Ordovician Kope formation Ohio

8/14/2019 Ordovician Kope formation Ohio

1/13

RESEARCH REPORTS 205

The Detection and Importance of Subtle Biofacies within a

Single Lithofacies : The Uppe r Ordovician Kope F ormationof the Cincinn ati, Ohio Re gion

STEVEN M. HOLLAND

Department of Geology, University of Georgia, Athens, GA 30602-2501

ARNOLD I. MILLER and DAVID L. MEYER

Department of Geology, University of Cincinnati, Cincinnati, OH 45221-0013

BENJ AMIN F . DATTILO

Departm ent of Geosciences, Weber S tate Uni versity, Ogden, U T 84408

PALAIOS, 2001, V. 16, p. 205217

Environmental controls on the distribution of fossils mostcom m only are foun d by recognizing th at certain distinctive

fossil assemblages are associated with particular lithofa-cies. Lack of change in lithofacies commonly is used as in-dicating a lack of significant environmental effects on thestratigraphic distribution of fossils. The results presentedhere challenge that view. The Upper Ordovician Kope For-mation of the Cincinnati, Ohio, area has long been consid-ered a single unit, both lithostratigraphically and in termsof depositional environment. Gradient analysis of over 1000

fossil assemblages reveals subtle environmental control onthe distribution of fossils, in the absence of obvious litholog-

ic change. This gradient analysis is used to construct anecological m odel of the Kope fauna , wit h values of preferreddepth, depth tolerance, and peak abundance estimated forthe most common fossils. This method, conducted within asingle lithofacies, offers the potential for reconstructing se-quence architecture because faunas can be more sensitiverecorders of environment than lithofacies. In addition, the

presence of subtle facies control as in the Kope raises the prospect that environmental controls on paleobiologic andbiostratigraphic patterns m ay be m ore pervasive than gen-erally acknowledged.

INTRODUCTION

Paleontologists have long recognized environmental

control on the distribution of fossils. However, most re-ports of environmental controls have come from cases inwhich lith ofacies obviously differ from one a nother , ma k-ing t he recognition of facies control on fossils str aightfor-ward (e.g., Bayer an d McGhee, 1985; Pat zkowsky, 1995).This approach has led elsewhere to suggestions that theabsence of changing lithofacies is evidence of a lack of en-vironmentally controlled fossil distributions. For example,evolutionary studies are preferred where facies are un-changing, presumably because environmental conditionswere relatively stable and, therefore, fossil faunas are notinfluenced by chan ges in environment (e.g., Sheldon,1996). Although severa l resear chers ha ve recognized thatfossils are more sensitive to environmental changes than

are lithofacies (e.g., Springer a nd Bamba ch, 1985; Miller,1988, 1997; Brett, 1998), empirical support for this argu-ment h as been limited.

The Upper Ordovician Kope Formation of the Cincin-nati, Ohio, region exemplifies this problem well. Althoughworker s in t he lat e 1800s and ea rly 1900s recognized fau-nal variat ions within the Kope (summarized in Caster etal., 1955), modern workers have tended to overlook them(but see Anstey and Perry, 1973; Anstey et al., 1987; An-stey and Rabbio, 1990). Upon broad inspection, the Kopegives the impr ession t ha t it cont ains a single suite of fossiltaxa that vary uns ys tem atically in their relative abun-dance. Here, Kope faunas are shown to be far more struc-tured than previously recognized, and this variation is in-terpreted to be environmentally driven. Strong evidenceexists that the composition of faunal collections from theKope chan ges with inferred water depth. Th e lack of obvi-ous lithofacies change in the Kope suggests that depthcontrol on faunas may be more common than currentlyrecognized. The existence of such widespread depth con-trol ha s significan ce for str at igraphic pat tern s of fossil oc-curren ces, in genera l, and for t he a pplicat ion of confidencelimits t o fossil ra nges.

BACKGROUND

The 65-m-thick Kope Format ion, and t he rest of th e typeCincinnatian Series, are interpreted t o have been depos-ited on a northwa rd-dipping, storm-domina ted, mixed car-bona te-siliciclastic ra mp (Tobin, 1982; Hollan d, 1993;J en-

nett e an d P ryor, 1993; Brett an d Algeo, 1999a). Althoughthe dip of the ram p m ay have had a s light eas tw ard orw es t wa r d com p on e n t , t h e g en e r a ll y n or t h w a r d d ipthr oughout the La te Ordovician is shown by the consisten toccurren ce of sha llower water facies in Kentucky anddeeper wat er facies in Ohio and In diana (Weir et al., 1984;Hollan d, 1993). The nort h-sout h a lignmen t of gutt er castsand the east-west orientation of megaripple crests withinthe Kope also are consistent with an east-west strike tothe ramp (Duke, 1990; Jennette and Pryor, 1993).

The Kope consists mostly of siliciclastic mudstone (6094%, median 80%), followed by varying amounts of calcis-iltite (016%, median 4%), skeletal packstone and grain-stone (035%, median 14%), with much smaller propor-


2/13

206 HOLLAND ET AL.

tions of lime mudstone and wackestone. Calcisiltite bedsdisplay a wide a rray of sedimentar y stru ctures includingsmall-scale hummocky and trough cross-lamination, pla-nar lam ination, vortex r ipples , current r ipples , guttercasts, prod marks, Kinneyia and millimeter ripples (Jen-nette and Pryor, 1993; Brett and Algeo, 1999a). Beds ofskeletal packstone and grainstone commonly have ero-siona l bases, and m ay display megaripples an d large-scalecross stratification (e.g., sets thicker than 5 cm). Siliciclas-tic mudstones are graded into 35 cm event beds, withslightly silty bases. Collectively, these features are inter-pret ed as storm beds (Kreisa et al., 1981; Tobin, 1982;J en-nette and Pryor, 1993; Holland et al., 1997; Brett and Al-geo, 1999a). The Kope has been int erpret ed as an offshorefacies affected only by the strongest storms (Hay, 1981;Tobin, 1982; Holland, 1993; Jennette and Pryor, 1993;Brett and Algeo, 1999b).

Several worker s ha ve recognized lith ologic cyclicity in

the Kope (Hay, 1981; Tobin, 1982; Jennette and Pryor,1993; Holland et a l., 1997; Brett an d Algeo, 1999a). Meter-

scale cycles in the Kope consist of alternations of mud-s tone-rich intervals and lim es tone-rich intervals , al-

though individual descriptions and interpretations of thiscyclicity vary widely (see discussions in Holland et al.,1997, 1999). It remains unclear whether these alterna-

tions reflect cyclic cha nges in wat er dept h or var iations instorm frequency and intensity. Furthermore, some have

questioned t he existence of these m eter-scale cycles by ar-guing that the pattern of lithologic change cannot be dis-

tinguished from r an domness (Wilkinson et al., 1997).Kope meter-scale cycles are bundled into lar ger-scale

20-m cycles in which the meter-scale cycles are initiallythicker than average, but upsection become average to

thinner than average. The origin of these 20-m cycles isalso uncertain, although the organization of their compo-

nent meter-scale cycles, coupled with clear stacking pat-tern s of limestone-rich and m udstone-rich int ervals in the

subsur face (Dat tilo, 1996) suggest th at t hey are t he resultof relative fluctuations in sea level (Jennette and Pryor,

1993; Hollan d et al., 1997). The organizat ion a nd delinea-tion of these 20-m cycles is not en tirely un ambiguous, anddifferent workers have developed somewhat different

characterizations. All of these disagreements over t hestructure, origin, and even the presence of both scales of

cyclicity within t he Kope under score t ha t t hese cyclic lith-ologic variations a re subtle.

