Associations of Vertebrate Skeletal Concentrations and...

[The Journal of Geology, 2000, volume 108, p. 131–154] q 2000 by The University of Chicago. All rights reserved. 0022-1376/2000/10802-0001$01.00

131

ARTICLES

Associations of Vertebrate Skeletal Concentrations and DiscontinuitySurfaces in Terrestrial and Shallow Marine Records:

A Test in the Cretaceous of Montana

Raymond R. Rogers and Susan M. Kidwell1

Department of Geology, Macalester College, 1600 Grand Avenue, St. Paul, Minnesota 55105, U.S.A.(e-mail: [email protected])

A B S T R A C T

Although lags of bones and teeth are commonly cited criteria for marine unconformities, the consistency of theassociation of vertebrate fossils and discontinuity surfaces, as well as the taphonomic (postmortem) controls on thisrelationship, are poorly understood. A field test across fluvial, paralic, and shallow marine facies in the CampanianTwo Medicine and Judith River formations of Montana indicates that the distribution of vertebrate skeletal concen-trations is poorly correlated with the inferred durations of erosional and omissional hiatuses. Instead, vertebrateconcentrations associated with discontinuities of all durations tend to be patchy and closely track the abundance offossil material in underlying and lateral facies. Based on the analysis of 83 measured sections, we found first thaterosional bases of channels and minor scours within channels yield vertebrate lags; tidally influenced fluvial depositsare more productive than are “upland” fluvial deposits. Second, erosional shoreface ravinements and their correlativetransgressive marine flooding surfaces (fourth-order sequence boundaries) have well-developed vertebrate lags onlyalong segments that cut across older shoreface deposits. Third, a nonerosional, widely traceable discontinuity, whichis interpreted as the nonmarine extension of a 75.4-Ma third-order transgressive surface, is completely lacking invertebrate concentrations. Despite being unfossiliferous itself, this discontinuity does mark a regional change in therichness of the vertebrate fossil record, with overlying beds characterized by a much greater abundance of skeletalmaterial. Fourth, a laterally extensive set of erosional surfaces, embedded within multistory fluvial sandstone sheets,is the nonmarine extension of an 80-Ma third-order sequence boundary in the marine basin and lacks vertebrateconcentrations. The strong dependence of vertebrate lag development on preexisting local sources of skeletal materialrather than on the magnitude of the erosional vacuity or the duration of the hiatus contrasts with skeletal concen-trations of invertebrates in marine successions, where exhumation is generally much less important than the pro-duction of new elements during the hiatus. These findings provide a guide to prospecting productive fossil horizonsin terrestrial records and underscore fundamental differences in the ways in which bioclastic material accumulatesin terrestrial and shallow marine settings, the qualities of paleobiologic data derived from such concentrations, andthe relative reliabilities of skeletal material as cues to stratigraphically significant discontinuities.

Introduction

Lags of bones and teeth are classic criteria for rec-ognizing marine unconformities (e.g., Krumbein1942), but little attention has been focused eitheron determining how systematic this relationship isin marine or terrestrial settings or on the relation-

Manuscript received June 14, 1999; accepted November 9,1999.

1 Department of Geophysical Sciences, University of Chi-cago, 5734 South Ellis Avenue, Chicago, Illinois 60637, U.S.A.;e-mail: [email protected].

ship between vertebrate fossils and discontinuitysurfaces in general. For example, although verte-brate hardparts should be durable to the rigors ofreworking, not all unconformities are mantled bysuch debris and, conversely, not all vertebrate con-centrations mark significant hiatuses. How consis-tently are widespread erosional unconformities ac-companied by relative concentrations of vertebrateskeletal material? Does the taphonomic signature(postmortem features) of these concentrations dif-fer significantly from those associated with ero-

132 R . R . R O G E R S A N D S . M . K I D W E L L

sional and omissional hiatuses of shorter durationand extent? In general, how strong an influence dopatterns of physical sedimentation—both across fa-cies tracts and through stratigraphic cycles—exerton the distribution and taphonomy of vertebratefossils?

Empirical data from the marine stratigraphic rec-ord demonstrate a strong association between ben-thic macroinvertebrate concentrations and discon-tinuity surfaces of many types, as well as con-gruence between taphonomic attributes and the in-ferred duration of the associated hiatus (Kidwell1991, 1993a, 1993b; Brett and Baird 1993; Brett1995; Garcia and Dromart 1997; Naish and Kamp1997; Abbott 1998; Gillespie et al. 1998; Kondo etal. 1998). It is thus reasonable to suspect that phys-ical patterns of sedimentation, specifically lateraland temporal variation in erosion, sedimentaryomission, and deposition, might exert a similarlystrong influence on the distribution and quality ofvertebrate fossil assemblages in both marine andterrestrial systems. Aepler and Reif (1971), Reif(1982), and Kidwell (1989) have published briefstratigraphic characterizations of some marinebone beds, and Behrensmeyer (1987, 1988) and Rog-ers (1993) have described how terrestrial bone as-semblages vary with subsidence-related shifts in al-luvial architecture. Still, the distribution andquality of vertebrate fossil assemblages in relationto physical patterns of sedimentation remainpoorly understood.

Here we present the results of a field test acrossa transect of fluvial, paralic, and nearshore marineenvironments in the Campanian Two Medicine andJudith River formations of Montana. The focus ison links between relative concentrations of skeletaldebris and the discontinuity surfaces essential tosequence analysis. This record is especially advan-tageous both because it is richly fossiliferous, thusgiving maximum likelihood of finding associations,and because an independent evaluation of discon-tinuity surfaces and their potential for sequencestratigraphic analysis has already been completed(Rogers 1994, 1998).

Expected Patterns in the Distribution ofVertebrate Fossils

In terms of the concentration and preservation ofvertebrate or invertebrate hardparts, intervals ofnondeposition and erosion are double-edged swords(Kidwell 1986). On the positive side, nondeposi-tional hiatuses create conditions of low sedimen-tary dilution for the “rain” of skeletal material pro-duced by contemporaneous vertebrate populations

and, thus, should foster relative concentrations ofvertebrate material by purely passive means (e.g.,along a soil surface during an interval of slow land-scape aggradation). The longer the period of theseconditions, the greater the absolute quantity ofhardparts supplied and, thus, the richer or laterallymore continuous the ultimate skeletal concentra-tion may be. Erosion (negative net sedimentation)augments the effect by preferentially removing sed-imentary matrix, leaving behind a true hydrauliclag of larger and denser skeletal material exhumedfrom previous deposits. Theoretically, the moredeeply erosion cuts into older strata, the greater thevolume of previously deposited vertebrate materialthat can be intersected and concentrated into lagform.

On the negative side, conditions of low or zeronet sedimentation increase the period that any co-hort of skeletal material is subject to destructivepostmortem processes operating at or near the dep-ositional interface. In terrestrial systems, these pro-cesses include trampling, scavenging, weathering(UV exposure, oxidation, fungal and other micro-bial attack, freeze-thaw), detrimental soil pro-cesses, and exhumation-burial cycles (e.g., via bed-form migration within channels; Behrensmeyer1991). Erosional downcutting during the hiatusmay simply aggravate already difficult conditionsfor preservation. Erosion could potentially increasethe local supply of hardparts to the discontinuitysurface through exhumation, but any cohort ofskeletal material will probably undergo a largernumber of reworking events than it would duringpurely aggradational conditions. Moreover, erosionitself represents an additional agent of skeletalbreakdown and transport out of the local system,and exhumed skeletal material may be less durablerather than equally or more durable than newlyproduced hardparts, so that exhumation has littlenet positive effect on hardpart supply. It is thuspossible that taphonomic processes may be moresevere during hiatuses (or at least during long hi-atuses or during erosional rather than purely omis-sional hiatuses) than during “background” times,so that any advantages of lowered dilution and/orinput of exhumed hardparts may be outweighed bytaphonomic culling.

If newly produced hardparts are relatively dura-ble under prevailing climatic and edaphic condi-tions, then skeletal material from local mortalitymay accrue to appreciable quantities during phasesof nondeposition or very slow net sedimentation(e.g., on stable soil surfaces or in laterally accretingchannel belts held at grade). The ultimate assem-blage that mantles the discontinuity would be time

Journal of Geology V E R T E B R A T E S K E L E T A L C O N C E N T R A T I O N S 133

averaged to a greater or lesser degree, depending onthe duration of the hiatus, but would date to theperiod of the hiatus itself (i.e., be a true hiatal con-centration, sensu Kidwell 1991). Local pocketsmight preserve census assemblages (e.g., mass mor-talities from drowning or other ecologically briefmortality events), but in general the material willbe time averaged. The longer the hiatus, the greaterthe opportunity for ecologically heterogeneous as-semblages (i.e., the mixing of noncontemporaneouscommunities or community states rather thansimply the mixing of generations from singlecommunities).

An alternative scenario, which is not mutuallyexclusive with the first, is that the bulk of skeletalmaterial is exhumed from underlying and lateralfacies during incision, planation, or both. Exhumedmaterial might already be in a prefossilized state(permineralized), which would presumably in-crease its durability. Skeletal material would be en-tirely older than the hiatal surface. The more severethe downcutting, the greater the age differential be-tween material in the lag and the hiatus that con-centrated it, and the lower the relevance of its tax-onomic composition to paleoenvironmental con-ditions during the hiatus. Taphonomic artifacts ofexhumation might include abrasion and rounding(although this modification feature is not diag-nostic), polish (A. K. Behrensmeyer, pers. comm.),angular as opposed to spiral breakage patterns (Mor-lan 1984), and environmentally mixed assemblages.Sedimentological features consistent with exhu-mation include stratigraphic evidence of incisionand exotic sedimentary matrices adhering to ex-humed skeletal debris.

The ultimate nature of the vertebrate fossil rec-ord, and in particular the association of vertebrateskeletal material with discontinuity surfaces, de-pends on the relative importance of these two con-tradictory aspects of low to negative sedimenta-tion—low dilution and the exhumation ofadditional supply on the one side and prolongedexposure and perhaps elevated rates of hardpart de-struction on the other. The accumulation of bonesand teeth in nonmarine conditions may follow sev-eral different scenarios, depending on the relativeimportance of hardpart sources (hiatal productionvs. exhumed debris) and the durabilities ofhardparts.

Stratigraphic and Taphonomic Framework

The Campanian Two Medicine and Judith Riverformations are widely exposed across much ofnorthwestern and north-central Montana and are

characterized by the intertonguing of marine andnonmarine strata at several scales, ranging fromdecimeter-scale transgressive-regressive cycles tothird-order depositional sequences (Rogers 1998;fig. 1). The ∼545-m-thick Two Medicine Formationaccumulated during two major cycles of the West-ern Interior Seaway (R7-T8 and R8-T9 of Kauffman1977) and comprises the proximal alluvial facies oftwo eastward-thinning clastic tongues. It consistspredominantly of pedogenically modified gray andgray-green silty claystone and siltstone with widelyspaced beds of fine- to medium-grained sandstone.At least 19 discrete bentonite beds are intercalatedin the type area of the Two Medicine Formation(Rogers et al. 1993). The ∼180-m-thick Judith RiverFormation comprises the distal reaches of a singleeastward-thinning clastic tongue of alluvial,coastal plain and shallow marine sediments thataccumulated during regression of the Claggett Sea(R8) and subsequent transgression of the BearpawSea (T9). It is a heterolithic composite of gray andtan silty claystones, siltstones, and fine- to me-dium-grained sandstones of fluvial, tidal, and shal-low marine origin, with subordinate beds of darkgray to black lignite and tan to orange ironstone.Several bentonite beds are also intercalated withinthe Judith River section in the type area (Rogers1995; Rogers and Swisher 1996). The Judith RiverFormation correlates to the west across the ero-sional Sweetgrass arch with nonmarine deposits ofthe middle and upper Two Medicine Formation andis bounded above and below by marine shales ofthe Bearpaw and Claggett formations (fig. 1).

