wilson & Palmer 1992, and references therein), and some vertical sequences contain many closely-spaced hardgrounds. They are very common in the Champlainian of the Upper Mississippi Valley, for example, and give the impression that sea-¯oor cementation was taking place continuously, with episodic deposition of beds of sediment that in turn immediately started to lithify (Levorson & Gerk 1972; Palmer 1978). As well as being widespread, early cement growth was probably rapid and capable of ®lling large areas of

LETHAIA 37 (2004)

Fig. 2. Early marine calcite fringing cement in Ordovician grainstone hardground interrupted (at arrow) during growth by gravitational ®ne sediment deposition within interparticle pore space. Residual pore space ®lled by later blocky cement. Kanosh Formation, Utah. Peel. Width of view = 3.2 mm.

pore space. In the Cincinnatian, for example, internal molds of orthocones occur that are composed of early cement over 10 mm in thickness, forming calcitic pebbles and cobbles that were then locally perforated by Trypanites. Sea-¯oor cementation was also the initial process in the generation of intraformational conglomerates (Fig. 3). These form an abundant and characteristic carbonate lithology of Cambrian and Ordovician shallow water successions, and are more likely to be recorded in the literature than hardgrounds because they are more conspicuous, giving a fuller picture of the extent of synsedimentary sea-¯oor lithi®cation at this time in the sedimentary record. They occur in beds centimetres to metres in thickness which may persist laterally for hundreds of meters. They have been most extensively described from North America (Davies & Walker 1974; Keith & Friedman 1977;

Fig. 3. Intraformational conglomerate of redeposited earlycemented carbonate clasts in vertical view. Kanosh Formation, Utah. Width of view = 19 cm.

Calcite and aragonite in Ordovician seas

419

Aitken 1978; Markello & Read 1981; Sepkoski 1982; Sepkoski et al. 1991; Brett et al. 1983; Cantrell & Walker 1985; Palmer & Wilson 1990; Chow & James 1992; Pratt 2002). They are also a feature of Ordovician strata in northeast Australia (Henderson 1976; Radke 1980; Kennard 1981), north Greenland (SoÈnderholm & Due 1985; Peel 1985; Ineson et al. 1986; Ineson & Peel 1987; Peel & Smith 1988), China (Kobayashi 1966); and Korea (Choi et al. 1993; Lee & Chan Kim 1992). Beds of intraformational conglomerate started as patches or layers that were weakly lithi®ed at or just below the sea bed. Principally, cement was precipitated directly from the overlying seawater to form crusts held together by a thin layer of marine cement in the pore spaces. This is the traditional way of cementing hardgrounds (see review by Wilson & Palmer 1992), and such layers are weak and easily broken by reworking in the early stages of their cementation history. More rarely, some clasts were formed as small concretions generated by early cementation in ®nergrained substrates (probably in bacterial decay aureoles) just below the sea bed (Wilson et al. 1992). Some such hardgrounds and horizons of nodules occur undisturbed in sequences rich in intraformational conglomerate horizons, but frequently nodules were exhumed, and weakly-cemented crusts were ripped up and smashed to give rise to locally generated lithic rudstone fabrics. These clasts vary in their degree of rounding, implying longer or shorter periods of intermittent reworking before their eventual burial. They tend to be ¯at, implying that initial cementation extended horizontally rather than vertically in response to poroperm characteristics or to the overlying source of the cement. Associated sedimentary structures, such as imbrication and vertical stacking, indicate deposition of the rudstone layers under the high-energy storm (Sepkoski et al. 1991; Mount & Kidder 1993) or tsunami-generated (Pratt 2002) conditions which must have been highly erosive in their early stages. The intraformational cobbles from the Upper Cambrian and Lower Ordovician of Utah show exactly the same range of early cement fabrics and encrusting fauna as the hardgrounds with which they are interbedded (Wilson et al. 1992). Evidence that several episodes of reworking by storms with continuing cementation in between is provided by clasts that are themselves pieces of earlier generations of conglomerate (Fig. 4). We have seen up to three generations of cementation, break-up, re-sedimentation and re-cementation in a single bed. The global abundance of this highly characteristic sediment type in Cambrian and early Ordovician time suggests a combination of circumstances that favored rapid and pervasive cementation, together with

420


Fig. 4. Intraformational conglomerate in plan view. Large clast in centre is derived by reworking of an earlier generation of the same lithology. Kanosh Formation, Utah. Width of view = 24 cm.

frequent major erosive events which broke up and resedimented the crusts before they became too rigid. Explanations for the abundance of intraformational conglomerates at this time have hitherto stressed the absence of burrowing organisms that would otherwise have jostled and separated grains during the cementation process (Sepkoski 1982; Sepkoski et al. 1991). However, early Palaeozoic bioturbators are now known to have been more diverse and to have formed deeper burrow systems than previously assumed (Miller & Byers 1984, Sheehan & Schiefelbein 1984; Bottjer & Droser 1994). We prefer to think of a mechanism that stresses the ubiquitous coincidence of both the erosive events and the cementation process. Maybe rapid and pervasive cementation resulted from high concentrations of bicarbonate ions in the seawater, recharged as fast as they precipitated from dissolution of the high concentrations of CO2 in the Ordovician atmosphere (Berner & Kothavela 2001), and maybe these same atmospheric properties correlated with a more severe climatic regime that left many characteristic markers in the sedimentary record. Even in the absence of ready textural and faunal evidence of sea-¯oor lithi®cation at the ®eld level, there is evidence that primary porosity could be rapidly occluded on the Ordovician sea-¯oor by LMC cement growth. Ordovician and other Palaeozoic limestones from the eastern USA show a variety of criteria, including cement morphology, inclusion characteristics, cement/cement intergrowths, and cement/grain relationships that can be used to distinguish calcitic cements of sea-¯oor origin from later diagenetic (meteoric) ones (Walker & Diehl 1985; Walker 1989; Steinhauff 1989; Johnson & Goldstein