Although the Kope contains a spectrum of lithologies,their interbedding within the Kope is repetitive and the

Kope displays only weak large-scale changes in the rela-tive proportions of these lithologies. This persistent fine-

scale interbedding of lithologies has frustrated most at-tem pts to s ubdivide the K ope into s m aller l ithostrati-

graphic or l ithofacies units (but s ee Brett and Algeo,1999c). As a result, the Kope is considered to be a single

lithostratigraphic unit (Weiss and Sweet, 1964; Weir etal., 1984; Tobin, 1986). For th e same r easons, th e Kope hasbeen regarded as either a single lithofacies or as two or

thr ee intimately a nd repetitively interbedded lithofacies(Hay, 1981; Tobin, 1982; Weir et al., 1984; Holland, 1993;

Jenn ette and Pryor, 1993). To see firsthand the difficultyof subdividing the Kope into member-scale lithostrati-

graphic units , the m eas ured s ections in J ennette and

Pryor (1993) and Holland et al. (1997) should be exam-ined.

Com pared to the s trata, K ope faunas have received

much less attent ion. Most stu dies have either treat ed theKope collectively (e.g., Lorenz, 1973; Ha rr ison and Mah an ,1981; Holland, 1993) or have focused on particular ele-ments, su ch as bryozoans (Anstey a nd Perry, 1973) orha rdground fauna s (Wilson, 1985). None of th ese ha ve ex-amined systematically the stra tigraphicchanges of the en-tire Kope fauna in detail.

The fauna of the K ope is divers e and abundant. A l-th ough Da lve (1948) lists 240 species from th e Kope, few ofthese are common (Table 1). Epifaunal suspension feed-ers, such as brachiopods, bryozoans, and crinoids, aredominant. Trilobites ar e also locally common, as are bi-valves, gastropods, cepha lopods, ostr acods, an d grapt o-lites.

METHODS

As part of a larger project on high-resolution correlationwithin the Kope Formation, five outcrops were measuredin detail (Fig. 1). All but the lowest 10 m of the Kope is ex-posed at t he K445 section, an d all of the Kope is exposed atHolst Creek, although poorly so through most of the mid-dle interval. Only portions of the Kope are exposed atMiamitown, Aurora, and Hume. At each locality, the ex-posed Kope Formation was measured, with the lithology,sedimentary structures, and fauna of every bed thickerth an 0.5 cm r ecorded. The details of the lithologic descrip-tions and their results have been published previously(Hollan d et al., 1997, 1999, 2000; Miller et al., 1997).

To describe the fauna, every fossiliferous horizon wasexcavated, w as hed in the fi eld, and the relative abun-dan ce of each t axon tallied. Beds ofsk eletal packstone andgrainst one, and t he soles of calcisiltite beds pr oduced mostof the fossils, although some mudstones were fossiliferousan d included in this stu dy. Relative abunda nce was scoredas r ar e (12 specimens per 1000 cm 2), common (310 spec-imens per 1000 cm2) , or abundant (10 specimens per1000 cm 2). Most taxa were identified t o genus whereverpossible, including brachiopods, crinoids, t rilobites, bi-valves, cephalopods, and gastropods. Most bryozoans wereclassified on zooarial morphology, including thin bifoliate(5 mm ), thick bifoliate (5 mm), thin ra mose (5 mm),thick ramose (5 mm), and encrusting forms. Cryptosto-me and fenestellid bryozoans were listed as such, and thedistinctive bryozoans Escharopora, Aspidopora, Prasopo-ra, Stomatopora a n d Parvohallopora were described to the

genus level.These fi eld descriptions res ulted in a relative abun-

dance mat rix consisting of 1949 sam ples and 57 taxa. Tominimize th e effects of extremely ra re t axa, all ta xa occur-ring in only one sa mple were r emoved. Likewise, samplescont aining only a single taxon were removed. This cullingresulted in a final matrix of 1337 samples and 46 taxa.

This matrix was analyzed with Detrended Correspon-dence Analysis, also known a s DCA or Decorana (Hill andGauch, 1980; McCune a nd Mefford, 1997). DCA is a mul-tivaria te st at istical technique widely used with ecologicaldata to ordinate taxa along underlying ecological gradi-ents. Although other techniques, such as CorrespondenceAnalysis, Principal Components Analysis, and Polar Or-


3/13

DETECTING SUBTLE BIOFACIES 207

TABLE 1Kope taxa analyzed in this study.

Low EpifaunalSuspension Feeders

High EpifaunalSuspension Feeders

Epifaunal Scavengers,

Grazers, an d DepositF eeder s Ca r n ivor es & P a r a sit es Ot h er

B r y o z o a n sEncr us t er s

Escharopora

FenestellidsAspidopora

S tomatopora

Inart i cul at esCraniopsSchizocrania

Art i cul at esDalmanella

Platystrophia

Plectorthis

PrasoporaRafinesquinaS owerbyella

Strophomena

Zygospira

Annel i daCornulites

B i val vi aAm bonychiaModiolopsis

B r y o z o a n sCryptostomesParvohallopora

Thick bifoliateThick ramoseThin bifoliateThin ramose

Cri noi deaCincinnaticrinus

Ectenocrinus

IocrinusMerocrinus

Glyptocrinus

U n i v a l v e dM ol l uscsCyclora

Unidentifiedgastropods(includes monoplaco-

phor ans )

TrilobitaCeraurus

Cryptolithus

Calymenids(includes Flexicalymene

a n d Gravicalymene)Isotelus

AcidaspisProetidella

Gast ropodaCyclonema

H y d r o z o a n s

Annel i daScolecodonts

C e p h a l o p o d s(includes pr imari-

ly Treptoceras ,but al s o Ordo-geisonoceras a n dendocerids)

Graptolites(Planktonic suspen-

sion feeders)

Pseudolingula

(Infaun al su spensionfeeding inarticu-lates)

Lepidocoleus

(Machaeridians ofuncertain life hab-its)

Nuculoids

(Infaun al depositfeeding bivalves)

Ostracods(nekto-benthic scav-

engers)

FIGURE 1Map of study area, showing the locations of the five mea-sured and sampled outcrops.

dinat ion, also can be used t o ordina te ecological dat a, th eytend to dis tort the data s uch that the principal gradientun derlying the da ta no longer corresponds directly to axis

1 (Minchin, 1987). Instead, this principal gradient is dis-tributed among axes 1 and 2, producing what is known asth e ar ch effect for Correspondence Analysis and t he h orse-shoe effect for Prin cipal Component s Analysis. This a rchis produced primarily when samples from the extremes ofaxis 1 contain such different assemblages that there is lit-

tle overlap in their taxonomic composition. DCA flattensthe Correspondence Analysis arch by dividing the archinto a series of segments and subtracting the m ean a xis 2

value for each segment from each score within tha t seg-ment . This procedure has th e side effect of tend ing to com-press faunal variation near the ends of axis 1. This un-wanted compression is then removed by rescaling axis 1scores such that there is a constant rate of turnover alongaxis 1. The end result is that axis 1 reflects the primarysource of ecological variation in the composition of faunas(for a m ore detailed explanation, s ee H ill an d G auch,1980). Axes 2 a nd 3 reflect a dditional sources of variat ionbeyond the principal gradient, but their int erpretation isoften much less clear because DCA is known to ta ke vari-at ion t hat should be expressed on axis 2 an d express someof it on a xis 3 (Kenk el a nd Orloci, 1986; Minchin, 1987).

PC-ORD, Version 3.0 (McCune and Mefford, 1997) wasused t o run DCA because of its simplicity of data entry,

culling, analysis, and plotting, as well as its a bility to han -dle large data sets. Default settings were used for rescal-ing axes (on), rescaling threshold (0), and number of seg-ments (26). The downweighting of rare taxa option wasused t o furt her minimize t he distortions they cause. Thisoption downweights in proportion to their abundance anytaxa rarer than Fmax/5, where Fmax is the frequency ofthe most common taxon. Taxa more abundant than Fmax/5 ar e un affected. Although only one of the DCA runs is re-ported her e, severa l others were per formed u sing differentsuites of samples or taxa, n ot using the rare taxon down-weighting option, and using a percent transform on thedat a. All of these pr oduced very similar r esults, indicat ingthe r obustness of the underlying structure in the dat a.