Two major stratigraphic discontinuities subdi-vide the largely nonmarine Two Medicine–JudithRiver record into regressive and transgressive de-positional systems (fig. 1). The lower of these twodiscontinuities, SB1, consists of a laterally exten-sive set of amalgamating erosional surfaces em-bedded within multistory fluvial sandstone sheetsand is considered to be the nonmarine (upstream)extension of a type 1 sequence boundary (Rogers1994, 1998). It correlates with a regionally traceablemarine disconformity embedded within the moredistal Eagle Formation (Hanson and Little 1989) andother marine units and formed in response to a dra-matic fall in relative sea level that occurred in theWestern Interior Basin at approximately 80 Ma(DeGraw 1975; Shurr and Reiskind 1984; Van Wag-oner et al. 1990). Shallow marine and paralic strataimmediately overlying the SB1 sequence boundaryaccumulated during the subsequent transgressionof the Claggett Sea (T8). These facies preserve a setof at least nine bentonite horizons that correlatestratigraphically and radioisotopically with the ma-


Figure 1. Schematic cross section of the Two Medicine–Judith River study interval restored across the Sweetgrassarch (SGA). Discontinuities referred to in the text (SB1, SB2, D1, D2, D3, D4) are indicated, as are the major third-order transgressive and regressive cycles of the Western Interior Seaway (Kauffman 1977). Changes in the density andstacking of fluvial channel sandstones across the study interval are schematically portrayed by the spacing of U-shaped channel symbols. All units, aside from those italicized, are included within the Montana Group. Modifiedfrom Gill and Cobban (1973).

rine Ardmore bentonite bed. Radioisotopic datingof two bentonites and a crystal tuff from this in-terval has yielded dates of Ma,79.6 5 0.1 79.8 5

Ma, and Ma, respectively (Rogers et0.1 80.0 5 0.1al. 1993). The shift from transgressive (T8) to re-gressive (R8) phases in the Two Medicine record isinterpreted to coincide with an observed shift tofacies more typical of inland fluvial and floodplainsettings. During the early stages of R8, the alluvialrecord of the Two Medicine Formation is domi-nated by fine-grained interchannel facies. Thechannel/floodplain ratio increases upsection with-in the nonmarine equivalent of the highstand dep-ositional system, and concurrent with this increasein sandstone abundance is a shift to more amal-gamated sheet geometries (fig. 1). The late high-stand depositional system in the more distal JudithRiver Formation shows a comparable record of fa-cies architecture, with the alluvial record domi-nated by relatively thick sandstone bodies withsheet geometries signifying a decrease in the rateof addition of accommodation (Rogers 1998).

The upper discontinuity, SB2, is a nonerosional,widely traceable discontinuity that is marked bythe abrupt appearance of lacustrine-carbonate de-

posits in upland areas (Two Medicine) and by asharp and regionally traceable decrease in channeldensity in the coastal plain (Judith River). It restsonly a few meters beneath a bentonite bed datedat 75.4 Ma in the Judith River Formation type area(Rogers and Swisher 1996) and is interpreted to cor-relate with the widespread addition of accommo-dation that coincided with the turnaround from theClaggett regression (R8) to the Bearpaw transgres-sion (T9) in the marine basin (fig. 1). In the easternpart of the Judith River Formation type area (fig. 1;Missouri Breaks), SB2 converges with a shorefaceravinement (D1 surface) at the base of three back-stepping fourth-order sequences that accumulatedduring the Bearpaw transgression. Bounding sur-faces of these fourth-order sequences (D1–D4 in fig.1) are defined by facies juxtapositions indicative ofabrupt deepening (flooding surfaces): the lower fewmeters of each sequence comprises a deepening-upinterval, which is why these packages are not cat-egorized as parasequences (Rogers 1995). Whereshoreface facies are superimposed immediately onthese flooding surfaces, there is clear evidence forerosion. Post-Cretaceous erosion over the Sweet-grass arch hinders our tracking of compara-


ble fourth-order (or higher order) transgressive-regressive cycles and their bounding surfaces to thewest in strata above the D4 surface, although theirpresence would be predicted.

Both the Two Medicine and Judith River for-mations are richly fossiliferous and contain a va-riety of skeletal concentrations set in a backgroundof dispersed skeletal material (Rogers 1993, 1995).These concentrations include dinosaur nests andnesting horizons (Horner and Makela 1979; Horner1982, 1994), dinosaur bone beds dominated by largeelements (Hooker 1987; Rogers 1990, 1995; Fiorillo1991; Varricchio 1995), and diverse vertebrate mi-crofossil localities (microsites) of both terrestrialand marine origin (Sahni 1972; Case 1978; Carranoet al. 1995; Rogers 1995). The vertebrate microsites,which typically consist of small physicochemicallyresistant elements such as teeth, scales, scutes, andcompact bones (phalanges, vertebrae), are particu-larly abundant. Loosely to densely packed fresh-water and marine shell beds crop out in the studyinterval as well (Stanton and Hatcher 1905; Sahni1972; S. M. Kidwell and R. R. Rogers, unpub. data).

One of the interesting taphonomic features of thestudy interval is that vertebrate skeletal debris andinvertebrate shell material commonly co-occur inboth terrestrial and marine concentrations, con-trary to the segregated patterns that are usually de-scribed from the fossil record. Another interestingtaphonomic aspect is the gradient in the quality ofvertebrate skeletal occurrences relative to proximal(Two Medicine) and distal (Judith River) positionsin the clastic wedge. The proximal record is dom-inated by high-resolution assemblages that showminimal evidence of time averaging, such as mono-or paucispecific dinosaur bone beds and nest sites,whereas the distal record is dominated by verte-brate microsites with high probable levels of timeaveraging (Rogers 1993). Still another large-scale ta-phonomic pattern in the study interval is variationin the abundance of vertebrate skeletal material rel-ative to third-order regressive-transgressive cycles.The vast majority of vertebrate skeletal material inthe Two Medicine and Judith River formations oc-curs in the transgressive record, particularly in thenonmarine facies equivalents of the T9 marine cy-cle (Rogers 1995).

Taphonomic Signatures ofDiscontinuity Surfaces

To test for relationships between taphonomic sig-nature and sedimentary hiatuses, stratigraphic sur-faces need to be ranked for hiatal length, using in-dependent physical stratigraphic criteria. In marine

records, Kidwell (1993b) ranked surfaces (fromshortest to longest) as (1) bedding planes and bed-set boundaries; (2) parasequence boundaries (flood-ing surfaces; the stratigraphically lowest floodingsurface within the transgressive systems tract is thetransgressive surface sensu stricto); (3) transgres-sive ravinements (landward segment of floodingsurface that includes an erosional vacuity createdby migration of the shoreface); (4) midsequencestarved intervals (omissional “surface of maximumtransgression” or “maximum flooding surface”;this is the master downlap surface within the dep-ositional cycle); and (5) sequence boundariesmarked by subaerial erosion (third order). Com-posites of the surfaces are possible, for example,transgressive ravinements that converge with se-quence boundaries over part of their extent. Thisranking assumes a general correspondence betweenthe lateral extent of the physical discontinuity sur-face and the temporal duration of the hiatus (atleast in settings of relatively broad diastrophism)and reflects the hierarchical reasoning of Campbell(1967) and the sequence stratigraphic terminologyof others (e.g., Van Wagoner et al. 1990; Mitchumand Van Wagoner 1991).

Hierarchical classifications of alluvial recordshave focused on sedimentary units rather than onthe surfaces that separate them (e.g., Miall 1991,1996), and the emphasis has largely been on finer-scale depositional units. To fill out the scale up tothird-order sequences, we have used the same rea-soning for channel-stacking and alluvial architec-ture as others (e.g., Allen 1978; Bridge and Leeder1979; Shanley and McCabe 1991, 1994; Posamen-tier and Allen 1993; Wright and Marriott 1993; Gi-bling and Bird 1994; Miall 1996). A comparable butnot precisely equivalent ranking of discontinuitiesfor nonmarine foreland basins based on observa-tions in the Campanian of Montana (Rogers 1994,1998) would be (1) bedding plane and bed-set bound-aries; (2) erosional bases and internal scours of sin-gle or multistory channels and paleosols developedlocally within single overbank units; (3) fluvial ero-sion surfaces, bypass (transport) surfaces, and omis-sion surfaces (including paleosols) that are morelaterally extensive amalgamations within an allu-vial succession or that lie between tracts with ap-preciable facies offset; (4) regionally extensive dis-continuities or narrow stratigraphic intervals inwhich alluvial architecture (primarily channel typeand spacing) indicates an abrupt increase in accom-modation within the foreland basin, with a possibleconcurrent landward (toward uplifted source area)shift in facies, analogous to turnaround from low-stand to transgressive systems tracts in the marine

Table 1. Taphonomic Signatures of Discontinuity Surfaces in the Two Medicine and Judith River Formations

Stratigraphic discontinuities Vertebrate Invertebrate Plant

Basal and internal scours of “upland”fluvial channels (n = 114;erosional; claystone,and caliche intraclasts)

Concentrations of skeletal debris generallyrare, ∼2.5% of channels surveyedpreserve lag concentrations of skeletaldebris, mixed faunal composition,variable preservational quality,mostly dense compact elements, somepolished bone pebbles

Unio present but not common,preservational quality variesfrom shell hash to intactdisarticulated and articulatedvalves, rotated out of lifeposition

Rare carbonaceous woodfragments and phyto-debris, rare silicifiedwood

Basal and internal scours of tidallyinfluenced fluvial channels(n = 47; erosional; claystone,and ironstone intraclasts)

Concentrations of skeletal debris relativelycommon, ∼9% of channels surveyedpreserve lag concentrations of skeletaldebris, mixed faunal composition,variable preservational quality, mostlydense compact elements, abundantpolished bone pebbles

Unio relatively common,abundant shell debris,disarticulated and articulatedvalves, typically rotatedout of life position

Abundant carbonaceouswood fragments andphytodebris, raresilicified wood

Marine flooding surfaces (D1, D2, D3, D4):Shoreface on paralic (n = 19;

erosional; claystone intraclasts)Very rare bone pebbles, fragmented and

abraded, poor preservational quality3–35-cm-thick shell beds,

oyster-dominated (Gyrostrea)with rare Corbicula andGoniobasis, pinch and swelllaterally, widely scatteredoyster valves, rare Uniofragments, leached debris,material varies fromintact valves to fragments,disarticulated, oxidized

Coaly stringers, coalyrip-ups

Shoreface on shoreface (n = 11;erosional; burrow rip-ups, reworkedconcretions, chert pebbles,ironstone intraclasts)

Widely scattered bones and teeth inter-spersed with localized concentrations ofshark teeth, fish bones, marine reptilebones, rare bones and teeth of terrestrialtaxa (mammals, dinosaurs), high degreeof fragmentation and abrasion, rounding

Rare fragmentary shell debris,white chalky preservation,partial and complete valves,rare gastropod steinkerns

None observed

Table 1. (Continued)

Stratigraphic discontinuities Vertebrate Invertebrate Plant

Offshore on paralic (n =12; no obviouserosion)

None observed Rare shell beds, up to 20 cmthick, dense packing,predominantly oysters(Gyrostrea) with rareCorbicula and Goniobasis,variable preservationalquality, intact andfragmentary valves

Coaly rip-ups

Offshore on shoreface (n = 4; noobvious erosion; ironstone andclaystone pebbles)

None observed Rare aragonitic shellfragments

None observed

Offshore on offshore (n = 2; noobvious erosion)

None observed None observed None observed

75.4-Ma transgressive surface (SB2; noobvious erosion):

Proximal None observed None observed None observedDistal None observed None observed None observed

80-Ma sequence boundary (SB1; erosional,abundant claystone and siltstoneintraclasts):

Proximal Rare bone fragments, isolated elements Rare gastropod steinkerns Carbonaceous plant debris,logs

Distal No data No data Carbonaceous plant debris,logs

Note. Observations are based on the analysis of 83 measured sections (Rogers 1995, 1998). Surfaces are ranked by inferred duration of hiatus. The n valuesreported in the stub column reflect the number of observations for a given type of discontinuity or facies juxtaposition (in the case of flooding surfaces).


realm (but not necessarily timed with such achange); (5) regionally extensive discontinuities ornarrow stratigraphic intervals in which alluvial ar-chitecture indicates a shift to decreasing accom-modation, analogous to turnaround from trans-gressive to highstand systems tract in the marinerealm; and (6) regionally extensive erosion surfaces,either planed or incised, analogous to third-ordersequence boundaries.