LETHAIA 37 (2004)

1999). Detailed studies in the Middle Ordovician of East Tennessee have shown that, at some levels, volumes of marine cements far exceed those of later diagenetic origin, and regularly account for 10% or more of total rock volume throughout sequences many tens of meters thick (Steinhauff et al. 1989). Taken together, these markers of early marine lithi®cation suggest that the Ordovician was a time in Earth history when large volumes of LMC precipitated from seawater. Both the calcitic mineralogy and the amounts of this cement represented conditions very different from those encountered in Recent carbonate environments, and in turn are likely to re¯ect levels of pCO2 and/or Mg:Ca ratios unlike those encountered in the equivalent environments today. The associated characteristics of the Ordovician sediments are thus unlikely to be those anticipated from a uniformitarian approach.

Behaviour of aragonite Previous accounts of Ordovician marine cementation have emphasized the precipitation of LMC cements. But in any system in which LMC is precipitating, and in which there is any source of particulate aragonite such as bioclasts, there must be a concentration gradient of the ions adjacent to the surface of the more soluble aragonite to any precipitating calcite. Some of the ions that end up in the early LMC cements will thus have come from the dissolution of nearby aragonite bioclasts, and the `normal' sources of dissolved CO3= and HCO3 for carbonate precipitation in the sea (from CO2 and CaCO3 dissolution outside the depositional environment) will be supplemented in shallow calcite seas by locally-generated ions.

Recognition of sea-¯oor aragonite dissolution in Ordovician seas Two principal problems arise in recognition of aragonite dissolution on ancient sea ¯oors. The ®rst is that, since the material is removed, it is a negative feature (the absence of the aragonite) that suggests dissolution and it is possible that the aragonitic organisms were absent in the ®rst place. For example, Cherns & Wright (2000) have inferred early dissolution of aragonite molluscs in the Silurian of Gotland, Sweden, because, though generally absent, they are present in abundance in beds where early silici®cation took place. A second problem is that aragonite is normally removed sooner or later in diagenesis, whether exposed to meteoric waters or buried, even at shallow depths (e.g. Hendry et al. 1995). Therefore

LETHAIA 37 (2004)

we have sought any evidence that the dissolution was taking place on the sea-¯oor. This is provided when the dissolution can be seen to be interacting contemporaneously with other sea-¯oor processes. We previously noted the phenomenon of early sea-¯oor aragonite dissolution in shallow calcite seas of Jurassic age (Palmer et al. 1988) but recent work in Utah and in the type Cincinnatian suggests that the same indicative fabrics are as common, or even more so, in the Ordovician. There are several variations on the theme of dissolution of aragonite bioclasts on the sea ¯oor. The bioclasts in question are largely cephalopods, gastropods, and bivalves. The latter two groups contain members which were entirely aragonitic, as well as others which displayed a calcitic outer shell layer over inner layers of aragonite (the normal condition for many pteriomorph bivalves; see Carter 1990). Collections have only been made from limestone and limestone/shale units in which primarily calcitic groups, such as corals, bryozoans, brachiopods, trilobites, and ostracods are well-preserved with full microstructural detail. Aragonite taxa are almost always preserved as molds (most commonly) or as sparry calcite casts (much more rarely) in such beds. The following criteria have been used to recognize cases when dissolution took place on the sea-¯oor, shortly after deposition, at the same time as early


421

cementation, sedimentation, and faunal colonization processes were also continuing. They are summarized in Figure 5. (1) Mold wall encrustation. ± This is the most straightforward criterion, and was that principally used to recognize early dissolution in Jurassic hardgrounds by Palmer et al. (1988). The aragonite shell was ®rst cemented into a hardground or cobble so that, following dissolution, the mold remained open. If the open mold intersected with the surface or was exposed by sea-¯oor erosion, then its surface could be colonized by small encrusters (Figs 5B1a, 6A). The most common encrusting groups in the Ordovician are trepostome or cryptostome bryozoans, cornulitids, pelmatozoan echinoderms, and Sphenothallus. This process is even more graphically illustrated if the shell was bored before dissolution. Then casts of the borings may also provide a substrate for the post-dissolution encrusters (Figs 5B1b, 6B). In the type Cincinnatian, aragonite bivalves such as Ambonychia are commonly found as molds within slabs of shelly rudstone that do not obviously form part of a hardground surface. Such molds are often encrusted by small taxa such as cornulitids and runner-type bryozoans, and they indicate the ubiquity of synsedimentary cementation even in the absence of the more conspicuous