4/13

208 HOLLAND ET AL.

FIGURE 2Axis 1 and axis 2 taxon scores (n 46) from detrendedcorrespondence analysis (DCA), with inset of sample scores (n 1337) in upper right. Axis 1 is interpreted to reflect water depths, withlow scores corresponding to deeper water and high scores corre-sponding to shallower water. Axis 2 may reflects a gradient in lifehabit, with low scores dominated by epifaunal suspension feeding taxaand high scores dominated by infaunal to semi-infaunal deposit andsuspension feeders, as well as planktonic and nektonic forms. Ac:Acidaspis. Am: Ambonychia. As: Aspidopora. ca: calymenids. Ce:Ceraurus. Ci: Cincinnaticrinus. cm: cryptostomes. Co: Cornulites. cp:cephalopods. Cr: Cyclora. Cs: Craniops. Ct: Cryptolithus. Cy: Cyclo-

nema. Da: Dalmanella. Ec: Ectenocrinus. en: encrusting bryozoans.Es: Escharopora. fe: fenestellids. ga: gastropods. Gl: Glyptocrinus. gr:graptolites. hy: hydrozoans. Io: Iocrinus. Is: Isotelus. kb: thick bifoliatebryozoans. kr: thick ramose bryozoans. Le: Lepidocoleus. Me: Mero-crinus. Mo: Modiolopsis. nb: thin bifoliate bryozoans. nr: thin ramosebryozoans. nu: nuculoids. os: ostracods. Pa: Parvohallopora. Pe:Plectorthis. Pl: Platystrophia. Ps: Prasopora. Pt: Proetidella. Pu: Pseu-dolingula. Ra: Rafinesquina. sc: scolecodonts. So: Sowerbyella. Sp:Stomatopora. St: Strophomena. Sz: Schizocrania. Zy: Zygospira.

FIGURE 3Typical sample compositions from selected points alongaxis 1. Relative abundance coded as a (abundant), c (common), andr (rare); see text for explanation of these classes.

FIGURE 4Photographs of slabs illustrating common assemblages, generally ordered along DCA axis 1, from lowest on axis 1 (A) to highest(H). Scale bar in each photo represents 1 cm. A: Cincinnaticrinus (Ci). B: Ectenocrinus (Ec). C: Sowerbyella (So), Cincinnaticrinus (Ci), andDalmanella (Da). D: calymenid (ca), thin ramose bryozoans (nr), thin bifoliate bryozoans (nb), Cryptolithus(Ct), and Acidaspis(Ac). E: gastro-pods (ga) and thick ramose bryozoans (kr). F: Dalmanella (Da) and Glyptocrinus(Gl). G: Dalmanella(Da), thick ramose (kr) and thick bifoliate(kb) bryozoans. H: Rafinesquina (Ra).

DCA RESULTS

Biofacies

DCA simultaneously calculates Kope taxon scores andsample scores, and plots them on the same a xes with thesame scales (Fig. 2). Kope faunas vary continuously across

axis 1 without evidence of abrupt breaks in faun al compo-sition. Nonetheless, it is possible to examin e th e composi-tion of represent at ive samples a long t his a xis (Figs. 3, 4).At extrem ely low values (050) on a xis 1, assemblages aredominated by the crinoids Cincinnaticrinus (Fig. 4A) and

Ectenocrinus (Fig. 4B), and may conta in a variety of oth er

taxa in m uch low er abundance, s uch as tr ilobites , m a-chaeridians, ostracods, a nd cornulitids. At somewhathigher values (100150), an assemblage consisting of thebrachiopod S owerbyella and the trilobite Cryptolithus oc-curs (Fig. 4C, D). The S owerbyella-Cryptolithus assem-blage also commonly has a wide variety of other taxa en-

countered in lower abundance. At higher axis 1 values(200250), the S owerbyella -Cryptolithus assemblage givesway to a n assemblage dominated by the brachiopod Dal-manella (Fig. 4F). The trilobites, Flexicalymene a n d Grav-icalymene, as well as t he t rilobite Isotelus, and bryozoans(chiefly encrusting, thin ramose and thin bifoliate forms)all commonly occur in the Dalmanella association. At axis1 values of 300350, Dalmanella gives way t o the bra chio-pod Zygospira, a diverse association of bryozoans, and asmall form of the brachiopod Platystrophia (Fig. 4G). Atthe highest observed values (400430), assemblages aredominated overwhelmingly by the brachiopods Rafines-quina a n d Strophomena (Fig. 4H). Based on several linesof evidence (see below), Axis 1 is interpreted to reflect wa-ter depth.

P revious analyses and replicate analyses described

above indicate that this u nderlying stru cture is robust. Ina cluster analysis of the uppermost 8 meters of the Kope,as well as t he overlying Fa irview and Bellevue Forma tion,Diekmeyer (1998) found compa rable assemblages, includ-ing an Onniella ( Dalmanella)-crinoid-trilobite, Onniel-la-bryozoan, Zygospira-graptolite-trilobite, Zygospira-


5/13



6/13

210 HOLLAND ET AL.

TABLE 2Variability around a single axis 1 sample score of 200

Sample #K528.88

C DalmanellaR thick remose bryozoanR calymenidR IsotelusR Ectenocrinus

Sample #H214.58C calymenidsC thin ra mose bryozoanR S owerbyellaR Zygospira

R PseudolingulaR Aspidopora

R thick ramose bryozoanR Am bonychiaR Acidaspis

R Isotelus

R Cincinnaticrinus

R Ectenocrinus

Sample #K527.42C Isotelus

R Dalmanella

R cephalopod

Sample #B219.56C DalmanellaR S owerbyella

R CycloraR thin ra mose bryozoan

Sample #K515.24R S owerbyellaR Zygospira

R AspidoporaR thin bifoliate bryozoanR thick ramose bryozoan

R CornulitesR Cryptolithus

R Iocrinus

bryozoan, Rafinesquina-bryozoan-crinoid, and Platystro-phia-bryozoan assemblages. Sim ilarly, seriat ion (Browerand Kile, 1988) ha s been u sed pr eviously to reconstructthe principal un derlying gradient in Kope a ssemblages(Holland et al., 2000). The ordering of ta xa a long DCA axis1 is correlated highly with that found through seriation(Spearma ns r 0.79, p0.001).

Although t hese ar e r epresentative assemblages withinth e Kope, wide variat ion in a ssemblage compositions canbe found at a single axis 1 score. For example, when sev-eral sa mples with a n a xis 1 score of 200 are compared (Ta-

ble 2), the assemblages h ave common Dalmanella in somecases and none in others, despite having a sample scoreclose to Dalmanella s score of 180. Some of th ese sa mpleshave a diverse bryozoan assemblage whereas others haveno bryozoans. This variation at a single axis 1 score mayhave at least four origins. First, variation is produced insome cases by th e origina tion, extinction, or m igration ofataxon in the study area. For example, the crinoid Merocri-nu s is known to occur only in the lowermost Kope. Thus,even a t sample scores of 150 (the t axon score for Merocri-nu s) high in th e Kope, Merocrinus is absent. Second, somecompositional variation may result from geographic vari-ation within the st udy area tha t is not correlated with wa-ter depth. For example, some taxa may be consistently

more abundant in some ar eas th an others, for reasons un-correlated to wat er depth. Third, differences in assem-blage composition m ay r eflect tempora l changes in factors

not correlated with water depth. Fourth, stochastic varia-tion is almost certainly present in the data. All of thesesources of var iation ma y be expressed on a xes 2 and 3, al-th ough how th eir effects combine t o produce these a xes isnot obvious.

The variation expressed at a single axis 1 score reflectssubstantial variation in the composition of assemblages,suggesting t hat assemblage compositions are not con-tr olled rigidly. To some extent , taxa mu st ha ve been free toform assemblages of widely var ying compositions, a n ob-servat ion at odds with th e concept of ecological lockingan d rigidly defined commu nities with limited membersh ip(Brett et al., 1996).