Several different types of stratigraphic disconti-nuities in the Two Medicine–Judith River recordare described in table 1, with emphasis on faciesassociations and taphonomic characteristics. Inthis report, a vertebrate skeletal concentration(bone bed or vertebrate microsite) is defined as adeposit preserving at least 5% bones and/or teethper unit volume (sensu Behrensmeyer 1999), withat least two or more individuals represented.

Minor Discontinuities Associated with Fluvial Chan-nels. Channel sandstones within the Two Medi-cine and Judith River formations can be character-ized as being either an “upland” type with purelyfluvial characteristics or a “tidally influenced” type(table 1). Tidal indicators include inclined hetero-lithic stratification, carbon drapes on fore- and toe-sets (in some instances distinctly paired), and thepresence of brackish and marine taxa (e.g., Tere-dolites, marine reptiles, sharks). Channel sand-stones are interstratified with a variety of flood-plain facies, including crevasse-splay deposits,diversely fossiliferous floodbasin pond/lake depos-its (dissociated plant, mollusk, fish, aquatic ver-tebrate, and lesser quantities of terrestrial dinosaurand mammal material), coastal swamp deposits(lignite and coal beds), and sporadic bentonite bedsand ironstone beds (see Rogers 1995 for more de-tailed facies descriptions). These floodplain faciesshow abundant evidence of pedogenesis.

Fluvial channels in both the Two Medicine andJudith River formations contain a multitude of mi-nor scours and nondepositional bedding-plane-scalehiatal surfaces, in addition to the more significanterosional surfaces that define channel bases and in-dividual stories (fig. 2). Skeletal debris is found onall of these types of surfaces throughout the TwoMedicine–Judith River study interval but is morecommon in tidally influenced channels. Using ourdatabase of 83 measured sections, we found that,of the 114 “upland” channels surveyed (most ofthese were in the Two Medicine Formation), only∼2.5% yield appreciable concentrations of verte-brate skeletal elements. Tidally influenced fluvialdeposits (most of these are in the Judith River For-mation) are considerably more productive, with

∼9% of the 47 channels surveyed yielding concen-trated lags of vertebrate debris (table 1).

Lags of vertebrate debris from both “upland” and“tidal” channel types tend to be taxonomically di-verse assemblages of elements, such as teeth,which are thoroughly dissociated and physico-chemically resistant, e.g., dinosaur, crocodile,champsosaur, mammal teeth; turtle shells; croco-dile scutes; gar scales; and assorted small densebones and bone fragments. The degree of fragmen-tation, abrasion, and weathering can vary widely.For example, fragile vertebrae with intact lateralprocesses are occasionally found in associationwith thoroughly rounded and highly polished bonepebbles. Carbonaceous plant debris and silicifiedwood, unionid and pisidiid bivalves, small high-spired gastropods, and rounded claystone and sid-erite pebbles are commonly found in associationwith fluvial concentrations of vertebrate debris (fig.2).

Shoreface Ravinements and Marine Flooding Sur-faces. Four discontinuities, referred to here as D1,D2, D3, and D4, are surfaces of conspicuous deep-ening that cut through paralic and shallow marinefacies of the Judith River Formation and mark thebases of small-scale (20–40 m thick), fourth-ordertransgressive-regressive cycles along the distal edgeof the nonmarine clastic wedge (figs. 1, 3). Each ofthese discontinuities is a composite surface com-prising both a fourth-order sequence boundary andits transgressive surface, which merges landwardwith a concave erosional ravinement surface. Thisravinement marks the maximum incursion of themarine shoreface in that cycle. None of the fourth-order cycles can be distinguished among the paralicstrata landward of the points where the ravinementdetaches from the sequence boundary sensu stricto.Coals are lenticular and unreliable as a means ofcycle definition and tracking in this particular rec-ord. The surfaces we define as D1–D4 thus divergefrom sequence boundaries in their proximal seg-ments and are locally diachronous but, throughouttheir extent, record omission and some amount ofshoreface erosion associated with marine floodingand thus have similar origins. We studied outcropexposures of skeletal material across these surfacesand found that biogenic debris is present in lowquantities in virtually all outcrops but that well-developed vertebrate lags are present only alongsegments that cut across older shoreface deposits.

The D1 discontinuity, which coincides over partof its extent with the 75.4-Ma discontinuity, is eas-ily identified in well logs but is poorly exposedalong the Missouri River. In the few localitieswhere we could examine it, the surface juxtaposes


Figure 2. Outcrop photographs of fossiliferous fluvial channel deposits. A, Lag of Unio shell debris and bone drapingbasal disconformity of a tidally influenced channel (hammer is 30 cm long). B, Internal scour surface mantled by anintraclast lag of claystone and ironstone pebbles and disseminated vertebrate elements. C, View of tidally influencedchannel deposit characterized by well-developed inclined heterolithic stratification and several meters of erosionalrelief. Channels in the study interval that show evidence of tidal influence are more likely to preserve lags of vertebrateskeletal material than are their “upland” counterparts. This presumably reflects the fossiliferous nature of surroundingsource facies in coastal settings, which are reworked during downcutting or lateral migration of channels to producelag concentrations. Arrow points to basal lag concentration illustrated in D. D, Basal lag includes disarticulated Uniovalves interspersed with carbonaceous debris and compact vertebrate elements (teeth, scales, phalanges, vertebrae),some of which exhibit polish.

shoreface strata on lignite, coal, or, more rarely,carbonaceous claystone. Dispersed coaly rip-upsmantle the surface along with widely scatteredshells and shell fragments (predominantly oystervalves). Based on admittedly scant exposures, thereis no appreciable vertebrate material associatedwith this surface (table 1).

The well-exposed D2 and D3 discontinuity sur-faces have been traced for several tens of squarekilometers in eastern portions of the Judith RiverFormation type area (figs. 3, 4). Along the seawardsegment of the D2 surface, offshore clay shales ofthe Bearpaw Formation rest on paralic beds of lig-nite or coal. With this facies association, the surface

is relatively unfossiliferous, although scattered coalrip-ups and very rare isolated shell fragments arepresent (table 1). Along western (more onshore) seg-ments of the D2 surface, where sandy shorefacestrata rest on paralic facies, bioclastic debris is dis-persed but more frequently encountered. Bioclastsinclude highly abraded and weathered bone pebbles(presumably dinosaur); isolated, weathered, andtypically fragmentary valves of the freshwater clamUnio; and small to large rip-ups of lignite. Thinstringers of chalky shell hash crop out locally a fewcentimeters above the D2 surface.

The D3 flooding surface is also overlain by in-creasingly shallower water facies to the west. The


Figure 3. Dip-parallel cross section in eastern portion of Judith River Formation type area showing the four dis-continuity surfaces (D1, D2, D3, D4) that bound three fourth-order transgressive-regressive cycles. Each discontinuityis a composite surface comprising both a fourth-order sequence boundary and its transgressive surface, which mergeslandward with a concave, erosional ravinement surface. Biogenic debris is present in low quantities in virtually alloutcrops of these surfaces, but well-developed vertebrate lags are present only along segments of the D3 surface thatcut across older shoreface deposits (note considerable erosional relief where the shoreface is juxtaposed). The D3discontinuity is in general extremely fossiliferous with regard to vertebrate skeletal hardparts; fossil debris is strewnas a thin pavement where the D3 surface is planar, and rich concentrations occur within local erosional scours.

D3 surface is extremely fossiliferous and is richestwhere shoreface strata cut across older shorefacedeposits (figs. 3, 4). In these segments, fossil debrisis strewn as a thin pavement where the D3 surfaceis planar; however, rich concentrations occurwithin small erosional scours (0.5–2 m; fig. 4).These lags are 1–5-cm thick on average, and insome localities, two discrete lag stringers spaced10–15-cm apart are present. Wet-sieving of matrixfrom a single scour can yield thousands of sharkteeth and fish elements, along with marine reptiles,rare terrestrial vertebrates, siderite and chert peb-bles, and reworked cemented burrow-fills (primar-ily simple tubes; see Case 1978 for a description ofselachian fossils from the D3 surface in the vicinityof Suction Creek).

The skeletal composition of D3 lags is over-whelmingly dominated by vertebrates: molluscanshell debris is minor, and there are no recordedplant occurrences. Vertebrate debris exhibits vari-able states of preservation, including rare elementsin pristine condition, but is dominated by taxo-nomically unidentifiable bone pebbles. Most ele-ments exhibit some degree of surficial degradation,with rounding and abrasion most common. The

composition of bioclasts varies from onshore to off-shore, probably due to the variable bioclast contentof underlying facies that were reworked during ero-sive transgression of the shoreface (cf. Fischer 1961;Swift 1968; Liu and Gastaldo 1992). In the vicinityof sections 4–6 in figure 3, vertebrate debris over-whelmingly dominates bioclasts, both in lags andin isolated occurrences; concentrations of this typecan also be traced for ∼17 km along depositionalstrike (N-S) from the Missouri River outcropsshown in figure 4. Approximately 5 km to the east,in the vicinity of section 8, the D3 lag contains farless vertebrate debris but does contain a diversearray of cemented burrow-fills (shallow-marineThallasinoides and Rosselia, among others) alongwith rare small nautiloids.

The widely exposed D4 surface is a transgressiveoverstep that can be traced throughout most of theJudith River Formation type area (Rogers 1995).This surface is extremely fossiliferous, as recog-nized since Stanton and Hatcher’s (1905) originalreport on the Judith River Formation. It is char-acterized by extensive oyster-dominated shell bedsrather than by vertebrate-dominated lags (fig. 4).Across most of its outcrop belt, the D4 surface is

Figure 4. A, View of D2 and D3 discontinuity surfaces along Woodhawk Creek drainage, on the right bank of theMissouri River, in the Judith River Formation type area. Shoreface strata (marked by trees) are sharply juxtaposedover coastal plain strata. The D3 discontinuity is embedded within shoreface strata, and the D3 arrow points to theapproximate position in figure 4B–4D. Based on regional thickness data, the position of D1 would be approximatelyat the base of the exposures. B, Local scour marking the D3 erosional discontinuity. This erosional surface has beencut between two large ironstone concretions that were already indurated on the Campanian sea floor when trans-gression ensued. The base of the scour preserves a concentration of shark teeth, marine reptile and fish bones (thelarge bone immediately adjacent to the D3 arrow is a plesiosaur limb element), rare shell debris, and cemented burrowcasts. C, Close-up of the D3 discontinuity. In the absence of skeletal debris, this sand-on-sand surface would bevirtually impossible to identify. The presence of a skeletal lag makes it possible to track this surface over an area ofapproximately 100 km2 in the Judith River Formation type area. D, Representative sample of vertebrate debris screenwashed from the D3 locality figured in figure 4B. Wet-sieving of matrix from a single scour can yield thousands ofshark teeth and fish elements, along with marine reptiles, rare terrestrial vertebrates, siderite and chert pebbles, andreworked cemented burrow fills (primarily simple tubes). E, Outcrop view of the D4 surface. At this particular locality,a ∼20-cm-thick oyster-dominated shell bed is sharply juxtaposed on an underlying bed of lignite. F, View of a tidallyreworked oyster coquina exposed along the rim of Knox Coulee in the western portion of the Judith River Formationtype area. Comparable beds, which accumulated in tidal channels and reach 3 m thick, presumably supplied oystershell debris to the D4 surface through the process of shoreface erosion.