Fig. 5. Evidence for the early timing of aragonite dissolution on Ordovician shallow sea ¯oors is provided when the resulting molds are affected by contemporaneous sea¯oor processes. & A. Before aragonite (stippled) dissolves. & B. After aragonite dissolves and surrounding sediment is cemented. 1a, biological encrustation of external and internal molds; 1b, encrustation of cemented sediment-®lled borings in the original aragonite shell; 2a, ®ne sediment deposited in mold that intersects bed surface; 2b, ®ne sediment piped down boring into mold beneath bed surface; 3, cemented internal mold (steinkern) becomes cobble, which may be encrusted; 4, bimineralic bivalve shows encrustation of the inner face of the outer, calcitic shell layer; 5, attachment faces of encrusters originally cemented to aragonite substrate become colonized by a post-dissolutional encrusting generation.

422


LETHAIA 37 (2004)

Fig. 6. Examples of encrustation relationships depicted in Fig. 5. Cincinnatian. & A. Encrusted mould wall. & B. Encrusted boring cast. & C. Encrusted face of calcite outer shell layer. & D. Revealed attachment face colonized by later encruster. Width of all views = 30 mm.

indicators such as hardgrounds and cobbles (Bodenbender et al. 1989). (2) Sediment-®lled molds. ± Grainstone hardgrounds frequently show evidence of ®ne sediment ®ltering down to occupy residual primary porosity that remains after the ®rst generation of marine cement has precipitated and held the grains together (e.g. Nelson & James 1999). Such interstitial sediment is often found ®lling examples of biomoldic pore space at hardground and cobble surfaces (Fig. 5B2). Locally, it may be heavily burrowed. Borings in the hardened surface may intersect with biomolds below the surface and carry sediment down into them. (3) Steinkern encrustation. ± Reworking of a partiallycemented layer following aragonite loss may separate out lithi®ed internal molds as discrete cobbles, which may in turn be encrusted or bored (Fig. 5B3). Striking encrusted steinkerns occur in the Kanosh Shale of Utah, derived from large endoceratid nautiloids. The internal molds of each of the chambers were cemented so that a series of nested molds became exposed following solution of the shell. Often these became

separated and were transported and incorporated as discrete clasts in the intraformational conglomerates, so the dissolution can be inferred even if they were not encrusted (Wilson et al. 1992). (4) Calcite inner face encrustation. ± Following the dissolution of the inner aragonite shell layer of bimineralic bivalves, the inner face of the outer calcite shell layer became exposed and was in turn encrusted. We have most often seen this with Pterinea spp. in the Cincinnatian (Figs 5B4, 6C). (5) Revealed attachment face encrustation. ± If an aragonite shell was encrusted by a calcitic epifauna (usually bryozoans, which cover a large surface area), the attachment faces (undersides) of this epifauna became exposed after dissolution of the shell. The attachment face was in turn directly encrusted by postdissolution encrusters, so that the pre- and postdissolution generations of epifauna are directly ventrally adpressed (Figs 5B5, 6D). (6) Marine cemented molds. ± Thin section examination has revealed examples of marine cements in

LETHAIA 37 (2004)


423

Fig. 7. Growth of early calcite cement within secondary porosity after aragonite dissolution. & A. Thin early calcite isopachous rind in primary porosity over all grains. & B. Aragonite dissolves; isopachous rind starts to grow within secondary porosity and continues in primary porosity.

hardground crusts partially ®lling molds after aragonite dissolution. A ®rst increment of submarine cement grew on the bioclast surface prior to dissolution, and this early cement layer provided support for the mold following dissolution. Subsequently precipitated marine cement both thickened the earlier layer and coated the inside of the mold (Figs 7 & 8). The fabrics described above are commonly encountered in the North American Ordovician sequences that we have examined, and all involve cementation as well as the aragonite dissolution (see also Kendall 1977). However, dissolution may have taken place without lithi®cation. In Europe, although hardgrounds are common in the Baltoscandian Ordovician (Janusson 1961; Holmer 1983), there is also some evidence of aragonite dissolution from soft sea-¯oors. The Swedish `Orthoceras Limestone' containing many large orthocone nautiloids is widely used as a polished internal ornamental stone and can be studied in many walls and ¯oors (e.g. surfaces in the Geological Survey

Fig. 8. Arrow marks original lower edge of aragonite shell. Early calcite cement rind is thick in primary porosity, but thinner where aragonite shell took time to dissolve. Cobble, Kanosh Formation, Utah. Peel. Width of view = 3.8 mm.

and University Geology Department in Copenhagen, Denmark). There are large open dwelling burrows of the Balanoglossites type that were excavated in the soft sediment. R.G. Bromley (personal communication, 1996; see also Ekdale & Bromley 2001) has noted burrows encountering orthocones that were sometimes de¯ected by the shell, but in other instances were unimpeded by it, passing from the external sediment, through the sediment ®lling the camerae, and out again with no apparent de¯ection. Thus it appears that the aragonite was dissolving on the sea-¯oor without any accompanying cementation, at least until after the ®nal episode of burrowing. Similarly, in the Russian Upper Volkhovian, which is also cephalopod-rich, wide sediment-®lled gaps corresponding to phragmocone walls and septa are often ®lled with contrasting sediment from the bed-matrix, but direct evidence for cementation is lacking.