Axis 2 appears to reflect differences in life habit andguild membersh ip. Almost all of th e ta xa with low (200)

axis 2 scores are sessile, low-tier epifaun al suspensionfeeders. High a xis 2 scores reflect t axa with a much widervariety of life h abits, including m obile taxa (several of thetrilobites, gastropods, cephalopods, ostracods), infaunalta xa (nuculoids, Modiolopsis, lingulids), deposit feeders(nuculoids, possibly t rilobites and gastr opods), car nivores(cephalopods, scolecodonts, hydrozoans), as well as plank-tic an d nekt ic forms (cephalopods, graptolites). This m ol-lusk-domina ted faun a with high axis 2 scores forms a dis-tinctive biofacies (Fig. 4E) tha t r ecurs at several horizonsth roughout t he Kope (Brett an d Algeo, 1999c).

DCA Axis 1 and Water Depth

As is common in m ar ine ecological da ta , axis 1 is corr e-lated with water depth, as su ggested by several lines of ev-idence. Several taxa with h igh axis 1 scoresStropho-mena, Rafinesquina, a n d Platystrophiaare known to bemore abun dan t in lithofacies shallower tha n th e Kope For-mation (e.g., the Fairview Formation, Tobin, 1982; Hol-land, 1993). These three genera lie at the extreme highend of axis 1, suggesting relatively shallow-water condi-tions. Because deeper water facies than the Kope ar e notexposed in Middle and Upper Ordovician rocks from theCincinna ti Arch, it is difficult to mak e a parallel argumentfor low values of axis 1. However, several taxa that haverelat ively low scores, su ch a s Cryptolithus a nd S owerbyel-la , have been interpreted to be relatively deep-water taxain the Ordovician of the Appalachian Basin (Bretsky,1970; Springer and Bambach, 1985). Similarly, the deep-water trilobite Triarthrus, too rare to be included in t his

ordination, als o occurred in s am ples w ith low axis 1scores.

As is true for most modern marine invertebrates, depthitself ra rely exerts direct cont rol on th e distribution of ta xa(see review in Holland, 2000). Instead, depth association isgoverned by physical and chemical factors that have astrong correlation with depth. Based on modern storm-dominated continental shelves, these factors in the Kopelikely include frequency of storm disturbance, substrateconsistency, t emperatur e, and oxygen levels, althoughother factors also may have contributed to the observeddepth r elationship. The type Cincinnat ian Series providesclear evidence that s torm dis turbance decreas ed intodeeper wat er, as would be expected (Tobin, 1982; J ennet te


7/13


FIGURE 5K445 section, including sequence stratigraphy, trends incycle thickness, and axis 1 scores from detrended correspondenceanalysis. Short lines next to measured section indicate boundaries ofmeter-scale cycles and numbered 20-m cycles. Central column is aFischer plot, except that points are placed at meter-scale cycle bound-aries, rather than being evenly spaced as in a conventional Fischerplot; this change accentuates the changes in cycle thickness. De flec-tions to the left indicate a series of thicker-than-average cycles, where-as deflections to the right indicate a series of thinner-than averagecycles. Heavy gray line in rightmost column indicates 21-point movingaverage used to highlight trends; value of 21 was found empirically tobalance minimizing local noise while preserving consistent trends.

FIGURE 6Systematic offset of axis 1 scores in depositionally updipHolst Creek outcrop and depositionally downdip K445 outcrop. Notethat Holst Creek is consistently offset towards higher values on DCA

Axis 1, as would be expected in a shallower water setting.

and Pryor, 1993). The increase in percent mudstone intodeeper facies in th e Cincinna tian also indicates t hat sub-str at e consistency was correla ted with wat er depth . Whenseveral variables are highly interrelated such that their

effects are difficult to tease apart, yet their combined ef-fects can be summarized by a single gradient, they areconsider ed collectively to be a complex gra dient (Cisne an dRabe, 1978). In ma rine set tings, depth commonly acts as acomplex gradient (Patzkowsky, 1995).

When plotted stratigraphically for the nearly completeKope section exposed a t K445, axis 1 sam ple scores sh ow aclear overall drift t owar ds higher values above 10 m, withinter vals of reversals to lower values (Fig. 5). This overalldrift is consistent with previous lithologically based int er-pretations of the Kope as overall shallowing-upward (To-bin, 1982; Holland, 1993; Jennett e and Pryor, 1993) andserves as a third line of evidence that axis 1 correlateswith water depth. Furthermore, the Fairview Formation

has higher axis 1 values than most of the Kope. This iscons istent w ith the w idespread interpretation that theF airview represents a s hallow er w ater facies th an theKope because of the F airviews lower sha le conten t, great -er proportion of proximal storm beds, greater number ofgrainstone beds, and the greater degree of abrasion, bor-ing, micritization, an d en crusta tion of fossils (Hay, 1981;Tobin, 1982; Holland, 1993; Jen nett e an d P ryor, 1993).

Superimposed on this long-term drift towards highervalues are several shifts toward lower values (at 39 m,2426 m, 3135 m, 5159 m). Some of these shifts corre-spond to a n interval of th icker-than-average meter-scalecycles that have been interpret ed as reflecting a relativerise in sea level (Jennette and Pryor, 1993; Holland et al.,1997). However, t rends in cycle t hickness and trends inaxis 1 scores gener ally do not para llel one a nother . As ac-commodat ion cha nges, it can be m anifested by changes insedimentation rate and rate of water depth change. Thetwo may not necessarily change in concert, a pattern seenhere. For example, under increasingly ra pid ra tes of ac-commodat ion, deepen ing m ay occur while cycles ma y getthicker or thinn er, depending on sedimentation rat es. Inthe case of the Kope, changes in cycle th ickness a nd a xis 1scores both provide information on changes in accommo-dation.

Finally, when axis 1 scores at K445 are compared tothose from the Holst Creek section, K445 scores are sys-tematically lower than Holst Creek scores (Fig. 6). These

two sections are hung on the Kope-Fairview contact, awell-defined horizon within a single meter-scale cycle thatcan be correlated thr oughout the region. This systematicoffset in a xis 1 scores confirms th at axis 1 corr elates withwater depth in that the depositionally downdip K445 sec-tion has consistent ly lower values tha n coeval strat a in thedepositionally updip Holst Creek section.

Although biofacies ana lysis normally would proceedwith a direct comparison of biofacies to lithology, such acompar ison her e reveals little in the way ofconsistent pat -terns (Fig. 5). F or example, t he limestone-rich intervalfrom 49 m coincides with a drift towards smaller axis 1scores, whereas the limestone interval from 6068 m re-flects increasing a xis 1 scores. Similar ly, some limestone-


8/13

212 HOLLAND ET AL.

FIGURE 7Modeling parameters of the distribution of taxa with re-spect to water depth. Preferred depth is equivalent to the mean of thedistribution, depth tolerance is equivalent to the standard deviation,and peak abundance is equivalent to the maximum probability withinthe distribution.

rich intervals (3033 m, 4850 m) correspond to locallyhigh a xis 1 values, wherea s other limestone-rich inter vals(1519 m, 2325 m) display no shift in a xis 1 values. Thislack of correspondence between axis 1 scores and litholog-ical trends also occurs on cross-plots not shown here,which display no linear t rends in axis 1 scores with r espectto either percent shale or percent limestone. This lack ofcorr espondence suggests t ha t Kope fau na s record en viron-ment al tr ends more sensitively tha n does lith ology.

AN ECOLOGICAL MODEL OF THE KOPE

Models of the stratigraphic distribution of fossils haveused bell-sha ped curves for each taxon tha t reflect the t ax-ons pr obability of occur rence with respect t o water depth(Holland, 1995a). Each curve is defined by three parame-ters (Fig. 7). Preferred depth (PD) is the depth at whichthe taxon is most abundant or most likely to be found.Depth tolerance (DT) is th e sta nda rd deviat ion of th e bell-shaped curve and reflects the degree to which the taxon isfound in other facies. Peak abundance (PA) is the proba-bility of collecting a taxon at its preferred depth and re-flects t he overall abunda nce of a ta xon.