Figure 5. SB2 discontinuity. A, Arrow points to base of abrupt facies juxtaposition in Two Medicine Formationseparating underlying fluvial facies from overlying lacustrine carbonate facies. B, Arrow indicates position in JudithRiver Formation of abrupt shift from a sandstone-dominated alluvial record to a mudstone-dominated alluvial record.No biogenic debris of any sort has been found in direct association with this discontinuity, although the relativeabundance of vertebrate skeletal debris (dinosaur nest sites, bone beds, vertebrate microsites) above and below thissurface does change dramatically in both the Two Medicine Formation and the Judith River Formation.

embedded between lignite or low-grade coal andseveral overlying meters of massive to swaley-bed-ded sandstone, commonly containing the shallow-marine trace fossil Ophiomorpha. This sandstonepasses rapidly upward into dark shales of the Bear-paw Formation. Pavements or densely packed beds(up to 50 cm thick) of disarticulated and fragmentaloyster shells (Gyrostrea subtrigonalis [Evans andShumard]; genus according to Malchus 1990) occurat the lignite-sandstone contact in most outcrops.Subordinate taxa at the D4 surface include simi-larly euryhaline corbulid and anomiid bivalves andthe gastropod Goniobasis. The oyster-dominatedshell bedforms a tabular but not perfectly contin-uous body that can be traced eastward from theerosional limit of the stratigraphic interval for morethan 40 km. On the east (downdip, distal edge), itthins to a single-valve pavement, which disappearsentirely at a point roughly coincident with thepinchout of paralic facies in the underlying JudithRiver Formation.

Locally, thin pavements of disarticulated oysters(and more rarely Anomia) occur within the under-lying lignite. Such pavements may have been thesource of some shell debris concentrated in the D4lag, but the lignites are clearly very difficult toerode. The shells within the lignites are probablyallochthonous material transported into the coastalswamps during storms rather than populations thatgrew in situ. Much of the oyster material in the D4lag is probably also allochthonous and, in this in-

stance, was exhumed and redistributed seawardthrough the process of shoreface erosion from au-tochthonous oyster banks in tidal environmentswhose sedimentary record has been otherwise de-stroyed by transgressive ravinement. Such a com-plex of small in situ oyster banks and tidally re-worked oyster coquinas are preserved in closeassociation with the D4 flooding surface along thewestern rim of Knox Coulee, in the western portionof the Judith River Formation type area (fig. 4).

Nonerosional SB2 Discontinuity (Nonmarine Exten-sion of 75.4-Ma Third-Order Transgressive Surface).The SB2 discontinuity marks an abrupt change infacies and channel-stacking patterns in both uplandand coastal plain portions of the Two Medi-cine–Judith River clastic wedge and is interpretedto correlate with the turnaround from regressionto transgression in the marine basin (fig. 1; Rogers1994, 1998). In the Two Medicine Formation typearea, the SB2 surface juxtaposes anomalous lacus-trine carbonate facies on fluvial channel and flood-plain facies (fig. 5) and also coincides with an in-crease in volcanic detritus and the appearance ofextrabasinal metamorphic pebbles. In the more dis-tal Judith River Formation, the SB2 surface marksan abrupt shift from sand- to mud-dominated al-luvial facies that can be traced in outcrop and ingeophysical logs throughout north-central Mon-tana and southern Alberta (fig. 5). The disconti-nuity itself shows no physical evidence of erosion,despite its striking appearance when viewed at a


distance. The nature of the facies change indicatesan increase in accommodation, coinciding with theonset of Bearpaw transgression (T9) in the marinebasin to the east. Thus, we refer to the SB2 dis-continuity as the nonmarine (upstream) expressionof that 75.4-Ma transgressive surface. The only doc-umented erosion along this transgressive surface isin the easternmost part of the study area, whereSB2 locally merges with the fully marine D1 ra-vinement. It may well be that there is virtually nohiatus of any type—erosional or omissional—associated with the SB2 discontinuity along itsnonmarine extent. The discontinuity lies in an ap-propriate position to be a sequence boundary, butapplying the operational definitions of Van Wag-oner et al. (1990), Christie-Blick and Driscoll (1995),and Van Wagoner (1995), among others, we avoidthat terminology in the absence of clear evidencefor a fall in baselevel (erosional downcutting or bev-eling). Instead, SB2 could be to referred to as a type2 unconformity (with minor subaerial exposure) orthe updip correlative conformity of a sequenceboundary (cf. Rogers 1998).

Facies on both sides of the SB2 discontinuity inboth the Two Medicine and Judith River formationscontain moderately to extremely abundant verte-brate and invertebrate fossils (Rogers 1995). How-ever, we have nowhere found a direct associationof skeletal material, of any quantity or quality, withthe discontinuity (table 1). The only taphonomi-cally notable aspect of the SB2 discontinuity is thedramatic increase in the abundance of bone con-centrations in facies above the contact (see below).These concentrations include dinosaur nest sites,bone beds, and vertebrate microfossil localities, aswell as scour-associated fluvial concentrations asdescribed above (Rogers 1995, 1998; Rogers andEberth 1996).

Erosional SB1 Discontinuity (Nonmarine Extension of80-Ma Third-Order Sequence Boundary). The SB1discontinuity is a laterally extensive set of ero-sional surfaces embedded within multistory fluvialsandstone sheets low within the Two Medicine For-mation (figs. 1, 6). This major through-going dis-continuity is interpreted as the nonmarine exten-sion of the 80-Ma third-order sequence boundary(Rogers 1994, 1998). It can be traced in nearly con-tinuous outcrop for tens of kilometers along theTwo Medicine River and Cut Bank Creek and istentatively correlated with a major erosional sur-face embedded within Hanson and Little’s (1989)“sequence 4” of the marine Eagle Formation. In theTwo Medicine type area, the discontinuity is con-tained within a 10–15-m-thick interval of inter-cutting fluvial scours and channels and is charac-

terized by a thick (up to 1.3 m) and laterallypersistent intraclast lag facies, pervasive and unu-sually intense oxidation, and a temporary shiftfrom fine- to medium/coarse-grained sandstone(fig. 6). The lag consists of rounded to angular peb-bles and blocks of intraformational claystone andsiltstone in a matrix of coarse-grained oxidizedsandstone. Vertebrate and invertebrate skeletal de-bris is exceptionally rare (table 1) and is in poorcondition (e.g., bone pebbles, gastropod steinkerns).In contrast, plant material is both abundant andwell preserved. Large carbonized logs and log im-pressions are common (fig. 6), as is finer phyto-debris. This is similar to the more distal Eagle For-mation, where large oriented logs and angularsandstone blocks are reported to rest on a regionallytraceable marine disconformity (Hanson and Little1989).

Facies surrounding the SB1 discontinuity in theTwo Medicine Formation are not particularly fos-siliferous. Underlying beds preserve scattered car-bonaceous material and root traces, and silicifiedlogs crop out locally. The first appreciable concen-trations of bones and shells crop out above the dis-continuity in beds that accumulated in a lowercoastal plain setting contemporaneous with theClaggett transgression (Rogers 1995, 1998).

Additional Taphonomic Observations. The ta-phonomic condition of vertebrate skeletal materialassociated with discontinuity surfaces varies rela-tively little across the study interval. Disarticu-lated and dissociated vertebrate elements consis-tently exhibit evidence for physical abrasion andprogressive faceting, regardless of whether they arein a marine or terrestrial setting. Moreover, thereis an apparent size sorting of skeletal debris in bothsettings, with long axes of vertebrate elements de-rived from a wide array of taxa, ranging from 1 to20 mm on average, with relatively few outliers (seefig. 4). Important taphonomic distinctions exist aswell. Polish is occasionally developed on facetedskeletal debris recovered from fluvial channels buthas not been found in the marine record. This ta-phonomic attribute may have important implica-tions for the origins of skeletal lags. In addition,there is a tendency for vertebrate material in fluvialchannels of the Two Medicine Formation to bemore indurated (less friable) than counterparts intidally influenced channels of the Judith River For-mation. This variation in induration may relate tothe differing geochemical environments of thechannel facies themselves (e.g., there is more or-ganic debris in the tidal beds, and presumably therewas saltwater influence as well). Alternatively, itmay reflect a different taphonomic history for the


Figure 6. Outcrop views of SB1 discontinuity (80-Ma sequence boundary) in the Two Medicine Formation. A, Viewof the unconformity along the Two Medicine River showing the pervasive oxidation that characterizes facies im-mediately surrounding the discontinuity. B, Close-up view of oxidized lag that marks the SB1 discontinuity. Arrowpoints to rounded bank-collapse block. The intraclast lag marking the discontinuity locally exceeds 1 m thick. C,Arrows point to downcutting erosional surface in fluvial sandstone sheet along the Two Medicine River. Here, thesurface exhibits up to 5 m of erosional relief. D, The only biogenic material commonly associated with the SB1discontinuity consists of plant debris, most notably large carbonized trees.

bioclasts altogether, with channels in one locale orthe other perhaps downcutting and/or eroding lat-erally and incorporating skeletal material from sur-rounding floodplain beds.

Origins of Vertebrate Skeletal Concentrations:Role of Preexisting Sources

Vertebrate skeletal material does occur in associ-ation with discontinuity surfaces within the TwoMedicine–Judith River record, but there is no clearcorrelation between the abundance of material andthe inferred duration of the actual hiatus, nor is thetaxonomic diversity or taphonomic condition ofbioclasts correlated with the type of discontinuity.The most significant discontinuities—the clearlyerosional SB1 set of surfaces and the less clearly

hiatal but widespread SB2 discontinuity—are vir-tually barren of skeletal debris. In contrast, manyhiatal surfaces of lesser extent and duration, rang-ing from marine flooding surfaces down to scourswithin individual fluvial channels, are mantled byrich skeletal concentrations. This indicates thatthese vertebrate skeletal concentrations are notsimply accumulations of elements generated bydeath during the hiatus—that is, they were not pro-duced solely by attritional input from communitiesthat lived contemporaneously with the formationof the discontinuity surface (Behrensmeyer andChapman 1993). Instead, all vertebrate concentra-tions in the Two Medicine–Judith River record arepatchy in their distribution, and even dispersedskeletal material varies in abundance along discon-tinuity surfaces. This variation closely tracks the


abundance of vertebrate material in underlying andlaterally disposed facies, providing evidence thatthis material was supplied primarily through ero-sional reworking. By this, we are not referring tothe continuous reworking of bedload that proceedsduring the life of a single channel but to the ex-humation of material from older strata in cut banksor during incision. Once remobilized, this exhumedmaterial could be rapidly reburied or, alternatively,admixed with newly produced hardparts, generat-ing time averaged and ecologically mixed fossilassemblages.