Dissolved aragonite as a local cement source Aragonite dissolves in seawater to give the same ions (Ca, HCO3 , CO3=) from which calcite cement precipitates. The dissolving aragonite, in an environment that is undergoing rapid and pervasive cementation by calcite, at the very least supplements the pool of dissolved carbonate and enhances cementation and may even provide the principal source, as some workers have suggested (Johnson & Goldstein 1999). Whether this process gives rise to a visible effect, causing cementation to take place preferentially or more rapidly adjacent to an aragonite source, will be in¯uenced by the relative importance of ¯ow versus diffusion in the transport of the liberated ions. An aragonite bioclast at the surface of the sediment pile and in contact with the overlying seawater, or a shell

424


fragment in a porous sediment, is likely to be ¯ushed rapidly compared with the rate of dissolution, so that there is no concentration build-up in the adjacent pore waters over and above the background seawater concentration. However, this situation is likely to be different for aragonite bioclasts that lie just below the sediment water interface, especially if surrounded by ®ne-grained sediment. In such cases, dissolved ions may build up to such concentrations adjacent to the source that calcite cement starts to precipitate in the pore spaces, leading to development of a concretion centered on the dissolving fossil. This carbonate will be in addition to any that is derived from the more traditional source of cement in concretions, such as HCO3 from microbial decay of organic matter. Intraformational conglomerates offer a natural experiment to test this scenario, and to con®rm that it is a relatively early sea-¯oor related effect. Reworked clasts in the Ordovician carbonate conglomerates of Utah contain many recognizable bivalves, gastropods and cephalopods that show thin layers of cemented matrix encasing the void where the shell would have been. One very common clast type consists of rounded and eroded (but still readily recognizable) internal gastropod molds made of cemented micrite (Fig 9), which we interpret as a ®ne-grained sediment ®ll within the gastropod that was cemented by the products of shell dissolution prior to reworking. Some large nautiloids, which would have yielded considerable volumes of potential cement on dissolution, were reworked and redeposited together with a covering sleeve of cemented sediment, up to 1 cm thick, adhering to the outside (Fig. 10). Locally, broken pieces of what appear to be such sleeves or similar broken pieces with molds of the external

Fig. 9. Plan view of complete and broken reworked and cemented sediment molds of formerly aragonite gastropods, dissolution of which probably provided the cement to lithify the adjacent sediment. Kanosh Formation, Utah. Width of view = 22 cm.

LETHAIA 37 (2004)

Fig. 10. Plan view of large nautiloid, reworked with external sleeve of early cemented sediment; cement is interpreted as having been derived from the dissolution of the nautiloid aragonite. Kanosh Formation, Utah. Width of view = 33 cm.

surfaces of gastropods or bivalves on one side are also found. In the Utah sequences, receptaculitids occur as cemented and reworked molds, or in larger lumps in which some of the surrounding original sediment is also present, con®rming that they too originally had at least a partially aragonitic composition. These reworked nodules centred on aragonite shells support the idea of locally derived aragonite as an additional cement source in calcite seas, and con®rm its early (pre-reworking) timing. Such clear examples suggest other processes that are likely to have operated. Many 1±5 cms thick beds in the Utah Ordovician, for example, contain abundant gastropods set in a micritic matrix. Some clasts of this lithology are encountered in the conglomerates, but many such beds are uneroded, so the timing of lithi®cation is not immediately apparent. Nevertheless, we suspect that all were early-cemented at least partially with aragonite-derived cement. It is also quite likely that, as well as recognizable aragonitic macrofossils, there were also quantities of unrecognizable aragonitic debris derived from broken shells. There may also have been aragonite remnants of more fragile groups such as algae. Beds containing appreciable amounts of such material would also become autocemented on the sea ¯oor, and might or might not subsequently become reworked into the intraformational conglomerates. Many of the clasts consist of uniformly ®ne-grained calcite-cemented fragments that require an explanation for the source of cement. Supply of external cement through circulation of seawater would have been inhibited by low permeability, but remobilization of evenly-distributed aragonite within the parent sediment would account for their rapid lithi®cation. We suspect that this kind of autocementation of micrites in shallow water sequences may be a telling indicator of calcite seas (see Lasemi & Sandberg 1984; Lasemi et al. 1990).

LETHAIA 37 (2004)