These paramet ers can be estimated from modern taxa,where water depth an d abundan ce of taxa ar e known fromsystematic sampling. The shapes of abunda nce distribu-tions of many types of modern marine taxa suggest that abell-shaped curve is a reasonable first approximation ofth e distribut ion of taxa a long ecological gradient s; indeed,man y ecological models proceed from this appr oximat ion

(ter Braak and Gremmen, 1987; Holland, in press). Prob-ably the most common departure from this bell-shapedcurve seen in modern ta xa is a t endency to have an asym-metrical distribution with respect to depth, with a longertail into deeper water.

Determining these parameters for fossil taxa has notbeen at tempt ed previously becau se several obstacles havestood in the way, perhaps t he m ost serious of which wasthe lack of a quant itative measure ofwater depth. BecauseDCA axis 1 scores a re correlated with wat er depth, t heysolve th is problem and can be u sed as a proxy for waterdepth within the study interval.

Pr eferred dept h (PD) can be estimat ed directly from th eaxis 1 score for each taxon because DCA calculates the

ta xon score as th e abun dan ce-weighted avera ge of samplescores of the samples in which it is found, and vice versa(Hill and Gauch, 1980). Because DCA axis 1 sample and

taxon scores are therefore on the same scale, depth toler-an ce (DT) of a t axon can be calculat ed as t he sta nda rd de-viation of all axis 1 sample scores in which the taxon oc-curs. If a taxon occurs in samples arrayed over a widerange along axis 1, its standard deviation will be large,wherea s if th e ta xon occurs only in sa mples with a na rrowra nge of axis 1 values, its sta nda rd deviation will be small.Finally, peak abundance (PA) can be estimated by calcu-lating the percentage of samples in which a taxon occurswithin one DT of the PD for the taxon. This percentage isthen rescaled to find the true PA, that is, the probabilityofoccurrence a t the PD. This r escaling factor is simply theratio of the PA to the average probability of occurrencewithin one DT of the PD (see Holland, 1995b for equ at iondescribing probabilit y ofoccurr ence). For all but extrem ely

abundant taxa, this scaling factor is 1.186. This approachfor estimating PD, DT, and PA is similar to that used inplant ecology for describing the response of plant taxa toenvironm ental param eters (cf. ter Braak and Loom an,1986).

From the values of PD, DT, and PA, a modeled abun-dance curve can be calculated for each taxon in the Kope(Figs. 89; Table 3). Calculated values of PD, given by theaxis 1 scores for each taxon, range from 44 to 479 (in axis1 units, unscaled to depth in meters or other units). Someof th ese lie outside t he observed ra nge of sample scores (0to 434), indicating that these taxa are most abundant inshallower or deeper conditions than found in the Kope. Ac-tua l values of preferred depth may be lower th an 44 orhigher th an 479 for these t axa. Similarly, taxa whose PDis near t he edge of th e sampled portion of axis 1 also mightactually be found to lie farther out on axis 1 with morecomplete sam pling of sha llower and deeper wat er settin gs.

Values of DT ran ge from 2 to 125, with a mea n of 69. Af-ter the poorly sam pled brachiopod Schizocrania, whichhas a DT of 2, the next lowest value of DT is 46. Thus, ofthe relatively abundant components of the Kope fauna,DT values vary by a factor of 3. J ust a s sampling sha llowerand deeper water environments could cause some valuesof PD t o shift, it a lso might cause va lues of DT to increase,particularly for taxa that are eurytopic. Finally, given thelength of th e sampled gradient in t he Kope and th e typicalDT values, man y taxa ha ve some non-zero likelihood ofoc-curren ce in much of the Kope, a fact r eflected by th e wide-spread view of the Kope as ha ving a single faun a.

PA ra nges from 0.3% to 164%, with a m ean of 22%. Ob-

viously, the probability of occurrence of a taxon cannot ex-ceed 100%, but calculated values of PA tha t exceed 100%reflect ta xa that are so common that they are guara nteedto be found not only at t heir P D, but also to some extent inshallower a nd deeper settings. In t he Kope, only the cri-noids Cincinnaticrinus a n d Ectenocrinus and the ubiqui-tous brachiopod Rafinesquina have PA greater tha n 100%.Even t hough observed values of PA in t he Kope vary by al-most three orders of magnitude, values of PA for all taxaever reported from th e Kope must va ry by an even great eramount, because many of these taxa were too rare to beencountered in this study.

Alth ough th e Kope was deposited on a storm-dominatedramp, theoretical considerations and field data suggest


9/13


FIGURE 8Reconstructed ecological distributions of abundant Kopetaxa along axis 1. See text for construction of these curves. Note thatscale at bottom indicates actual sampled part of gradient; curves be-yond these limits are extrapolations and their accuracy is currentlyunknown.

tha t post-mort em shell tra nsport is minor. In offshore set-tings, storm-generated water movement consists of an on-shore-offshore wave component tha t results in little net

trans port of s edim ents and an isobath-parallel currentcomponent resulting from geostrophic flow t hat is th eprincipal agent of sediment transport (Duke, 1990). Al-though E kman turn ing of this geostrophic flow causes aslight (25) basinward transport of bedload, most largergrains a re moved only short distan ces dur ing storm events(Kachel and Smith, 1986). Fu rth ermore, multiple fieldstudies indicate that out-of-habitat transport is rare inlevel-bottom sublittoral environments, although many ormost individuals un dergo some within-habitat tra nsport(Kidwell and Bosence, 1991; Kidwell and Flessa, 1996).Because out-of-habitat shell tra nsport in t hese settingsgenerally involves sm all n umber s of individua ls (Kidwelland Bosence, 1991), transport is more likely to raise val-ues of DT tha n t o cau se shifts in PD values. Finally,Miller

(1997) found that small-scale faunal patchiness was pre-served in the shallower-water and more storm-influencedFairview Formation and that it could not be explained bypost-mortem transport.

DISCUSSION

The most important conclusion to be drawn from thisstud y is tha t su btle facies cont rol on t he distr ibution of fos-sils easily can go undetected unless it is sought specificallywith analyses designed for that purpose. The Kope haslong been treat ed a s a single lithofacies an d, as such, itsfauna s ha ve likewise been viewed simply as variat ions onan offshore biota. Instead, Kope faunas vary predictablywith inferred water depth to a much finer degree than canbe detected lithologically. As has been argued elsewhere,benthic faunas can be more sensitive indicators of envi-ronmental change t han lithofacies (Miller, 1988; Brett,1998).

Quantified changes in faunal abundance may be used ins om e cas es t o infer s equence architecture in a m annersimilar t o coenocorrelation (Cisne an d Rabe, 1978). For ex-ample, the stra tigraphic patt ern seen in the m oving aver-age of axis 1 sample scores (Fig. 5) matches the textbookw ell-log pattern of progradational s tacking in that i tshows a net drift towards shallower values, with well-de-fined shorter inter vals of deepening (i.e., flooding su rfac-es). Detrended correspondence analysis and other multi-variate techniques can supply a single proxy variable forw ater depth that can be us ed in this w ay for s equencestrat igraphic interpretat ions, pa rticularly in monotonous

offshore lithofacies such as the Kope. Furthermore, theoriginal data need not consist of actual counts of speci-mens. The data u sed here were based on a consistently ap-plied scheme of relat ive abun dan ce th at was lat er convert -ed to numerical values. Thus, even semiquantitative orrank data can be used to make these interpreta tions. Sim-ilar a pproaches ha ve been used for single ta xa or rat ios oftaxa (e.g., Sweet, 1979; Armentrout and Clement, 1991),but m ultivariate approaches h ave an advantage in thatthey combine informat ion from all of th e ta xa into a singlemetric.

Multivariate approaches a lso have the advanta ge of al-lowing the estimation of three parameters describing fos-sil distributions: preferred depth, depth tolerance, and


10/13

214 HOLLAND ET AL.

TABLE 3Calculated values of Preferred Depth, Depth Tolerance,and Peak Abundance.