For example, the D3 flooding surface in the ma-rine portion of the Judith River record demonstra-bly truncates several meters of underlying shore-face strata along the discontinuities mostfossiliferous segment (figs. 3, 4), and the sharkteeth, fish and marine reptile bones, and scatteredchert pebbles that mantle the surface can also befound, albeit more widely dispersed, in underlyingbeds. Additional evidence for exhumation consistsof reworked steinkerns of burrows (induratedcasts), which are found in abundance on the D3surface and are of the shoreface ichnogenera com-monly found in underlying beds. Their concen-trated occurrence on the surface is strong evidenceof local erosional exhumation without significantlateral transport. Finally, the taphonomy of theskeletal debris concentrated on the D3 surface isconsistent with physical reworking, in that the ma-jority of the vertebrate elements exhibit evidenceof breakage and/or abrasion (e.g., more than 70%of fish vertebrae show evidence of abrasion androunding).

Evidence for the exhumation of skeletal debris isalso apparent in the terrestrial portion of the studyinterval, where the recurrent skeletal lags thatmantle fluvial erosion surfaces closely track theabundance of skeletal debris in surrounding flood-plain facies. Fossil preservation in the floodplainenvironment ranges from isolated skeletal ele-ments to concentrated dinosaur bone beds and ver-tebrate microsites. Sedimentologic and tapho-nomic data indicate that floodplain micrositesformed via the attritional mortality of predomi-nantly aquatic taxa in coastal plain ponds where,over time, physicochemically resistant elementssuch as teeth, scales, scutes, and small dense bonesaccumulated to fossiliferous levels. Floodplain mi-crosites are particularly abundant above the SB2discontinuity, with nine of the 13 localities thatcan be placed with confidence relative to SB2 in-tercalated in overlying strata. Channel-hosted mi-crosite assemblages are also most abundant above

SB2, with 11 of 13 localities situated above the dis-continuity (table 2).

The positive correlation in the stratigraphic dis-tribution of floodplain- and channel-hosted mi-crosite localities in the Judith River Formation isnot a simple coincidence, and we have found evi-dence that suggests that Judith River channel as-semblages include material reworked from preex-isting floodplain microsites. The intersection of afossiliferous channel deposit (site UC-8302; Rogers1995) with an underlying floodplain microsite (siteUC-8302A; Rogers 1995) in the Judith River For-mation type area provides a good example (fig. 7).The 45-cm-thick floodplain microsite (UC-8302A)is a massive clay-rich siltstone containing turtle,crocodile, fish, and dinosaur debris interspersedwith pisidiid bivalves, small thin-shelled gastro-pods, fragmentary Unio debris, and carbonaceousstringers. The overlying channel deposit is char-acterized by low-angle inclined sets of massive tofaintly cross-stratified fine-grained sandstone. Thincarbon and clay partings drape most set boundariesand some foresets; small- to medium-scale troughcross-bedding, ripple lamination, and rare climbingripples are present near the top of this tidally in-fluenced fluvial deposit. The skeletal lag in thebasal 20 cm of this unit (UC-8302) crops out im-mediately above UC-8302A and contains a com-parable assemblage of vertebrate elements in as-sociation with abundant fragmentary shells ofchannel sand-dwelling Unio bivalves and flood-plain-derived smaller invertebrates (mud-dwellingpisidiid bivalves and gastropods), ironstone andclaystone pebbles, and coal stringers. The UC-8302assemblage crops out immediately above an ero-sional scour developed on the underlying UC-8302A stratum (fig. 7a), and this association is in-terpreted to reflect the erosional exhumation andreworking of a preexisting source bed, with sub-sequent generation of a fossiliferous fluvial lagdeposit.

Locality UC-8349 (Rogers 1995) provides a sec-ond example of a fossiliferous channel deposit con-sistent with the reworking of a preexisting fossil-rich bed (fig. 7). This ancient channel depositconsists of 3.2 m of fine-grained sandstone withmedium- to large-scale trough cross-bedding andlocalized planar bedding. A lag of vertebrate skel-etal debris crops out along the basal contact, andseveral more lag stringers mantle internal scoursurfaces that mark set boundaries. The skeletal de-bris amassed on these surfaces consists of teeth (di-nosaur, champsosaur, crocodile, mammal), fishscales, crocodile and turtle scutes, and small densebones (vertebrae, phalanges). Quality varies from


Table 2. Stratigraphy of 55 Vertebrate Skeletal Concentrations in the Terrestrial Two Medicine–Judith River RecordRelative to the SB2 Discontinuity and Third-Order Transgressive-Regressive (T-R) Cycles of the Western InteriorSeawaya

Locality General site categories Microsite typeb T vs. R

Two Medicine Formation:TM-055 Hadrosaur-dominated bonebed ) RTM-067 Hadrosaur-dominated bonebed ) RTM-041 Hadrosaur-dominated bonebed ) TTM-019 Hadrosaur-dominated bonebed ) TTM-023 Ceratopsian-dominated bonebed ) TTM-046 Ceratopsian-dominated bonebed ) TGM-1c Ceratopsian-dominated bonebed ) TTM-068 Multitaxic dinosaur bonebed ) TTM-066 Dinosaur nest site ) TTM-018 Dinosaur nest site ) TTM-068A Dinosaur nest site ) TTM-097 Dinosaur nest site ) TTM-052 Dinosaur nest site ) TTM-038 Dinosaur nest site ) TTM-061 Dinosaur nest site ) TTM-088 Vertebrate microsite a TTM-141 Vertebrate microsite a TTM-020 Vertebrate microsite a TTM-053d Vertebrate microsite a T

Judith River Formation:UC-92J Hadrosaur-dominated bonebed ) TUC-939 Multitaxic dinosaur bonebed ) TJR-142 Dinosaur nest site ) TJR-122 Dinosaur nest site ) TUC-9312 Vertebrate microsite a RUC-932 Vertebrate microsite a RUC-934d Vertebrate microsite a RUC-935d Vertebrate microsite a RUC-929d Vertebrate microsite b RUC-8325d Vertebrate microsite a RUC-9314d Vertebrate microsite a TUC-9311 Vertebrate microsite a TUC-9313 Vertebrate microsite a TUC-8326d Vertebrate microsite a TUC-914d Vertebrate microsite a TUC-8302Ad Vertebrate microsite a TUC-8322Bd Vertebrate microsite a TUC-8303d Vertebrate microsite a TCBHd,e Vertebrate microsite a TUC-938 Vertebrate microsite b TUC-937d Vertebrate microsite b TUC-9210d Vertebrate microsite b TJR-142Ad Vertebrate microsite b TUC-9110d Vertebrate microsite b TUC-8322Ad Vertebrate microsite b TUC-941d Vertebrate microsite b TUC-915d Vertebrate microsite b TUC-917d Vertebrate microsite b TUC-919d Vertebrate microsite b TUC-8302d Vertebrate microsite b TUC-8315d Vertebrate microsite a ?UC-9310d Vertebrate microsite a ?UC-931d Vertebrate microsite a ?UC-936d Vertebrate microsite a ?UC-913d Vertebrate microsite b ?UC-8439d Vertebrate microsite b ?

Note. TM and JR indicate Museum of the Rockies (Bozeman, Mont.) localities. UC designations were used by Rogers (1995) todistinguish among sites in the field.a Rogers 1995, 1998.b a = Floodplain microsite, subaqueous (pond/lake) settings ; b = channel-hosted microsite.c Brachyceratops bonebed of Gilmore 1917.d Sample available at Macalester College (St. Paul, Minn.).e Clambank Hollow site of Sahni 1972.


Figure 7. A, Channel-lag locality UC-8302. The skeletal lag in the basal 20 cm of this tidally influenced sandstonebody contains vertebrate debris in association with abundant fragmentary shells of channel sand-dwelling Uniobivalves and floodplain-derived smaller invertebrates (mud-dwelling pisidiid bivalves and gastropods), ironstone andclaystone pebbles, and coaly stringers. Arrows indicate position of lag deposit and demonstrate the downcutting ofthe channel to the right. The underlying 45-cm-thick source bed (UC-8302A) is a tan-to-brown massive clay-richsiltstone containing turtle, crocodile, fish, and dinosaur debris interspersed with pisidiid bivalves, small thin-shelledgastropods, fragmentary Unio debris, and carbonaceous stringers. B, View of locality UC-8349, which preserves severallag stringers of vertebrate skeletal debris on internal scour surfaces that mark set boundaries (arrows). The recurrenceof skeletal lags in this deposit suggests that a preexisting source of skeletal debris supplied the channel as bedformsmigrated and the channel filled. C, Close-up of typical floodplain microsite “source facies” showing scattered shellfragments (white) and a small dinosaur centrum (top center) in a matrix of massive, silty claystone. The sedimentologyof these tabular beds indicates that deposition occurred in shallow floodbasin ponds or lakes. Skeletal debris presum-ably accumulated via attritional processes. Bioclasts typically include fragmentary and intact pisidiid bivalves, smallhigh-spired gastropods, and disseminated vertebrate elements (teeth, scales, scutes, phalanges, vertebrae, etc.). D, Thethree polished pebbles of trabecular bone above the scale bar (=1 cm) were collected from the basal lag of a fluvialchannel deposit in the Judith River Formation (UC-9301), and they presumably reflect the physical abrasion ofpermineralized material. The bone pebbles below the scale bar are typical of low-energy floodplain sites (see fig. 7C)and show no evidence of polish.

pristine elements to rounded and unidentifiablebone pebbles. Claystone pebbles and carbonizedwood fragments, constituting rip-ups from flood-plain facies, are also concentrated on basal and in-

ternal scours. The recurrent nature of skeletal lagsat this particular site is interpreted to reflect theerosion of a nearby source bed that repeatedly sup-plied the channel with skeletal debris. Bones and


teeth were apparently delivered to the active chan-nel from an eroding bank somewhere upstreamfrom the present site, and skeletal material wasconcentrated in the scour pits of three-dimensionalbedforms as they migrated downstream.

Finally, features of the lag concentrations them-selves lend credence to the reworking hypothesis.First, the taxonomic makeup of channel-hosted as-semblages shows a general correspondence to thatof floodplain localities (Rogers et al. 1995; Carranoet al. 1995, 1997; Blob et al. 1997), with the rareexception of the marine taxa (plesiosaurs, sharks)that may be found in tidally influenced channelfacies. Moreover, channel-lag assemblages containthe same basic assortment of elements as the flood-plain microsites—teeth, vertebrae, phalanges,scutes, and scales—but some elements are polished(fig. 7). Polish is an enigmatic bone modificationfeature (Morlan 1984; Behrensmeyer et al. 1989),and its origin (or origins) is less than clear. It canpresumably be imparted to permineralized (“pre-fossilized”) material on exhumation and physicalabrasion (A. K. Behrensmeyer, pers. comm. 1994);in fact, an initial period of burial and diagenesiswould seem to be prerequisite for abrasion to resultin polishing rather than rounding alone. The pres-ence of polished bones in the concentrations thatdrape fluvial scours suggests that at least a portionof the skeletal fraction was reworked in a prefos-silized condition. Interestingly, polish is not evi-dent on reworked skeletal material recovered frommarine flooding surfaces in the study interval.Whether this indicates a more limited depth ofstratigraphic incision or different fossilization dy-namics in the two depositional settings, is unclear.