Effects on the Ordovician epifauna As well as local early diagenetic consequences, the readiness with which aragonite shells dissolved on the sea-¯oors of the early Palaeozoic is likely to have had a bearing on evolutionary history at a critical time in the radiations of many of the groups involved. It has often been pointed out that invertebrate faunas at this time were dominated by groups with calcite skeletons, and the tendency of the seas to precipitate calcite has been implicated (Stanley & Hardie 1998). But maybe it was the tendency for such seas to dissolve aragonite that acted to select for calcite skeletons rather than the fact that organisms were adopting the normal inorganic precipitate for their skeletal minerals. The exceptions test the rule: some of the molluscs undoubtedly were aragonite, but many of these were infaunal (Pojeta 1971) and thus removed from the most rapid effects of shell dissolution. Many of the epifaunal molluscs, both bivalves and (unusually for the group as a whole) gastropods (e.g. the Cincinnatian Cyclonema), evolved outer shell layers of calcite. Possibly the dominance of the pteriomorphs and the abundance of families with outer calcite shell layers in marine epifaunas throughout the Phanerozoic had their roots in the Ordovician seas that threatened to dissolve their ancestors. Harper et al. (1997) have surveyed the times of origin of the major bivalve clades throughout the Phanerozoic, and have noted the statistically signi®cant origination of forms with outer calcite shell layers during calcite sea times, of which the Ordovician and the Jurassic were the most important. We have also long been curious about why some of the epifaunal aragonitic Ordovician genera in the Cincinnatian such as Modiolopsis are preserved with a thick black outer shell covering (e.g. Pojeta 1971, pl. 15, ®g. 6). It now seems likely that this was a hypertrophied periostracum that conferred some protection against dissolution during life, similar to the situation seen in Recent unionids that are susceptible to dissolution in their fresh-water habitats. The rapid dissolution of skeletal aragonite and the precipitation of calcite cements in Ordovician sediments resulted in many cases of extensive hardground formation. Some muddy sea-¯oors were converted relatively quickly into rocky substrates. Soft-sediment infaunal organisms were replaced by borers and nestlers, and epifaunal organisms adapted to a rigid surface. Wilson et al. (1989), Guensburg & Sprinkle (1992), Wilson et al. (1992), Wilson & Palmer (1990, 1992), and Rozhnov & Palmer (1996) have noted the temporal correlation between the increasing abundance of hardgrounds in the Early Ordovician and the radiation of attaching echinoderms and have suggested that hardgrounds were the critical physical


425

factor. Likewise, the Early Ordovician radiation of bryozoans may be connected with the same abundance of hardgrounds. Hu & Spjeldnaes (1991) describe some of the earliest bryozoans from the BaltoScandian area as having large attachment bases, which Wilson & Palmer (1992) proposed were adaptations to widespread hardgrounds.

Conclusions (1) Calcite seas were often characterized by the rapid inorganic precipitation of calcite cement in shallow carbonate sediments, especially those of the Ordovician and Jurassic. (2) Aragonitic shells in calcite seas dissolved rapidly after the deaths of the creating organisms. (3) This early diagenesis and the response of contemporary organisms left a characteristic fabric in the sediments, most notably mold wall encrustation, sediment-®lled molds, encrusted steinkerns, encrusted calcitic faces of bimineralic shells, and attachment surface encrustation. (4) The calcium carbonate ions liberated by aragonite dissolution may have been, at least locally, immediately incorporated in the precipitating calcite cements, forming haloes of cemented sediment adjacent to the dissolving aragonite source. The early timing of this process is con®rmed where these incipient nodules were reworked within intraformational conglomerates. (5) The rapid dissolution of aragonite on the Ordovician sea-¯oors may have in¯uenced the evolution of shelled marine invertebrates by encouraging the development of calcitic or bimineralic shells or shells protected by thick periostraca. (6) The rapid dissolution of aragonite and precipitation of calcite cements in shallow marine sediments formed extensive hardgrounds which may have played critical roles in early echinoderm and bryozoan radiations. Acknowledgements. ± Partial support for this work was received from the donors of the Petroleum Research Fund, administered by the American Chemical Society, as well as from the Luce Fund for Distinguished Scholarship at The College of Wooster.

References Aissaoui, D.M. & Purser, B.H. 1983: Nature and origins of internal sediments in Jurassic limestones of Burgundy (France) and Fnoud (Algeria). Sedimentology 30, 273±283. Aitken, J.D. 1978: Revised models for depositional grand cycles, Cambrian of the southern Rocky Mountains, Canada. Bulletin of Canadian Petroleum Geology 26, 515±542.