Taxon

PreferredDepth

(PD)

DepthTolerance

(DT)

PeakAbundance

(PA)

Acidaspis

Am bonychia

Aspidopora

calymenidscephalopods

138.7169.9111.2162.2164.5

65.977.146.969.460.7

13.63.38.8

42.61.5

Ceraurus

Cincinnaticrinus

CornulitesCraniops

Cryptolithus

182.143.9

81.5133.1

68.0

59.671.570.182.355.7

0.4164.2

7.61.0

27.9cryptostomesCyclonema

Cyclora

Dalmanella

Ectenocrinus

179.6138.7185.2204.7

48.2

70.980.156.969.1

63.3

13.90.35.1

71.4

106.4encrusting bryozoans

Escharopora

fenestellidsgastropodsGlyptocrinus

169.2262.1310.7133.9

94.9

96.382.4

124.695.357.4

2.83.10.91.50.4

graptoliteshydrozoans

Iocrinus

Isotelus

Lepidocoleus

109.4369.7

84.7200.5

39.2

57.063.472.572.860.5

9.82.0

10.648.2

5.7Merocrinus

Modiolopsis

nuculoidsostracodsParvohallopora

112.8176.6202.3138.719.7

61.979.466.862.6

123.8

4.80.71.98.60.7

Platystrophia

Plectorthis

PrasoporaProetidella

Pseudolingula

408.9278.5

144.3125.6233.2

66.3101.7

49.049.753.1

24.41.0

0.50.61.5

RafinesquinaSchizocrania

scolecodontsS owerbyellaStomatopora

Strophomena

430.479.6

285.8102.012.3478.8

74.52.2

46.148.062.087.7

106.2*

2.746.9

4.414.8

thick bifoliate bryozoansthick ra mose bryozoansthin bifoliate bryozoansthin ramose bryozoans

Zygospira

405.0362.5282.4279.0300.7

70.969.376.772.569.0

3.763.4

9.759.773.0

* too few samples to be calculated; poor fit to normal distribution.

FIGURE 9Reconstructed ecological distributions of rarer Kope taxaalong axis 1. See Figure 8 caption for explanation. Note 10x exag-geration in vertical scale relative to Figure 8, necessary because thesetaxa are much rarer than those in Figure 8.

peak abun dance. The latter two of these parameters pr o-vide an additional layer of information in that they allowthe facies specificity an d abundan ce of each t axon to becalculat ed. In part icular , alth ough some ta xa already weresuspected to be more facies controlled t han others, esti-mation of depth tolerance allows this degree of control tobe quan tified. Similarly, estimation of peak a bunda nce al-


11/13


lows an estimate of the abundance of a taxon that is inde-

pend ent of facies effects.

The ecological m odel presented here, in which values of

PD, PA, and DT ar e calculated for a r an ge of ta xa, has sev-

eral potential uses. These values could be used t o search

for long-term strat igraphical, geographical, a nd macro-

ecological pattern s in the distribution, ecological toler-

ance, and abundan ce of taxa (cf. Brown, 1984; Brown,

1995). They could be used, in conjunction with axis 1 sam-

ple scores, to define fossil recovery potential functions to

place confidence limits on fossil ran ges, in which th e con-

fidence limits would account for changes in facies (Mar-

shall, 1997).

Improved estimat es of PD, PA, an d DT also can be used

to better constr ain m odels oft he fossil record (Hollan d and

Pa tzkowsky, 1999). For example, a conserva tive estimat e

of the ma ximum depth ra nge between th e shallowest and

deepest facies oft he Kope would be 65 m, equivalent to th e

present compacted thickness of the unit. This can be seen

if accommodation was held constant and the Kope simply

filled in the available accommodation while not compact-

ing: water depth would decrease by an amount equivalent

to the thickness of sediments deposited. However, this

simplistic appr oach neglects the opposing effects of com-

paction, which would have caused the Kope to have been

originally much thicker, and isostatic subsidence, which

would diminish t he tota l amount of sha llowing during the

deposition of the Kope. Backst rippin g to allow for compa c-

tion and isostatic subsidence suggests that the Kope re-

cords a t m ost 2540 m of sha llowing. Thu s, the dept h dif-

ference between t he highest (434) an d lowest (0) axis sam-

ple scores would equate to 2540 m. By t his scaling, the

average observed value of depth tolerance (69) would

equal a pproximately 46 m, over an order of magnitude

smaller than that used in Holland and Patzkowsky(1999),

suggesting th at facies effects on the strat igraphic distri-

bution of fossils ar e far more severe t han previously mod-

eled. Values of PA in t his stu dy are also lower th an previ-

ously modeled, suggesting that phenomena such as the

clust ering of first a nd la st occurr ences at flooding sur faces

and sequence boundaries may be masked by the overall

rar ity of taxa.

The existence of subtle, environment ally controlled var -

iations in the composition of fossil assemblages poses a

challenge to biostrat igraphic and paleobiologic studies.

Facies control itself is already well known. What should

now be appreciated is the degree to which it can be devel-

oped, how subtly it can be manifested, and how easily itcan escape detection. Targeting a single lithofacies in the

hope th at it will eliminate facies cont rol on fossil distribu-

tions may only reduce the magnitude of the problem. Par-

ticularly where facies belts are broad and encompass a

wide r ange of water depths, such a s offshore and deeper

envir onmen ts, subt le facies contr ol on fossils easily may go

undetected. To eliminate this type of facies control, sam-

ples should be collected from horizons with similar a xis 1

scores, tha t is, from similar positions a long th e depth gra -

dient. If th is type of subtle facies cont rol is common, it rais-

es the possibility that facies control affects more paleobio-

logic and biostrat igraphic interpreta tions than current ly

thought.

CONCLUSIONS

(1) Fossil assemblages within the Kope Formation dis-

play subtle environmental contr ol. This control is inter-preted to be dept h-related because of (A) the known faciesdistribut ion of some fossils near th e edges of DCA axis 1,(B) the correlation of DCA axis 1 to lithologically-baseddepth interpretations in vertical section, and (C) system-atic differences between DCA axis 1 sample scores fromdepositionally updip a nd downdip sections.

(2) A model of depth-related control of Kope faunas isproposed. This model describes th e preferred dept h, depthtolerance, and peak abundance of the 35 most commontaxa examined in this study, principally belonging to th ebrachiopods, br yozoans, trilobites, mollusks, an d echino-derms. Values of depth tolerance appear to be over an or-der of magnitude smaller tha n previously modeled, sug-gesting that facies effects on the stratigraphicdistribution

of ta xa (e.g., clustering of first a nd last occurrences atflooding surfaces) are more severe than previously mod-eled.

(3) The detection of facies control where none previouslyha d been documen ted r aises th e possibility th at facies con-trol can be more subtle than generally accepted. If so, fa-cies effects may exert a much greater influence on somebiostr at igraphic and paleobiologicp at tern s than previous-ly thought, par ticularly in offshore and deeper water fa-cies th at ar e difficult to su bdivide based on lithology.

ACKNOWLEDGMENTS

We appreciate the assistance of S. Diekmeyer and T.Reardon in the collection of th e field dat a. We tha nk T. Al-geo and another anonymous PALAIOS reviewer for their

reviews of this ma nuscript. This research was supportedby NSF Grants EAR-9204445 to S.M. Holland, and EAR-9204916 to A.I. Miller a nd D.L. Meyer.

REFERENCES

ANSTEY, R.L., and PERRY, T.G., 1973, Eden Shale bryozoans: A nu-merical st udy (Ordovician; Oh io Valley): Michigan Stat e Un iver-sity Paleontological Series, v. 1, p. 180.

ANSTEY, R.L., and RABBIO, S.F., 1990, Regional bryozoan biostrat ig-raphy and taphonomy of the Edenian Stratotype (Kope Forma-tion, Cincinnat i Area): Graphic correlat ion a nd gra dient an alysis:PALAIOS, v. 4, p. 574584.

ANSTEY, R .L., RABBIO, S . F . , a n d TUCKEY, M.E., 1987, Bryozoanbathymetric gradients within a Late Ordovician epeiric sea: Pa-leoceanography, v. 2, p. 165176.