Comparison with Marine InvertebrateSkeletal Concentrations

The strong dependence of vertebrate lag develop-ment on preexisting local sources of skeletal ma-terial, rather than on the magnitude of the hiatus,contrasts with skeletal concentrations of inverte-brates in marine successions. The marine inver-tebrate record includes many kinds of shell beds(relative concentrations of skeletal elements ≥2mm), which are distinguished by variations inphysical scale, composition, preservational state,and inferred origins (Kidwell 1991; Brett and Baird1993; Doyle and MacDonald 1993; Fursich andOschmann 1993; Meldahl 1993). The majority ofshell beds are (1) relatively simple event concen-trations created by short, discrete, and localizedevents of skeletal input or reworking (e.g., stormbeds, turbidites, shell-filled feeding pits, clusters of

gregarious benthos) and (2) generally larger-scale,amalgamated or accretionary composite concentra-tions that reflect more complex multiple-event his-tories of local colonization, physical and biogenicwinnowing, and/or input from exotic sources (e.g.,facies-scale bioclastic shoals, subtidal channel fills,washover fans; classification scheme of Kidwell1991). Basin-scale studies have shown that thesetypes of concentrations constitute the backgroundpattern of skeletal accumulation in aggradationalshallow marine environments and are strongly fa-cies controlled. In terms of stratigraphic disconti-nuities, they are most commonly associated withbedding planes and bed-set boundaries but alsomantle some parasequence-bounding flooding sur-faces (Kidwell 1993b).

Marine records are also characterized by tapho-nomically more complex and “time rich” (3) hiatalconcentrations, dominated by skeletal hardpartsgenerated during stratigraphically significant hia-tuses in sedimentation (e.g., sediment starvation ofsubmarine paleohighs, distal portions of basins, andtransgressive shelves; sediment bypassing of shal-low-water environments at grade), and (4) lag con-centrations, dominated by hardparts exhumedduring stratigraphically significant erosion or cor-rosion of older deposits (e.g., erosional beveling dur-ing transgressive ravinement, planing, or incisionduring forced regression of shoreline; see classifi-cation scheme of Kidwell 1991; and see Brett andBaird 1993; Fursich and Oschmann 1993; Kidwell1993b). Hiatal concentrations, like composite con-centrations, are products of multiple events of col-onization and winnowing but are stratigraphicallycondensed (demonstrably thinner than coeval, lessfossiliferous strata) and typically bear additional ev-idence of prolonged low net sedimentation, such asecological condensation and admixtures of hard-parts with diverse taphonomic or diagenetic his-tories. Lags, in contrast, are typically much thinner(limited to a few centimeters or decimeters ratherthan ranging up to several meters in thickness),patchy (rather than laterally extensive, tabular bod-ies) and composed of true hydraulic and diageneticresidua and have compositions that are ecologi-cally, biostratigraphically, and/or taphonomicallydiscordant with the sedimentary facies that hoststhem. Many studies demonstrate that, although notall marine discontinuity surfaces are mantled byskeletal concentrations, taphonomically complexhiatal and lag concentrations are consistently as-sociated with significant discontinuities, such asparasequence-bounding flooding surfaces, trans-gressive surfaces, midcycle surfaces of maximumtransgression, and third- or fourth-order sequence


boundaries (references above and Brett 1995; Ab-bott 1998; Gillespie et al. 1998; Kondo et al. 1998).

Some marine skeletal deposits are compounds ofhiatal and lag concentrations. The 1–10-m-thickshelly sands exposed on many modern continentalshelves provide excellent modern analogs of the hy-brid concentrations that mantle transgressive sur-faces. These skeletal sands contain either a mixtureof lag material exhumed during shoreface ravine-ment of estuarine facies (e.g., black-stained oysters)or older bedrock (e.g., shells with unusual sedi-mentary infills, or styles of shell replacement) andhiatal shells newly produced under open marineconditions (environmentally condensed macro-benthic and marine vertebrate debris derived fromthe bathymetric array of communities that mi-grated across the shelf during the past 20,000 yr ofinundation and effective sediment starvation, ver-ified through radiocarbon dating; Fischer 1961;Swift 1968; Swift et al. 1991; Anderson and Mc-Bride 1996; Stanley and Bernasconi 1998). Volu-metrically, hiatal material exceeds lag material insuch compound accumulations, both here and inthe fossil record, and numerically, hiatal concen-trations appear to be far more common than lags(Kidwell 1991).

Thus, in the genesis of skeletal concentrationsassociated with significant marine discontinuitysurfaces, at least those dominated by mollusks, theproduction of new elements during the hiatus is ofequal or greater importance than the gains fromerosional exhumation, and the negative aspects ofhiatuses (e.g., retarded permanent burial, repeatedsmall-scale burial-exhumation cycles, possible el-evated attack from other taphonomic agents) do notoutweigh the positive aspects of low sedimentarydilution. Apparently, even in sites of sediment star-vation, episodes of burial and/or thin veneers ofmobile sediment can protect accumulating shellmaterial. Indeed, the progressive concentration ofskeletal material during a hiatal interval may aidskeletal preservation by buffering pore waters, ar-moring against erosion, and stimulating further col-onization by shell-producing benthos (Davies et al.1989; Kidwell 1989).

Hiatal skeletal concentrations in the marinetaphonomic sense are extremely rare in the TwoMedicine–Judith River record. Paleosol-hosted as-semblages would constitute hiatal concentrationsof terrestrial taxa, but none of these were encoun-tered in the study area. This is not entirely unex-pected, given the harsh conditions that can accom-pany pedogenesis. Given the type and rate ofvertebrate material supplied by local mortality inthis Cretaceous example, the destructive effects of

nondepositional hiatuses on the land surface ap-parently outweighed the positive effects of low sed-imentary dilution. Vertebrate concentrations as-sociated with Two Medicine–Judith River overbankdeposits instead consist of time averaged floodplainpond assemblages and snapshot-type assemblagesgenerated by rapid burial of vertebrates on land sur-faces (mono- or paucispecific bone beds and nestsites).

Transgressive-Regressive Differencesin Taphonomy

The stratigraphy of 55 vertebrate concentrations inthe Two Medicine–Judith River record was ascer-tained relative to the SB2 discontinuity (table 2),and despite being unfossiliferous itself, the discon-tinuity does mark a regional change in the richnessof the vertebrate fossil record. Overlying beds,which are interpreted to have accumulated duringthe Bearpaw transgression (T9), are characterizedby a much greater abundance of vertebrate skeletalmaterial. In the Two Medicine Formation, 17 of 19sites surveyed definitively occur in the transgres-sive record above SB2. In the Judith River Forma-tion, 24 of the 30 sites that can be placed relativeto SB2 with certainty crop out in transgressive-phase deposits (table 2).

Some potential controls on this regional taphon-omic pattern can be ruled out. In particular, prox-imal versus distal location in the foreland basin isnot a factor—upland (Two Medicine) and lowland(Judith River) facies both exhibit the regressive-transgressive contrast in preservation. Also, overalloutcrop area can be ruled out as a factor becauseregressive-phase deposits are at least as well andperhaps better exposed than transgressive-phase de-posits. However, several other explanations for thislarge-scale taphonomic pattern are possible, fromboth the supply side (rate and quality of bone input)and the removal side of the taphonomic equation.

Accessibility of Productive Facies. One possiblecontrol on the large-scale distribution of fossils isthe relative richness of facies available for rework-ing. In the Two Medicine and Judith River forma-tions, tidally influenced channels and floodbasinponds are the most productive terrestrial facies,whereas shoreface deposits are the richest of theshallow marine facies. In “upland” exposures of theTwo Medicine Formation, the contrast in fossilabundance across the SB2 discontinuity does notcoincide with any significant change in alluvial fa-cies, only in the channel/floodplain ratio (Rogers1998). Fluvial and floodplain deposits of the high-stand depositional system (R8) are broadly similar


to those of the transgressive depositional system(T9). However, in the lowland coastal plain recordof the Judith River Formation type area, the shiftto a more fossiliferous transgressive record coin-cides with an increase in the abundance of tidallyinfluenced channel and floodbasin pond deposits.The increased abundance of these two particularfacies in the transgressive portion of the JudithRiver record clearly influences the large-scale strat-igraphic pattern of fossil occurrence in the typearea. Outcrops are not available to test whether theregressive record also increases in fossil content to-ward the paleocoastline.

Burial Rate and Likelihood of Reworking. Higheraccommodation of deposition during the risingbaselevel conditions of transgression should alsoresult in higher net rates of burial of any given dep-ositional surface—i.e., fewer burial-exhumation cy-cles per stratigraphic increment. Although this en-hanced burial potential is probabilistic and thuswould not affect every carcass or bone on an an-cient landscape, it should reduce destructive cyclesof exhumation and reworking within facies, as wellas increase the preservation potential of faciesthemselves. This transgressive versus regressivereasoning is the same as used in explaining proxi-mal-distal differences in fossil quality and abun-dance within the Two Medicine–Judith River rec-ord (Rogers 1993): proximal portions of the forelandbasin with highest subsidence and accommodationof deposition (Two Medicine) do in fact containmore high-resolution, non-channel-lag type con-centrations than in distal (Judith River) portions ofthe clastic wedge.

Relative burial potential for the stratigraphic in-tervals under scrutiny can be estimated using 40Ar/39Ar ages and thickness data (Rogers et al. 1993;Rogers 1994, 1995, 1998; Rogers and Swisher 1996).The R8 regressive depositional system yields rockaccumulation rates ranging from 6 to 7 cm/1000yr, whereas the T9 transgressive depositional sys-tem is characterized by a net rate of rock accu-mulation on the order of 10–11 yr. Thesecm/1000values are consistent with higher net rates of skel-etal burial in the transgressive record.

Differential Diagenesis. In marine transgressive-regressive cycles, carbonates within regressivephases experience more severe meteoric diagenesisthan their counterparts in transgressive deposits,leading to significantly different cementation lev-els and grain preservation state (Heckel 1983;Braithwaite 1993). If there is a nonmarine analogto this diagenetic asymmetry, then we would alsoexpect more favorable (less destructive) diagenesisof vertebrate skeletal material in transgressive than

in regressive phases, ignoring relative likelihoodsof destruction during later diagenesis and tecton-ism. Phreatic diagenesis (below the water table)will presumably play a larger role in nonmarinefacies deposited during transgression (baselevel riseand high accommodation) than during regression(steady or falling baselevel), where the same faciesare more likely to experience near surface vadosediagenesis (above the water table).

The likelihood of bone preservation in a sedi-ment profile is governed by several parameters, in-cluding the movement of pore waters (potential formobilization and flushing of soluble ions), pH andEh values, and the abundance of destructive micro-organisms (Newesely 1989). In terrestrial deposi-tional systems, destructive diagenetic processeswould presumably be most severe in the vadosezone, where pore waters would be apt to circulatemore quickly through the sediment profile (Tucker1991), Eh values would presumably be higher, andmicroorganisms (e.g., fungi, bacteria) would beabundant and more diverse or active. Undersatur-ation with respect to CaCO3 is also typical of somevadose pore fluids, especially in the upper reachesof the soil zone, and this would potentially en-courage bone corrosion and dissolution. In contrast,the fully phreatic environment would be typifiedby more restricted movement of pore fluids (exceptin the immediate vicinity of the water table), lowerEh values, and probably less skeletal corrosion andleaching.

Environmental and Ecological Contrasts. The spa-tial mosaic of sedimentary environments on theTwo Medicine–Judith River alluvial-coastal plain,which was bounded to the west by tectonic high-lands and to the east by the Cretaceous InteriorSeaway, would have fluctuated dramatically duringmajor transgressive-regressive cycles. During trans-gression, the regional gradient would have almostcertainly diminished (especially if transgressionwas subsidence related). This reduction in regionalgradient would have presumably led to significantlandward expansion of lower coastal plain environ-ments. In fact, Eberth (1996) has estimated thattidal influence on the Judith River coastal plain ofsouthern Alberta extended inland for more than200 km during transgression of the Bearpaw Sea.Based on our taphonomic data and comparable ob-servations from southern Alberta and Saskatche-wan (Brinkman 1990; Eberth 1990; Eberth et al.1990), this broad belt of coastal and paralic envi-ronments was extremely conducive to the preser-vation of vertebrate skeletal concentrations (pre-dominantly vertebrate microsites and ceratopsianbone beds). This preservational bias may reflect the


productivity of aquatic and terrestrial coastal com-munities, the favorable geochemistry of hydro-morphic coastal environments, or both.