426


Bates, N.R. & Brand, U. 1990: Secular variation of calcium carbonate mineralogy; an evolution of ooid and micrite chemistries. Geologische Rundschau 79, 27±46. Berner, R.A. & Kothavela, Z. 2001: GEOCARB III: a revised model of atmospheric CO2 over Phanerozoic time. American Journal of Science 301, 182±204. Bodenbender, B.E., Wilson, M.A. & Palmer, T.J. 1989: Palaeoecology of Sphenothallus on an Upper Ordovician hardground. Lethaia 22, 217±225. Bottjer, D.J. & Droser, M.L. 1994: The history of Phanerozoic bioturbation:. In Donovan, S.K. (ed.): The Palaeobiology of Trace Fossils, 155±176. Johns Hopkins University Press, Baltimore, Maryland. Brett, C.E. & Brook®eld, M.E. 1984: Morphology, faunas and genesis of Ordovician hardgrounds from southern Ontario, Canada. Palaeogeography, Palaeoclimatology, Palaeoecology 46, 233±290. Brett, C.E., Liddell, W.D. & Derstler, K.L. 1983: Late Cambrian hard substrate communities from Montana/Wyoming; the oldest known hardground encrusters. Lethaia 16, 281±289. Burton, E.A. 1993: Controls on marine carbonate cement mineralogy: review and reassessment. Chemical Geology 105, 163±179. Burton, E.A. & Walter, L.M. 1987: Relative precipitation rates of aragonite and Mg calcite from seawater: temperature or carbonate ion control? Geology 15, 111±114. Cantrell, D.L. & Walker, K.R. 1985: Depositional and diagenetic patterns, ancient Oolite: Middle Ordovician, eastern Tennessee. Journal of Sedimentary Petrology 55, 518±531. Carter, J.G. 1990: Skeletal biomineralization: patterns, processes and evolutionary trends. 832 pp. Van Nostrand & Reinhold, New York. Cherns, L. & Wright, V.P. 2000: Missing molluscs as evidence of large-scale, early skeletal aragonite dissolution in a Silurian Sea. Geology 28, 791±794. Choi, Y.S., Kim, J. & Lee, Y.I. 1993: Subtidal ¯at-pebble conglomerates from the Early Ordovician Mungok Formation, Korea: origin and depositional process. Journal of the Geological Society of Korea 29, 15±29. Chow, N. & James, N.P. 1992: Synsedimentary diagenesis of Cambrian peritidal carbonates: evidence from hardgrounds and surface palaeokarst in the Port au Port Group, western Newfoundland. Bulletin of Canadian Petroleum Geology 40, 115±127. Davies, I.C. & Walker, R.G. 1974: Transport and deposition of resedimented conglomerate: the Cap EnrageÂ Formation, Cambro-Ordovician, GaspeÂ, Quebec. Journal of Sedimentary Petrology 44, 1200±1216. Ekdale, A.A. & Bromley, R.G. 2001: Bioerosional innovation for living in carbonate hardgrounds in the Early Ordovician of Sweden. Lethaia 34, 1±12. Fischer, A.G. 1984: The two Phanerozoic supercycles:. In Berggren, W.A. & Van Couvering, J.A. (eds): Catastrophes and Earth History: The New Uniformitarianism, 155±176. Princeton University Press, Princeton, New Jersey. Guensburg, T.E. & Sprinkle, J. 1992: Rise of echinoderms in the Paleozoic Evolutionary Fauna: signi®cance of paleoenvironmental controls. Geology 20, 407±410. Harper, E.M., Palmer, T.J. & Alphey, J.R. 1997: Evolutionary response by bivalves to changing Phanerozoic sea-water chemistry. Geological Magazine 134, 403±407. Henderson, R.A. 1976: Stratigraphy of the Georgina Limestone and a revised zonation for the early Upper Cambrian Idamean Stage. Journal of the Geological Society of Australia 23, 423±433. Hendry, J.P., Ditch®eld, P.W. & Marshall, J.D. 1995: Two stage metamorphism of Jurassic aragonitic bivalves: implications for early diagenesis. Journal of Sedimentary Research A65, 214±224. Hintze, L.F. 1973: Lower and Middle Ordovician stratigraphic sections in the Ibex area, Millard County, Utah. Brigham Young University, Geology Studies 20 (4), 3±36. Holmer, L.S. 1983: Lower Viruan discontinuity surfaces in central Sweden. Geologiska FoÈreningens i Stockholm FoÈrhandlingar 105, 29±42.

LETHAIA 37 (2004) Hu, Z.±X. & Spjeldnaes, N. 1991: Early Ordovician bryozoans from China:. In Bigey, F.P. (ed.): Bryozoaires Actuels et Fossiles: Bryozoa Living and Fossil, 179±185. Bulletin de la SocieÂteÂ des Sciences Naturelles de l'Ouest de la France, hors seÂrie 1. Ineson, J.R. & Peel, J.S. 1987: Cambrian platform ± outer shelf relationships in the NordenskioÈld Fjord region, central North Greenland. Rapports Grùnlands geologiske Undersùgelse 133, 13± 26. Ineson, J.R., Peel, J.S. & Smith, M.P. 1986: The Sjaelland Fjelde Formation; a new Ordovician formation from eastern North Greenland. Rapports Grùnlands geologiske Undersùgelse 132, 27± 37. Jaanusson, V. 1961: Discontinuity surfaces in limestones. Bulletin of the Geological Institutions of the University of Uppsala, new series 40, 221±241. Johnson, W.J. & Goldstein, R.H. 1993: Cambrian sea water preserved as inclusions in marine low-magnesium calcite cement. Nature 362, 335±337. Johnson, W.J. & Goldstein, R.H. 1999: Shallow marine low-Mg calcite cementation promoted by aragonite dissolution. American Association of Petroleum Geologists, Annual Meeting Expanded Abstracts, A69. Keith, B.D. & Friedman, G.M. 1977: A slope-fan-basin-plain model, Taconic Sequence, New York and Vermont. Journal of Sedimentary Petrology 47, 1220±1241. Kendall, A.C. 1977: Origin of dolomite mottling in Ordovician limestones from Saskatchewan and Manitoba. Bulletin of Canadian Petroleum Geology 25, 480±504. Kennard, J. 1981: The Arrinthrunga Formation: Upper Cambrian epeiric carbonates in the Georgina Basin, central Australia: Australia,. Bureau of Mineral Resources, Geology and Geophysics Bulletin 211, 1±30. Kobayashi, T. 1966: The Cambro-Ordovician formations and faunas of South Korea. Part X. Stratigraphy of the Chosen Group in Korea and south Manchuria. Section B. The Chosen Group of North Korea and northeast China. Journal of the Faculty of Science, University of Tokyo, Section II 16, 209±311. Lasemi, Z. & Sandberg, P.A. 1984: Transformation of aragonitedominated lime muds to microcrystalline limestones. Geology 12, 420±423. Lasemi, Z., Sandberg, P.A. & Boardman, M.R. 1990: New microtextural criterion for differentiation of compaction and early cementation in ®ne-grained limestones. Geology 18, 370±373. Lee, Y.I. & Kim, J. 1992: Storm-in¯uenced siliciclastic and carbonate ramp deposits, the Lower Ordovician Dumugol Formation, South Korea. Sedimentology 39, 951±969. Levorson, C.O. & Gerk, A.J. 1972: A preliminary stratigraphic study of the Galena Group of Winneshiek County, Iowa. Proceedings of the Iowa Academy of Sciences 79, 111±122. Lowenstein, T.K., Timofeeff, M.N., Brennan, S.T., Hardie, L.A. & Demicco, R.V. 2001: Oscillations in Phanerozoic seawater chemistry: evidence from ¯uid inclusions. Science 294, 1086± 1088. Mackenzie, F.T. & Pigott, J.D. 1981: Tectonic controls of Phanerozoic sedimentary rock cycling. Journal of the Geological Society of London 138, 183±196. Markello, J.R. & Read, J.F. 1981: Carbonate ramp to deeper shale shelf transitions of an Upper Cambrian intrashelf basin, Nolichucky Formation, southwest Virginia Appalachians. Sedimentology 28, 573±597. Miller, M.F. & Byers, C.W. 1984: Abundant and diverse early Palaeozoic infauna indicated by the stratigraphic record. Geology 12, 40±43. Morse, J.W. & Mackenzie, F.T. 1990: Geochemistry of Sedimentary Carbonates, 707 pp. Developments in Sedimentology 48, Elsevier, Amsterdam. Mount, J.F. & Kidder, D. 1990: Combined ¯ow origin of edgewise intraclast conglomerates; Sellick Hill Formation (Lower Cambrian) South Australia. Sedimentology 40, 315±329. Nelson, C.S. & James, N.P. 1999: Environment of aragonite dissolution at hardground surfaces in mid-Tertiary cool-water carbonates, New Zealand and southern Australia:. American