ARMENTROUT, J.M., and CLEMENT, J.F., 1991, Biostratigraphic cali-

bration of depositional cycles: a case study in High IslandGal-vestonEast Breaks areas, offshore Texas: in ARMENTROUT,J.M.,and PERKINS, B.F., eds., Sequence stratigraphy as an explorationtool: Concepts and practices: Gulf Coast Section of SEP M, Hous-ton, p. 2151.

BAYER, U., and MCGHEE, G.R., 1985, Evolution in marginal epiconti-nent al basins: The role of phylogenetic and ecologic factors (Am-monite replacements in the German Lower and Middle Jurassic):in BAYER, U ., and SEILACHER, A., eds., Sedimentary and Evolu-tionary Cycles: Springer-Verlag, New York, p. 164220.

BRETSKY, P.W., 1970, Late Ordovician ecology of the central Appala-chians: Peabody Museum of Natura l History Bulletin, v. 34, p. 1150.

BRETT, C.E., 1998, Sequence stratigraphy, paleoecology, and evolu-tion: Biotic clues and responses to sea-level fluctuations: P A-LAIOS, v. 13, p. 241262.


12/13

216 HOLLAND ET AL.

BRETT, C.E., and ALGEO, T.J., 1999a, E vent beds and small-scale cy-cles in E denian to lower Maysvillian st rat a (Upper Ordovician) ofnorthern Kentucky: identification, origin, and temporal con-

straints: in ALGEO, T.J., and BRETT, C.E., eds., Sequence, cycle &event stratigraphy of Upper Ordovician & Silurian strata of theCincinna ti Arch region: 1999 Field Conference of the Great LakesSection of SEPM, Cincinnati, p. 6592.

BRETT, C.E., and ALGEO, T.J., 1999b, Sequence stratigraph y ofUpperOrdovician and Lower Silurian strata of the Cincinnati Arch re-gion: in ALGEO, T.J . , and BRETT, C.E., eds., Sequence, cycle &event stratigraphy of Upper Ordovician & Silurian strata of theCincinnati Arch region:1999 field conference of the Great LakesSection of SEPM, Cincinnati, p. 3446.

BRETT, C.E., and ALGEO, T.J., 1999c, Stratigraphy of the Upper Or-dovician Kope Formation in its type ar ea (northern Kent ucky), in-cluding a revised nomenclature: in ALGEO, T.J., and BRETT, C.E.,eds., Sequence, cycle & event stra tigraphy of Upper Ordovician &Silurian strata of the Cincinnati Arch region:1999 Field confer-ence of the Great Lakes Section of SEPM, Cincinnat i, p. 4764.

BRETT, C.E., IVANY, L.C., and SCHOPF, K.M., 1996, Coordinated st a-sis: An overview: Pa laeogeograph y, Palaeoclimatology, Pa laeoe-

cology, v. 127, p. 120.BROWER, J.C., and KILE , K.M., 1988, Seriation of an original data ma-

trix as applied to paleoecology: Lethaia, v. 21, p. 7993.BROWN, J.H., 1984, On the relationship between abundance and dis-

tribution of species: The American Natu ralist, v. 124, p. 255279.BROWN, J.H., 1995, Macroecology: University of Chicago Press, Chi-

cago, 269 p.CASTER, K.E., DALVE, E.A., and POPE , J.K., 1955, Elementary guide

to the fossils and st rat a of the Ordovician in t he vicinity ofCincin-nat i Ohio: Cincinnati Museu m of Natur al History, 47 p.

CISNE, J.L., and RABE, B.D., 1978, Coenocorrelation: Gradient analy-

sis of fossil communities and its applications stratigraphy: Le-tha ia, v. 11, p. 341364.

DALVE, E., 1948, The fossil faun a of the Or dovician in the Cincinnat i

region: Un iversity Museum, Depa rtmen t of Geology and Geogra-phy, Cincinnati, Ohio, 56 p.

DATTILO, B.F., 1996, A quan titat ive paleoecological a pproach to high-

resolution cyclic and event stra tigraphy: the Upper Ordovician

Miamitown Shale in the type Cincinnatian: Lethaia, v. 29, p. 2137.

DIEKMEYER, S.C.S.L., 1998, Kope to Bellevue Forma tions:t he RiedlinRoad / Mason Road site (Upper Ordovician, Cincinnati, Ohio re-gion): in DAVIS, R.A., and CUFFEY, R.J., eds., Sampling the layer

cake tha t isnt: The stra tigraphy a nd pa leontology of the type-Cin-cinnatian: Ohio Division of Geological Survey Guidebook No. 13,Columbu s, p. 1035.

DUKE , W.L., 1990, Geostrophic circulation or shallow marin e tu rbid-ity currents? The dilemma of paleoflow patterns in storm-influ-enced prograding shoreline systems: Journal of Sedimentary Pe-

trology, v. 60, p. 870883.HARRISON, W.B., and MAHAN, T.K., 1981, Lower Cincinna tian Kope

Formation: in ROBERTS, T.G., ed., GSA Cincinnat i 81 field trip

guidebooks: American Geological Institute, Falls Church, p. 3645.

HAY, H.B., 1981, Lithofacies and forma tions of the Cincinnatian Se-

ries (Upper Ordovician), southeastern Indiana and southwestern

Ohio: Unpu blished Ph.D. Dissertat ion, Miami Un iversity,Oxford,Ohio, 236 p.

H ILL, M.O., and GAUCH, H.G., J R., 1980, Detrended correspondenceanalysis: an improved ordination technique: Vegetatio, v. 42, p.4758.

HOLLAND, S.M., 1993, Sequence stratigraphy of a carbonate-clasticramp: The Cincinnatian Series (Upper Ordovician) in its typear ea: Geological Society of America Bulletin , v. 105, p. 306322.

HOLLAND, S.M., 1995a, Depositional sequences, facies control, andthe distribution of fossils: in HAQ, B.U., ed., Sequence stratigra-phy and depositional response to eustatic, tectonic and climatic

forcing: Kluwer Academic Publishing, Dordrecht, p. 123.HOLLAND, S.M., 1995b, The stratigraphic distribution of fossils: Pa-

leobiology, v. 21, p. 92109.

HOLLAND, S.M., 2000, The quality of the fossil recorda sequencestratigraphic perspective: in Erwin, D.H., and Wing, S.L., eds.,

Deep time: Paleobiologys perspective: Paleontological Society,Lawrence, p. 148168.

HOLLAND, S.M., MEYER, D.L., AND MILLER, A.I., 2000, High-resolu-

tion correlation in apparently monotonous rocks: Upper Ordovi-cian Kope Formation, Cincinnati Arch: PALAIOS, v. 15, p. 7380.HOLLAND, S.M., MILLER, A.I., DATTILO, B.F., MEYER, D.L., and DIEK -

MEYER, S.L., 1997, Cycle ana tomy and variability in the storm-dominated type Cincinnat ian (Upper Ordovician):Coming to gripswith cycle delineation and genesis: Journa l of Geology, v. 105, p.135152.

HOLLAND, S.M., MILLER, A.I., and MEYER, D.L., 1999, Sequence stra -tigraphy of the Kope-Fairview interval (Upper Ordovician, Cincin-nati, Ohio area): in ALGEO, T.J., and BRETT, C.E., eds., Sequence,cycle & event str atigraph y of Upper Ordovician & Silurian st rat aof the Cincinnat i Arch region: 1999 field conference of the GreatLakes Section of SEPM, Cincinnati, p. 93102.

HOLLAND, S.M., and PATZKOWSKY, M.E., 1999, Models for simulatingth e fossil record: Geology, v. 27, p. 491494.

JENNETTE , D.C., and PRYOR, W.A., 1993, Cyclic alternation of proxi-mal and distal storm facies: Kope and Fairview Formations (Up-per Ordovician), Ohio and Kent ucky: Journa l of Sedimentar y Pe-

trology, v. 63, p. 183203.KACHEL, N.B., and SMITH , J.D., 1986, Geological impact of sediment

transporting events on th e Washington continental shelf: inKNIGHT, R . J . , a n d MCLEAN, J .R ., eds . , Shel f s ands and s and-stones: Canadian Society of Petroleum Geology, p. 145162.

