RESEARCH REPORTS 559 Comparative Taxonomic Richness and Abundance of Late Ordovician Gastropods and Bivalves in Mollusc-rich Strata of the Cincinnati Arch PHILIP M. NOVACK-GOTTSHALL Department of Biology, Box 90338, Duke University, Durham, NC 27708-0338, E-Mail: [email protected] ARNOLD I. MILLER Department of Geology, PO Box 210013, University of Cincinnati, Cincinnati, OH 45221-0013 PALAIOS, 2003, V. 18, p. 559±571 roevolutionary rates (Sepkoski, 1978; Stanley, 1979; Raup and Sepkoski, 1984; Gilinsky, 1994; Foote, 2000). Howev- Using a ®eld analysis of Upper Ordovician mollusc-rich er, while such studies are essential for understanding the faunas of the Cincinnati Arch, this study tests whether the ecological underpinnings of evolution, they are limited by large-scale patterns of Ordovician gastropods and bivalves their large-scale observational framework. In other words, observed in a companion study are maintained at the ®ner they cannot be used to test ecological causes of these pat- scales of individual strata and localities, and when utiliz- terns without using ®ner-scale data, particularly without ing abundance data in addition to taxonomic richness. using abundance data (see also Lidgard, 2002). Non-metric multidimensional scaling and several statisti- With the exception of terrestrial plant paleoecology cal analyses show that the taxonomic richness and abun- where the use of abundance data is commonplace (Pfeffer- dance of these classes within samples were signi®cantly korn and Thomson, 1982; DiMichele et al., 1985; Di- negatively correlated, such that bivalve-rich settings were Michele and Wing, 1988; Lupia et al., 1999), use of abun- only sparsely inhabited by gastropods and vice versa. There dance data in other branches of paleontology historically also were important environmental differences between has been limited. Because of this de®ciency, many pale- these classes. Gastropods were most dominant in shallow, ontological hypotheses based on taxonomic richness have carbonate-rich, and generally low-turbidity settings. Gas- assumed implicitly that abundance patterns mirror taxo- tropods also occurred in restricted lagoons, where bivalves nomic richness in local settings (but see Sepkoski and were only minor elements. In contrast, bivalves were most Miller, 1985; Miller, 1989). However, it is critical that pa- dominant in deep subtidal, siliciclastic shales with high leontologists test these assumptions, and there are good levels of turbidity. Both in terms of abundance and taxo- reasons to be skeptical. For example, Wing et al. (1993) nomic richness, these results strongly support those ob- demonstrated that, while several Late Cretaceous terres- served at the larger scales of paleocontinents and the globe. trial associations were strongly dominated in terms of tax- Taken together, these results argue that, despite similar tax- onomic richness by angiosperms, these plants were only onomical diversi®cation patterns of these classes at the minor components of the preserved fossils when abun- global scale and heterogeneous patterns among paleoconti- dance was considered. Lupia et al. (1999) went a step fur- nents and among regions within Laurentia, gastropods ther to demonstrate with richness and abundance data and bivalves had quite different, yet unchanging, environ- that the ascension of angiosperms during the Cretaceous mental distributions throughout the Ordovician, and that was accompanied by a major, and presumably competi- these classes did not co-occur to a signi®cant degree, either tively driven, decline in free-sporing plants; in contrast, in terms of taxonomic richness or abundance. the gymnosperm biota underwent only minor changes, de- spite traditional expectations guided by richness-only data. Among marine paleontologists, McKinney et al. (1998) showed that, despite only minor changes in taxo- INTRODUCTION nomic richness of cheilostome bryozoans following the K/T The advantages of taxonomic data in paleontology are mass extinction, there was a long-term decline in the per- self-evident, whether for documenting ®rst and last strati- cent skeletal mass of cheilostomes in shelf settings over graphic occurrences of taxonomic lineages (Sepkoski, the same interval. 