LETHAIA 37 (2004) Association of Petroleum Geologists, Annual Meeting Expanded Abstracts, A100. Opdyke, B.N. & Wilkinson, B.H. 1993: Carbonate mineral saturation state and cratonic limestone accumulation. American Journal of Science 293, 217±234. Palmer, T.J. 1978: Burrows at certain omission surfaces in the Middle Ordovician of the Upper Mississippi Valley. Journal of Paleontology 52, 109±117. Palmer, T.J. 1982: Cambrian to Cretaceous changes in hardground communities. Lethaia 15, 309±323. Palmer, T.J., Hudson, J.D. & Wilson, M.A. 1988: Palaeoecological evidence for early aragonite dissolution in ancient calcite seas. Nature 335, 809±810. Palmer, T.J. & Palmer, C.D. 1977: Faunal distribution and colonisation strategy in a Middle Ordovician hardground community: Lethaia 10, 179±199. Palmer, T.J. & Wilson, M.A. 1990: Submarine cementation and the origin of intraformational conglomerates in Cambro-Ordovician calcite seas. 13th International Sedimentological Congress 1990, Abstracts Volume, 171±172. Peel, J.S. 1985: Cambrian±Silurian platform stratigraphy of eastern North Greenland:. In Gee, D.G. & Sturt, B.A. (eds): The Caledonide Orogen±Scandinavia and Related Areas, 1078±1094. John Wiley and Sons, Chichester. Peel, J.S. & Smith, M.P. 1988: The Wandel Valley Formation (Early±Middle Ordovician) of north Greenland and its correlatives. Rapports Grùnlands geologiske Undersùgelse 137, 61±92. Pojeta, J. 1971: Review of Ordovician pelecypods:. United States Geological Survey, Professional Paper 695, 1±46. Pratt, B.R. 2002: Storms versus tsunamis: dynamic interplay of sedimentary, diagenetic, and tectonic processes in the Cambrian of Montana. Geology 30, 423±426. Radke, B.M. 1980: Epeiric carbonate sedimentation of the Ninmaroo Formation (Upper Cambrian±Lower Ordovician), Georgina Basin. BMR Journal of Australian Geology and Geophysics 5, 183±200. Reid, R.P., Macintyre, I.G. & James, N.P. 1990: Internal precipitation of microcrystalline carbonate: a fundamental problem for sedimentologists. Sedimentary Geology 68, 163±170. Ross, R.J., James, N.P., Hintze, L.F. & Poole, F.G. 1989: Architecture and evolution of a Whiterockian (early Middle Ordovician) carbonate platform, basin and ranges of western U.S.A. In Crevello, P.D., Wilson, J.L., Sarg, J.F. & Read, J.F. (eds): Controls on Carbonate Platform Development and Basin Development, 167±185. Society of Economic Paleontologists and Mineralogists Special Publication 44. Rozhnov, S.V. & Palmer, T.J. 1996: The origin of the ecosystem of hardgrounds and the Ordovician benthic radiation. Paleontological Journal 30, 688±692. Sandberg, P.A. 1975: New interpretation of Great Salt Lake ooids and ancient non-skeletal carbonate mineralogy. Sedimentology 22, 497±538. Sandberg, P.A. 1983: An oscillating trend in Phanerozoic nonskeletal carbonate mineralogy. Nature 305, 19±22. Sandberg, P.A. 1985: Non±skeletal aragonite and pCO2 in the Phanerozoic and Proterozoic. In Sundquist, E.T. & Broecker, W.S. (eds): The Carbon Cycle and Atmospheric CO2: Natural Variations, Archean to Present, 585±594. Geophysical Monographs of the American Geophysical Union 32. Schlager, W. & James, N.P. 1978: Low-magnesian calcite limestones forming at the deep-sea ¯oor, Tongue of the Ocean, Bahamas. Sedimentology 25, 675±699. Sepkoski, J.J. 1982: Flat±pebble conglomerates, storm deposits, and the Cambrian bottom fauna. In Einsele, G. & Seilacher, A. (eds): Cyclic and Event Strati®cation, 371±385. Springer-Verlag, Berlin. Sepkoski, J.J., Bambach, R.K. & Droser, M.L. 1991: Secular changes