KENKEL, N.C., and ORLOCI, L., 1986, Applying met ric and nonmetr icmultidimensional scaling t o ecological st udies: Some n ew results:Ecology, v. 67, p. 919928.

KIDWELL, S.M., and BOSENCE, D.W.J., 1991, Taphonomy and time-averaging of marine shelly faunas: in ALLISON, P.A., and BRIGGS,D.E.G., eds., Taphonomy: Releasing t he Da ta Locked in the Fossil

Record: Plenum Press, New York, p. 115209.KIDWELL, S.M., and FLESSA, K.W., 1996, The qua lity of the fossil re-

cord: Populations, species, and communities: Annual Review of

Ear th and Planet ary Sciences, v. 24, p. 433464.KREISA, R.D., DOROBEK, S.L., ACCORTI, P.J., and GINGER, E.P., 1981,

Recognition of storm-generated deposits in the Cincinnatian Se-

ries, Ohio: Geological Society of America Abstracts with Pro-

grams, v. 13, p. 285.LORENZ, D.M., 1973, Edenian (Upper Ordovician) benthic commu nity

ecology in north-central Kentucky: Dissertation Abstracts Inter-nat ional, v. 34B, p. 2816.

MARSHALL, C.R., 1997, Confidence interva ls on str atigraph ic ranges

with nonran dom distributions of fossil h orizons: Pa leobiology, v.23, p. 165173.

MCCUNE, B., and MEFFORD, M.J., 1997, PC-ORD. Multivariate Anal-

ysis of Ecological Data, Version 3.0: Gleneden Beach, Oregon,MjM Software Design.

MILLER, A.I., 1988, Spatial resolution in subfossil molluscan remains:

Implications for paleobiological an alyses: Pa leobiology, v. 14, p.91103.

MILLER, A.I., 1997, Count ing fossils in a Cincinna tian storm bed: spa-

tial resolution in the fossil record: in BRETT, C.E., and BAIRD, G.C.,eds., P aleontological event s: St rat igraphic, ecological, and evolu-tionary implications: Columbia University Press, New York, p.

5772.

MILLER, A.I., HOLLAND, S.M., DATTILO, B.F., and MEYER, D.L., 1997,Stratigraphic r esolution and perceptions of cycle architecture:

Variations in meter-scale cyclicity in the type Cincinnatian Series:Journal of Geology, v. 105, p. 737743.

MINCHIN, P.R., 1987, An evaluation of the relative robustness oftech-

niques for ecological ordination: Vegetatio, v. 69, p. 89107.PATZKOWSKY, M.E., 1995, Gradient analysis of Middle Ordovician

brachiopod biofacies: Biostratigraphic, biogeographic, and macro-

evolutionary implications: PALAIOS, v. 10, p. 154179.SHELDON, P.R., 1996, Plus Aa chan gea m odel for stasis a nd evolu-

tion in different environments: P alaeogeography, P alaeoclimatol-

ogy, Pala eoecology, v. 127, p. 209228.SPRINGER, D.A., and BAMBACH, R.K., 1985, Gradient versus cluster

analysis of fossil assemblages: A comparison from the Ordovician

of south western Virginia: Lethaia, v. 18, p. 181198.SWEET, W.C., 1979, Conodonts and conodont biostratigraphy of post-


13/13


Tyrone Ordovician rocks of the Cincinnati region: United StatesGeological Survey Professional Paper, v. 1066-G, p. 126.

TER BRAAK, C.J.F., and GREMMEN, N.J.M., 1987, Ecological ampli-

tudes of plant species an d t he int erna l consistency of Ellenbergsindicator values for moisture: Vegetatio, v. 69, p. 7987.TER BRAAK, C.J.F., and LOOMAN, C.W.M., 1986, Weighted averaging,

logistic regression and t he Gau ssian response model:Vegetatio,v.65, p. 311.

TOBIN, R.C., 1982, A model for cyclic deposition in the CincinnatianSeries of southwestern Ohio, n orthern Kentucky an d southeast-ern Indiana: Unpublished Ph.D. Dissertation, University of Cin-cinna ti, 483 p.

TOBIN, R.C., 1986, An assessment of the lithostra tigraphic and inter -pretive value ofth e tr aditional biostra tigraphyof the type UpperOrdovician ofNorth America: American Journ al ofScience, v.286,p. 673701.

WEIR, G.W., PETERSON, W.L., SWADLEY, W.C., an d POJETA, J., 1984,Lithostratigraphy of Upper Ordovician strata exposed in Ken-tucky: United States Geological Sur vey Pr ofessional P aper, v.1151-E, p. 1121.

WEISS, M.P., and SWEET, W.C., 1964, Kope F ormation (Upper Ordo-

vician)Ohio and Kentucky: Science, v. 145, p. 1296, 13011302.WILSON, M.A., 1985, Distu rba nce an d ecologic succession in an Upper

Ordovician cobble-dwelling hardground fauna: Science, v. 228, p.575577.

ACCEPTE D DECEMBER 14, 2000

APPENDIX: LOCALITY DESCRIPTIONS

Hum e: Roadcut a long west side ofUS Route 42, 0.3 mile north ofin-tersection with Kent ucky State Route 14 at town of Hum e near inter-

section with Kent ucky State Route 338, just south of Beaverlick, Ken-tucky. Verona, KY 7 1/2 quadrangle. 38 50 48 N, 84 43 00 W.

Holst Creek: Series of roadcuts on northeast side of AA Highway

(Kentucky State Route 9) at intersection with Kentucky State Route1019 and proceeding downhill to the northwest for 0.6 miles to inter-section with Holst Creek Road, south of Foster, Kentucky. Moscow,OH-KY 7 1/2 quadrangle. 38 46 33 N, 84 12 20 W.

K445 Composite. Composite outcrop consist s offour sections: K445,CON1, CON2, and CON3. K445: Roadcut on both sides of KentuckyState Route 445, 0.2 km west of intersection with Kentucky StateRoute 8, immediately northwest of the I-275 bridge over the Ohio Riv-er near Old Coney Amusement Park. Newport, KY-OH 7 1/2 quan-drangle. 39 03 22 N, 84 26 10 W. CON1: First roadcut on north-west side of westbound I-275, 0.5 km southwest of intersection of I-275 and the Kentucky bank of Ohio River near Old Coney Amuse-ment Par k. Newport, KY-OH 7 1/2 quadrangle. 39 03 15N, 84 2620 W. CON2: Second roadcut on north west side of westbound I-275,0.6 km southwest of intersection of I-275 and the Kentucky bank ofOhio River near Old Coney Amusement Park. Newport, KY-OH 7 1/2 quadrangle. 39 03 13 N, 84 26 24 W. CON3: Third roadcut onnorthwest side of westbound I-275, 0.8 km southwest of inter sectionof I-275 and t he Kentu cky bank of Ohio River near Old ConeyAmuse-

ment Par k. Newport, KY-OH 7 1/2 quadrangle. 39 03 10N, 84 2630 W.

Aurora: Hillside exposure behin d th e River Creek Village shoppingplaza at int ersection ofWilson Creek Road an d US Route 50,2.3 milessouth of intersection of US Route 50 and Indiana Route 48 at thebridge over Tann ers Cr eek. Aurora, IN-KY 7 1/2 quadrangle. 39 0437 N, 84 53 45 W.

Miamitown:Composite of two sections. (1) Roadcut along south andnorth sides of eastbound off-ram p from I-74 at exit #7 for Ohio StateRoute 128, south of Miamitown, Ohio. Addyston, OH-KY 7 1/2 quad-rangle. 39 12 21 N, 84 42 34 W. (2) Roadcut a long n orth side ofwestbound I-74, at end of onramp at exit for #7 for Ohio State Route128. Addyston, OH-KY7 1/2 quadrangle. 39 12 28N, 84 42 43W.

Documents

Ordovician Kope formation Ohio