Although admittedly more speculative, trans-gression would also have potential taphonomic im-pacts on inland settings. During the transgressivephase, as coastal plain environments backsteppedinland toward the mountain front, upland environ-ments on the Two Medicine alluvial plain wouldpresumably become more areally restricted. Withdiminishing habitat, taxa adapted for inland set-tings may have experienced stress due to more lim-ited resources and perhaps heightened competition.Resource limitations in turn might have led toheightened mortality and perhaps widespread massmortality. This prediction is consistent with theabundance of mono- and paucispecific dinosaurbone beds preserved in the transgressive deposi-tional system of the Two Medicine Formation (Rog-ers 1990, 1993, 1995; Varricchio 1995). These“event” dinosaur bone beds indicate that massmortality was a common occurrence on the TwoMedicine alluvial plain during transgression of theBearpaw Sea.

Conclusions

Based on our taphonomic analysis of the Campan-ian Two Medicine–Judith River record, there areclearly fundamental differences in the ways inwhich bioclastic material accumulates in marineand terrestrial settings, and it is also clear that ver-tebrate skeletal concentrations are less reliablecues to stratigraphically significant surfaces. Thedistribution of vertebrate skeletal material is notpredictable with regard to the scale or duration ofstratal discontinuities in either marine or terres-trial portions of the study interval, based on ourexamination of third- and fourth-order sequenceboundaries, the nonmarine extension of a third-order transgressive surface, marine flooding sur-faces, and localized fluvial scours. Instead, the ver-tebrate skeletal signatures of hiatal surfaces of alldurations closely track the fossil content of under-lying and lateral facies: relative concentrations ofbones and teeth occur where discontinuity surfacescut across preexisting beds with appreciable skel-etal content.

The distribution of vertebrate fossils on hiatalsurfaces in the Two Medicine–Judith River recordis thus controlled by erosion potential and the bio-clastic composition of facies that may be geneti-cally unrelated to those of the hiatus. This resulthas important implications for stratigraphic anal-ysis in nonmarine depositional systems, where the

heterolithic mix of facies would seem to precludethe widespread development of skeletal-lag-drapederosional discontinuities. The spatial patchiness ofvertebrate populations in most terrestrial settings,combined with overall lower fecundity relative tomost marine vertebrate and invertebrate counter-parts, may actually inhibit the generation of ap-preciable skeletal concentrations on hiatal surfacesunder any conditions, especially concentrationsthat are laterally persistent.

Documenting the basic patterns of fossil distri-bution and assessing their controls is also key toestimating the possible bias of paleobiological in-formation and to developing more effective strat-egies for prospecting vertebrate occurrences. Ouranalysis indicates that the vast majority of verte-brate skeletal concentrations associated with dis-continuity surfaces are time averaged lag-typemixtures of ecologically and, in some instances,taphonomically, disparate skeletal material. Thedegree of time averaging in such concentrations hasbeen estimated to range in excess of 100,000 yr(Rogers 1993); this level of temporal mixing rendersit plausible that noncontemporaneous vertebratecommunities are admixed in at least some lag as-semblages. Interestingly, although such lags wouldtypically have relatively low levels of temporal acu-ity, they may have relatively high degrees of spatialfidelity. The direct association of lag concentra-tions and floodplain source beds in alluvial settings(fig. 7), coupled with evidence of lateral variationin the composition of skeletal lags on fourth-ordermarine flooding surfaces (presumably a function ofunderlying facies content), strongly suggests thatlittle lateral transport of skeletal material occurredafter exhumation.

Finally, our results reveal some broad and un-expected trends in the abundance of vertebrate fos-sil assemblages, both across environments andthrough stratigraphic cycles. Among environmentsin this study interval, tidally influenced channeldeposits and shoreface deposits are consistentlymore productive than surrounding strata (table 1).At the basin scale, strata deposited during the R8-T9 third-order cycle show a striking contrast inpreservation potential (table 2), with the transgres-sive depositional system considerably more fossil-iferous than the underlying regressive depositionalsystem, in both shallow marine and terrestrial set-tings. These findings highlight the control of facieson the distribution of vertebrate fossils and dem-onstrate the importance of this association at scalesranging from the individual deposit to the third-order cycle. The link between physical patternsof sedimentation and patterns of vertebrate fossil


distribution is unequivocal within the TwoMedicine–Judith River record, and it is certainlyreasonable to conclude that spatial and temporalvariations in erosion, omission, and depositionmight exert a similarly strong influence on the dis-tribution and quality of vertebrate fossil assem-blages in other marine and terrestrial systems.

A C K N O W L E D G M E N T S

We thank C. Badgley, K. Curry Rogers, R. Dunn, J.Thole, and two anonymous reviewers for theirhelpful comments on the manuscript. Logistical

support in the Missouri Breaks was provided by L.Eichhorn and J. Mitchell (Bureau of Land Manage-ment, Lewistown District). The Blackfeet Nationkindly granted access to the Two Medicine typearea. Funding for fieldwork was provided by Sigma-Xi, the Geological Society of America, the Amer-ican Museum of Natural History (Theodore Roo-sevelt Fund), the Dinosaur Society, and the Donorsto the Petroleum Research Fund of the AmericanChemical Society, whose support is gratefully ac-knowledged. Funding from the Pew Midstatesshort-term consultation program expedited com-pletion of this report.

R E F E R E N C E S C I T E D

Abbott, S. T. 1998. Transgressive systems tracts and on-lap shellbeds from mid-Pleistocene sequences, Wang-anui Basin, New Zealand. J. Sediment. Res. 68:253–268.

Aepler, R., and Reif, W.-E. 1971. Origin of bone-beds.Abstr. 8th International Sedimentology Congress.

Allen, J. R. L. 1978. Studies in fluviatile sedimentation:an exploratory quantitative model for the architectureof avulsion-controlled suites. Sediment. Geol. 21:129–147.

Anderson, L. C., and McBride, R. A. 1996. Taphonomicand paleoenvironmental evidence of Holocene shell-bed genesis and history on the northeastern Gulf ofMexico shelf. Palaios 11:532–549.

Behrensmeyer, A. K. 1987. Miocene fluvial facies andvertebrate taphonomy in northern Pakistan. Soc.Econ. Paleontol. Mineral. Spec. Publ. 39:169–176.

———. 1988. Vertebrate preservation in fluvial channels.Palaeogeogr. Palaeoclimatol. Palaeoecol. 63:183–199.

———. 1991. Terrestrial vertebrate accumulations. In Al-lison, P. A., and Briggs, D. E. G., eds. Taphonomy:releasing the data locked in the fossil record. Top. Geo-biol. 9:291–335.

———. 1999. Bonebeds through geologic time. J. Vertebr.Paleontol. 19(suppl. 3):31A–32A.

Behrensmeyer, A. K., and Chapman, R. E. 1993. Modelsand simulations of time-averaging in terrestrial ver-tebrate accumulations. In Kidwell, S. M., and Beh-rensmeyer, A. K., eds. Taphonomic approaches to timeresolution in fossil assemblages. Paleontol. Soc. ShortCourses Paleontol. 6:125–149.

Behrensmeyer, A. K.; Gordon, K. D.; and Yanagi, G. T.1989. Non-human bone modification in Miocene fos-sils from Pakistan. In Bonnichsen, R., and Sorg, M.H., eds. Bone modification. Proc. First InternationalConference on Bone Modification. Orono, Maine,Cent. Stud. First Am., p. 99–120.

Blob, R. W.; Carrano, M. T.; Rogers, R. R.; Forster, C. A.;and Espinoza, N. R. 1997. New taxonomic and ta-phonomic data from the herpetofauna of the Judith

River Formation (Campanian), Montana. J. Vertebr.Paleontol. 17(suppl. 3):32A–33A.

Braithwaite, C. J. R. 1993. Cement sequence stratigraphyin carbonates. J. Sediment. Petrol. 63:295–303.

Brett, C. E. 1995. Sequence stratigraphy, biostratigraphy,and taphonomy in shallow marine environments. Pa-laios 10:597–616.

Brett, C. E., and Baird, G. C. 1993. Taphonomic ap-proaches to temporal resolution in stratigraphy: ex-amples from Paleozoic marine mudrocks. In Kidwell,S. M., and Behrensmeyer, A. K., eds. Taphonomic ap-proaches to temporal resolution in fossil assemblages.Paleontol. Soc. Short Courses Paleontol. 6:250–274.

Bridge, J. S., and Leeder, M. R. 1979. A simulation modelof alluvial stratigraphy. Sedimentology 26:617–644.

Brinkman, D. B. 1990. Paleoecology of the Judith RiverFormation (Campanian) of Dinosaur Provincial Park,Alberta, Canada. Palaeogeogr. Palaeoclimatol. Pa-laeoecol. 78:37–54.

Campbell, C. V. 1967. Lamina, laminaset, bed, and bed-set. Sedimentology 8:7–26.

Carrano, M. T.; Blob, R. W.; Flynn, J. J.; Forster, C. A.;and Rogers, R. R. 1995. Additions to the fauna of theJudith River Formation (Campanian) type area, north-central Montana, with possible range extensions oftwo genera of eutherian mammals. J. Vertebr. Paleon-tol. 15(suppl. 3):21A.

Carrano, M. T.; Blob, R. W.; Flynn, J. J.; Rogers, R. R.;and Forster, C. A. 1997. The mammalian fauna of theJudith River Formation type area (Campanian, centralMontana) revisited. J. Vertebr. Paleontol. 17(suppl. 3):36A.

Case, G. R. 1978. A new selachian fauna from the JudithRiver Formation (Campanian) of Montana. Palaeon-togr. Abt. A 160:176–205.

Christie-Blick, N., and Driscoll, N. W. 1995. Sequencestratigraphy. Annu. Rev. Earth Planet. Sci. 23:451–478.

Davies, D. J.; Powell, E. N.; and Stanton, R. J., Jr. 1989.Relative rates of shell dissolution and net sedimentaccumulation—a commentary: can shell beds form by


the gradual accumulation of biogenic debris on thesea floor? Lethaia 22:207–212.

DeGraw, H. M. 1975. The Pierre-Niobrara unconformityin western Nebraska. In Caldwell, W. G. E., ed. TheCretaceous system in the western interior of NorthAmerica. Geol. Assoc. Can. Spec. Pap. 13:589–606.

Doyle, P., and MacDonald, D. I. M. 1993. Belemnite bat-tlefields. Lethaia 26:65–80.

Eberth, D. A. 1990. Stratigraphy and sedimentology ofvertebrate microfossil localities in uppermost JudithRiver Formation (Campanian) of Dinosaur ProvincialPark. Palaeogeogr. Palaeoclimatol. Palaeoecol. 78:1–36.

———. 1996. Origin and significance of mud-filled in-cised valleys (Upper Cretaceous) in southern Alberta,Canada. Sedimentology 43:459–477.

Eberth, D. A.; Bramen, D. R.; and Tokaryk, T. T. 1990.Stratigraphy, sedimentology, and vertebrate paleon-tology of the Judith River Formation (Campanian) nearMuddy Lake, west-central Saskatchewan. Bull. Can.Pet. Geol. 38:387–406.

Fiorillo, A. R. 1991. Taphonomy and depositional settingof Careless Creek Quarry (Judith River Formation),Wheatland County, Montana, U.S.A. Palaeogeogr. Pa-laeoclimatol. Palaeoecol. 81:281–311.