427

in Phanerozoic event bedding and the biological overprint. In Einsele, G., Ricken, W. & Seilacher, A. (eds): Cycles and Events in Stratigraphy, 298±312. Springer Verlag, Berlin. Shearman, D.J., Twyman, J. & Karimi, M.Z. 1970: The genesis and diagenesis of oolites. Proceedings of the Geologists' Association 81, 561±575. Sheehan, P.M. & Schiefelbein, D.R.J. 1984: The trace fossil Thalassinoides from the Upper Ordovician of the eastern Great Basin: deep burrowing in the early Paleozoic. Journal of Paleontology 58, 440±447. Sùnderholm, M. & Due, P.H. 1985: Lower and Middle Ordovician platform carbonate lithostratigraphy of Warming Land, Wulff Land and Nares Land, North Greenland. Rapports Grùnlands geologiske Undersùgelse 126, 31±46. Stanley, S.M. & Hardie, L.A. 1998: Secular oscillations in the carbonate mineralogy of reef-building and sediment-producing organisms driven by tectonically forced shifts in seawater chemistry. Palaeogeography, Palaeoclimatology, Palaeoecology 144, 3±19. Steinhauff, D.M. 1989: Marine cements. In Walker, K.R. (ed.): The Fabric of Cements in Paleozoic Limestones, 37±53. University of Tennessee, Studies in Geology 20. Steinhauff, D.M., Jernigan, D.G. & Walker, K.R. 1989: Stratigraphic patterns of cementation in Middle Ordovician Holston and Rockdell Formations, Tennessee. In Walker, K.R. (ed.): The Fabric of Cements in Paleozoic Limestones, 118±138. University of Tennessee, Studies in Geology 20. Walker, K.R. 1989: Petrographic criteria for interpreting pore®llings: In Walker, K.R. (ed.): The Fabric of Cements in Paleozoic Limestones, 4±10. University of Tennessee, Studies in Geology 20. Walker, K.R. & Diehl, W.W. 1985: The role of marine cementation in the preservation of lower Paleozoic assemblages. Philosophical Transactions of the Royal Society of London, series B, Biological Sciences 311, 143±153. Wilkinson, B.H., Janecke, S.U. & Brett, C.E. 1982: Low-magnesian calcite marine cement in Middle Ordovician hardgrounds from Kirk®eld, Ontario. Journal of Sedimentary Petrology 52, 47±57. Wilkinson, B.H. & Given, K.R. 1986: Secular variation in abiotic marine carbonates: constraints on Phanerozoic atmospheric carbon dioxide contents and oceanic Mg/Ca ratios. Journal of Geology 94, 321±333. Wilkinson, B.H., Owen, R.M. & Carroll, A.R. 1985: Submarine hydrothermal weathering, global eustacy, and carbonate polymorphism in Phanerozoic marine oolites. Journal of Sedimentary Petrology 55, 171±183. Wilkinson, B.H. & Walker, J.C.G. 1989: Phanerozoic cycling of sedimentary carbonate. American Journal of Science, 289, 525± 548. Wilson, M.A. & Palmer, T.J. 1990: A review of evolutionary trends in carbonate hardground communities:. In Miller, W. (ed.): Paleocommunity Temporal Dynamics: the long-term development of multispecies assemblies, 137±152. The Paleontological Society, Special Publication 5. Wilson, M.A. & Palmer, T.J. 1992: Hardgrounds and hardground faunas. University of Wales, Aberystwyth, Institute of Earth Studies Publications 9, 1±131. Wilson, M.A., Palmer, T.J., Guensburg, T.E. & Finton, C.D. 1989: Sea ¯oor cementation and the development of marine hard substrate communities: new evidence from Cambro-Ordovician hardgrounds in Nevada and Utah. Geological Society of America, Abstracts With Programs 21, 253±254. Wilson, M.A., Palmer, T.J., Guensburg, T.E., Finton, C.D. & Kaufman, L.E. 1992: The development of an Early Ordovician hardground community in response to rapid sea±¯oor calcite precipitation. Lethaia 25, 19±34.