1984; Niklas et al., 1983; Benton, 1985), or as faunal lists Thus, there are good reasons to question whether taxo- detailing the taxonomic richness of a particular strati- nomic richness data alone are adequate to identify and ex- graphic unit within a geographically con®ned region (Mill- plain ecological processes in the fossil record. At the same er and Foote, 1996; Miller, 1997; Alroy et al., 2001). Pat- time, there are ample indications that fossil assemblages terns in such data often are used to infer particular ecolog- can maintain reasonable proxies of the relative abundanc- ical processes, whether involving environmental prefer- es of readily preservable taxa, based primarily on analyses ence (Bretsky, 1969; Bottjer and Jablonski, 1988), of macroscopic molluscan associations (Miller, 1988; Kid- competition (Van Valen, 1973; Gould and Calloway, 1980; well and Bosence, 1991; Kidwell and Flessa, 1996; Kid- Thayer, 1983; Rosenzweig and McCord, 1991; Vermeij, well, 2001). In fact, the preservation of relative abundance 1987), resource utilization (Bambach, 1983, 1985), or mac- has enabled recognition of biologically meaningful faunal Copyright Q 2003, SEPM (Society for Sedimentary Geology) 0883-1351/03/0018-0559/$3.00 560 NOVACK-GOTTSHALL & MILLER gradients in fossil associations that show similar patterns GEOGRAPHIC, STRATIGRAPHIC, AND to those found in living communities (Ziegler, 1965; Whit- DEPOSITIONAL CONTEXT taker, 1970; Springer and Bambach, 1985; Miller, 1988; Patzkowsky, 1995; Olszewski and Patzkowsky, 2001; Ol- Sedimentation in the Cincinnati Arch generally consists szewski and Kidwell, 2002). of mixed carbonate-siliciclastic lithologies in which ®ne- In a companion study, Novack-Gottshall and Miller grained siliciclastics are interbedded with storm-in¯u- enced packstones, wackestones, and calcisiltites (Holland, (2003) used a literature-derived database of Ordovician 1993; Jennette and Pryor, 1993; Holland et al., 1997; Web- fossil occurrences of gastropods and bivalves to determine ber, 2002). Recent sequence stratigraphic work in the type whether these two molluscan classes, which were promi- Cincinnatian con®rms that primary controls on sedimen- nent members of Sepkoski's (1981) Modern Evolutionary tation were eustatic sea-level changes and storm in¯uence Fauna, displayed similar environmental and geographic in an equatorial epeiric sea with occasional, large pulses of diversity dynamics during the Ordovician Radiation at siliciclastic sediments from eastern Taconic sources (Hol- scales ranging from global to regional. At a global scale, land, 1993; Dattilo, 1996; Holland and Patzkowsky, 1996; both classes displayed similar diversity trajectories. How- Holland et al., 1997; Webber, 2002). Because of this vari- ever, at ®ner scales, their dynamics were more disparate. ability, there is an array of depositional environments in Bivalves were taxonomically richest in high-latitude pa- the region, ranging from peritidal and restricted dolomites leocontinents dominated by siliciclastic sedimentation, and limestones, shoal coquinites, and shelf grainstones, to whereas gastropods were taxonomically richest in carbon- deeper-water packstones, shales, and mudstones, al- ate-rich, equatorial paleocontinents. Similar patterns also though shelf conditions predominated. In general, shal- were expressed among regions within Laurentia, with sta- lower-water lithologies are found to the southeast, the pre- tistically distinct diversity trajectories within regions, in sumed upslope direction of a paleoramp (Holland, 1993; which bivalves dominated deeper, more siliciclastic-rich Cuffey, 1998; but see Miller et al., 2001) with a maximum environments than gastropods. depth of 25 to 50 m (Frey, 1987; Cuffey, 1998). The database for the companion study, however, did not The macrofauna is composed primarily of brachiopods include relative-abundance data for fossil occurrences be- and bryozoans, with less abundant trilobites, molluscs, cause of its scarce documentation in the literature. Fur- and echinoderms. Total molluscan fossil abundance in thermore, the database was limited to comparisons at these strata is typically less than 10% throughout the re- gion (Meyer et al., 1981; Goldman, 1993), but there are a stratigraphic scales larger than individual horizons at a variety of strata in which bivalves (Pojeta, 1971; Frey, single locality. Thus, the conclusions were limited to pat- 1987; Miller, 1989) and gastropods (Swadley, 1979; Frey, terns of taxonomic richness at scales that could mask im- 1987; Wahlman, 1992; Dattilo, 1996) exceed 50% of the portant underlying ecological trends. In particular, it preserved fossil material. The use of abundance data to could not be determined whether abundance data yielded compare bivalves and gastropods in such settings is war- patterns contrasting with those observed from taxonomic ranted from a taphonomic perspective because these clas- richness. More importantly, in cases where bivalves and ses share similar preservation potentials (Valentine, 1989; gastropods appeared to co-occur at the spatial and tempo- Palmqvist, 1991; Kidwell and Flessa, 1996; Foote and Sep- ral resolution of the database, the nature of the data pre- koski, 1999; Kidwell, 2001). Furthermore, mollusc-rich cluded determination whether