Fischer, A. G. 1961. Stratigraphic record of transgressingseas in light of sedimentation on Atlantic coast ofNew Jersey. Am. Assoc. Petrol. Geol. Bull. 45:1656–1666.

Fursich, F. T., and Oschmann, W. 1993. Shell beds as toolsin basin analysis: the Jurassic of Kachchh, westernIndia. J. Geol. Soc. Lond. 150:169–185.

Garcia, J. P., and Dromart, G. 1997. The validity of twobiostratigraphic approaches in sequence stratigraphiccorrelations; brachiopod zones and marker-beds in theJurassic. Sediment. Geol. 114:55–79.

Gibling, M. R., and Bird, D. J. 1994. Late Carboniferouscyclothems and alluvial paleovalleys in the SydneyBasin, Nova Scotia. Geol. Soc. Am. Bull. 106:105–117.

Gill, J. R., and Cobban, W. A. 1973. Stratigraphy andgeologic history of the Montana Group and equivalentrocks, Montana, Wyoming, and North and South Da-kota. U.S. Geol. Surv. Prof. Pap. 776, 37 p.

Gillespie, J. L.; Nelson, C. S.; and Nodder, S. D. 1998.Post-glacial sea-level control and sequence stratigra-phy of carbonate-terrigenous sediments, Wanganuishelf, New Zealand. Sediment. Geol. 122:245–266.

Gilmore, C. W. 1917. Brachyceratops—a ceratopsian di-nosaur from the Two Medicine Formation of Mon-tana. U.S. Geol. Surv. Prof. Pap. 103, 45 p.

Hanson, M. S., and Little, L. D. 1989. Origins, stackingconfigurations, and facies distributions of genetic se-quences, Eagle Sandstone, 1989, Field ConferenceGuidebook. Billings, Montana. Mont. Geol. Soc., p.141–150.

Heckel, P. H. 1983. Diagenetic model for carbonate rocksin midcontinent Pennsylvanian eustatic cyclothems.J. Sediment. Petrol. 53:733–759.

Hooker, J. S. 1987. Late Cretaceous ashfall and the de-

mise of a hadrosaurian “herd.” Geol. Soc. Am. Abstr.Program 19:284.

Horner, J. R. 1982. Evidence of colonial nesting and sirefidelity among orhithischian dinosaurs. Nature 297:675–676.

———. 1994. Comparative taphonomy of some dinosaurand extant bird colonial nesting grounds. In Carpen-ter, K.; Hirsch, K. F.; and Horner, J. R., eds. Dinosaureggs and babies. Cambridge, Cambridge UniversityPress, p. 116–123.

Horner, J. R., and Makela, R. 1979. Nest of juvenilesprovides evidence of family structure among dino-saurs. Nature 282:296–298.

Kauffman, E. G. 1977. Geological and biological over-view: western interior Cretaceous basin. Mt. Geol. 14:75–99.

Kidwell, S. M. 1986. Models for fossil concentrations:paleobiologic implications. Paleobiology 12:6–24.

———. 1989. Stratigraphic condensation of marine trans-gressive records: origin of major shell deposits in theMiocene of Maryland. J. Geol. 97:1–24.

———. 1991. The stratigraphy of shell concentrations.In Allison, P. A., and Briggs, D. E. G., eds. Taphonomy:releasing the data locked in the fossil record. Top. Geo-biol. 9:211–290.

———. 1993a. Patterns of time-averaging in shallow ma-rine fossil assemblages. In Kidwell, S. M., and Beh-rensmeyer, A. K., eds. Taphonomic approaches to tem-poral resolution in fossil assemblages. Paleontol. Soc.Short Courses Paleontol. 6:275–300.

———. 1993b. Taphonomic expressions of sedimentaryhiatuses: field observations on bioclastic concentra-tions and sequence anatomy in low, moderate andhigh subsidence settings. Geol. Rundsch. 82:189–202.

Kondo Y.; Abbott, S. T.; Kitamura, A.; Kamp, P. J. J.;Naish, T. R.; Kamataki, T.; and Saul, G. S. 1998. Therelationship between shellbed type and sequence ar-chitecture: examples from Japan and New Zealand.Sediment. Geol. 122:109–128.

Krumbein, W. C. 1942. Criteria for subsurface recogni-tion of unconformities. Bull. Am. Assoc. Pet. Geol.26:36–62.

Liu, Y., and Gastaldo, R. A. 1992. Characteristics of aPennsylvanian ravinement surface. Sediment. Geol.77:197–213.

Malchus, N. 1990. Revision der Kreide-Austern (Bivalvia:Pteriomorphia) Agyptens (Biostratigraphie, Systema-tik). Berl. Geowissensch. Abh. A 125, 231 p.

Meldahl, K. H. 1993. Geographic gradients in the for-mation of shell concentrations: Plio Pleistocene ma-rine deposits, Gulf of California. Palaeogeogr. Palaeo-climatol. Palaeoecol. 101:1–25.

Miall, A. D. 1991. Hierarchies of architectural units inclastic rocks, and their relationship to sedimentationrate. In Miall, A. D., and Tyler, N., eds. Three-dimen-sional facies architecture of terrigenous clastic sedi-ments, and its implications for hydrocarbon discoveryand recovery. Soc. Econ. Paleontol. Mineral. Conf.Sediment. Paleontol. 3:6–12.


———. 1996. The geology of fluvial deposits. New York,Springer, 582 p.

Mitchum, R. M., Jr., and Van Wagoner, J. C. 1991. High-frequency sequences and their stacking patterns: se-quence stratigraphic evidence of high-frequency cy-cles. Sediment. Geol. 70:131–160.

Morlan, R. E. 1984. Toward the definition of criteria forthe recognition of artificial bone alterations. Quat.Res. 22:160–171.

Naish, T., and Kamp, P. J. J. 1997. Sequence stratigraphyof sixth-order (41 k.y.) Pliocene-Pleistocene cycloth-ems, Wanganui Basin, New Zealand: a case for theregressive systems tract. Geol. Soc. Am. Bull. 109:978–999.

Newesely, H. 1989. Fossil bone apatite. Appl. Geochem.4:233–245.

Posamentier, H. W., and Allen, G. P. 1993. Variability ofthe sequence stratigraphic model: effects of local basinfactors. Sediment. Geol. 86:91–109.

Reif, W.-E. 1982. Muschelkalk/Keuper bone-beds (MiddleTriassic, SW-Germany)—Storm condensation in a re-gressive cycle. In Einsele, G., and Seilacher, A., eds.Cyclic and event stratification. Berlin, Springer, p.299–325.

Rogers, R. R. 1990. Taphonomy of three dinosaur bonebeds in the Upper Cretaceous Two Medicine Forma-tion of Montana: evidence for drought-related mor-tality. Palaios 5:394–413.

———. 1993. Systematic patterns of time-averaging inthe terrestrial vertebrate record: a Cretaceous casestudy. In Kidwell, S.M., and Behrensmeyer, A. K., eds.Taphonomic approaches to time resolution in fossilassemblages. Paleontol. Soc. Short Courses Paleontol.6:228–249.

———. 1994. Nature and origin of through-going discon-tinuities in nonmarine foreland basin strata, UpperCretaceous, Montana: implications for sequence anal-ysis. Geology 22:1119–1122.

———. 1995. Sequence stratigraphy and vertebrate ta-phonomy of the Upper Cretaceous Two Medicine andJudith River formations, Montana. Unpub. Ph.D. dis-sertation, University of Chicago.

———. 1998. Sequence analysis of the Upper CretaceousTwo Medicine and Judith River formations, Montana:nonmarine response to the Claggett and Bearpaw ma-rine cycles. J. Sediment. Res. 68:615–631.

Rogers, R. R., and Eberth, D. A. 1996. Stratigraphic utilityof vertebrate microfossil assemblages in the Campan-ian of Montana and Alberta. J. Vertebr. Paleontol.16(suppl. 3):61A.

Rogers, R. R.; Forster, C. A.; Blob, R. W.; and Carrano,M. T. 1995. Late Cretaceous vertebrates by the sea:diverse non-marine assemblages in coastal facies ofthe Campanian Judith River Formation, north-centralMontana. Geol. Soc. Am. Abstr. Program RockyMountain Sect. 27:52.

Rogers, R. R., and Swisher, C. C. 1996. The Claggett and

Bearpaw transgression revisited: new 40Ar/39Ar dataand a review of possible drivers. Geol. Soc. Am. N.-Cent. Sect. Abstr. Program 28:62.

Rogers, R. R.; Swisher, C. C.; and Horner, J. R. 1993. 40Ar/39Ar age and correlation of the non-marine Two Med-icine Formation (Upper Cretaceous), northwesternMontana. Can. J. Earth Sci. 30:1066–1075.

Sahni, A. 1972. The vertebrate fauna of the Judith RiverFormation, Montana. Am. Mus. Nat. Hist. Mus. Bull.147:321–412.

Shanley, K. W., and McCabe, P. J. 1991. Predicting faciesarchitecture through sequence stratigraphy—an ex-ample from the Kaiparowits Plateau, Utah. Geology19:742–745.

———. 1994. Perspectives on the sequence stratigraphyof continental strata. Am. Assoc. Pet. Geol. Bull. 78:544–568.

Shurr, G. W., and Reiskind, J. 1984. Stratigraphic frame-work of the Niobrara Formation (Upper Cretaceous)in North and South Dakota. In Stott, D. F., and Glass,D. J., eds. The Mesozoic of middle North America.Can. Soc. Pet. Geol. Mem. 9:205–219.

Stanley, D. J., and Bernasconi, M. P. 1998. Relict andpalimpsest depositional patterns on the Nile shelf re-corded by molluscan faunas. Palaios 13:79–86.

Stanton, T. W., and Hatcher, J. B. 1905. Geology and pa-leontology of the Judith River beds. U.S. Geol. Surv.Bull. 257:1–128.

Swift, D. J. P. 1968. Coastal erosion and transgressivestratigraphy. J. Geol. 76:444–456.

Swift, D. J. P.; Phillips, S.; and Thorne, J. A. 1991. Sedi-mentation on continental margins. IV. Lithofacies anddepositional systems. Spec. Publ. Int. Assoc. Sedi-ment. Geol. 14:89–152.

Tucker, M. E. 1991. The diagenesis of fossils. In Donovan,S. K., ed. The processes of fossilization. New York,Plenum, p. 84–104.

Van Wagoner, J. C. 1995 Sequence stratigraphy and ma-rine to nonmarine facies architecture of foreland basinstrata, Book Cliffs, Utah, U.S.A. In Van Wagoner,J. C., and Bertram, G. T., eds. Sequence stratigraphyof foreland basin deposits. Am. Assoc. Pet. Geol.Mem. 64:137–223.

Van Wagoner, J. C.; Mitchum, R. M.; Campion, K. M.;and Rahmanian, V. D. 1990. Siliciclastic sequencestratigraphy in well logs, cores, and outcrops: conceptsfor high-resolution correlation of time and facies. Am.Assoc. Pet. Geol. Methods Explor. Ser. 7, 55 p.

Varricchio, D. J. 1995. Taphonomy of Jack’s Birthday Site:a diverse dinosaur bonebed from the Upper CretaceousTwo Medicine Formation of Montana. Palaeogeogr.Palaeoclimatol. Palaeoecol. 114:297–323.

Wright, V. P., and Marriott, S. B. 1993. The sequencestratigraphy of fluvial depositional systems: the roleof floodplain sediment storage. Sediment. Geol. 86:203–210.

Associations of Vertebrate Skeletal Concentrations and...

Documents

Transcript of Associations of Vertebrate Skeletal Concentrations and...