Crinoid biofacies in Upper Carboniferous cyclothems, midcontinent North America: faunal tracking and the role of regional processes in biofacies recurrence

E L S E V I E R Palaeogeography, Palaeoclirnatology, Palaeoecology 127 (1996) 47-81

Crinoid biofacies in Upper Carboniferous cyclothems, midcontinent North America: faunal tracking and the role of

regional processes in biofacies recurrence

Peter F. Holterhoff *

Department of Geosciences, Gould-Simpson Building, University of Arizona, Tucson, AZ 85721, USA

Received 8 June 1995; accepted 24 August 1995

Abstract

Coordinated stasis is a pattern of long-term, multispecies evolutionary and ecological stability bounded by rapid turnover events. One class of models proposed to explain the paleoecological pattern of biofacies stability, typified by ecological locking, states that biotas within habitats become integrated into ecological units that, through population and community-level processes, rebound and persist through short-term disturbances. This high frequency community resilience allows the units to persist through geologic time, tracking their preferred habitats during transgressions and regressions. Because species are tightly woven into the fabric of the ecosystem, there is little opportunity for speciation; this also buffers them against extinction. This ecological locking by ecosystem homeostatic processes produces long intervals of regionally stable biofacies in the fossil record. In these models, the primary structuring mechanisms are at the local, community level, controlled by species autecology, species interactions, and local disturbance regime.

An opposing class of models states that species are fundamentally independent and that species interactions are not strong nor spatio-temporally persistent enough to forge communities into integrated units with long-term persistence. Community recurrence would be expected as long as the various habitat types recur in time and space, species autecologies remain stable, the pool of available species remains intact and the recruitment of species into local habitats is not interrupted. The focus of these models is dominantly at the regional level, controlled by biogeographic history and dispersal among, as well as within, communities.

The results of this study of crinoids from the Upper Carboniferous Lansing Group of midcontinent North America show that they are distributed among five recurrent paleocommunity types, or biofacies, which are generally arrayed along an onshore-offshore gradient. Two of these biofacies are restricted to the offshore, dysaerobic horizons of the most extensive cyclothem in the group. The other three biofacies recur consistently among the cycles. However, this repeated pattern includes a regionally distributed offshore biofacies, unique to late transgressive deposits, that is not found in analogous regressive deposits. While the distribution of the two shoreward biofacies is consistent with faunal tracking, the recurrence of a biofacies unique to late transgressive horizons is not consistent with an integrated paleocommunity that tracks shifting habitats through transgression and regression. These results favor the alternative models, namely that biofacies develop in place via dispersal from a persistent species pool in response to prevailing environmental conditions.

The recurrence of apparently ephemeral biofacies through time underscores the importance of regional processes in controlling local paleocommunity composition. The establishment of refuges, or "source" populations, during environmental bottlenecks (i.e. lowstands) is critical to maintaining the regional species pool through time for continued recruitment into local "sink" paleocommunities during more amenable times (i.e. highstands). Metapopulation dynamics among habitats are equally important in maintaining species within a region through time.

* Email: [email protected].

0031-0182/96/$15.00 © 1996 Elsevier Science B.V. All rights reserved PII S0031-0182(96) 00088-0

48 P.F. Holterhoff/Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996) 47-81

Conflicting observations of stochastic (individualistic) and deterministic (integrated) behaviors of the same ecological systems probably result from differences in the scales of analyses and therefore the processes being studied.

Keywords: Pennsylvanian; crinoid; North America; palaeocommunity; palaeoecology

1. Introduction

Many paleontologists have asserted that the fossil record can be subdivided into long intervals in which multiple lineages have correlated patterns of low taxonomic turnover, ending abruptly in short intervals of accelerated turnover, then giving way to the next interval of stability (see Boucot, 1990). Recent workers have emphasized that lineages not only show coordinated patterns of turnover, but that paleoecological patterns, such as paleocommunity and biofacies membership and relative abundances within biofacies, are also stable through intervals of correlated stasis. These large scale packages of stasis have been referred to as Ecological-Evolutionary Units (Boucot, 1983, 1986, 1990; Sheehan, 1991) and the smaller packages of ecological and evolutionary stability as Ecological-Evolutionary Subunits (Brett et al., 1990; Brett and Baird, 1995). Indeed, the recognition of temporally restricted, regionally deployed, habitat-specific ecosystems is the very goal of ecostratigraphy (Hoffman, 1981). Thus, these units are defined by two separate, but interrelated, patterns: morphologic and taxonomic coordinated stasis (and turnover) among multiple lineages and stability of paleocommunity/biofacies structure through intervals of stasis (Brett and Baird, 1995; Morris et al., 1995).

The resurgence of this community evolution paradigm arose largely from the application of explicit ecological models which ascribe causal significance to paleocommunity stability in driving coordinated stasis (Brett et al., 1990; Schopf et al., 1992; Morris et al., 1992, 1995). As synthesized here, these models explain the macroevolutionary pattern of coordinated stasis as the result of hierarchical ecosystem processes driven by high frequency disturbance events, species autecologies, and species interactions (Morris et al., 1995). High frequency disturbance events produce a spatio- temporally shifting mosaic of habitat patches

which are populated by a similarly shifting, but limited, set of species (Miller et al., 1988). Limited membership into these local ecosystems is produced by species autecologies (substrate prefer- ences, feeding type, etc.) as well as species interactions (substrate competition, predation, etc.). This limited membership into local ecosystems produces resilient community-level responses at larger spatio-temporal scales. In other words, the composition of several neighboring patches or an individual patch through time may not be stable, but the composition of the community through time would be as it is at this level at which the disturbances are incorporated into the system (O'Neill et al., 1986). Ecosystem processes, dominantly species interactions, bind species within communities together, locking them into nets of interactions which limit deviations in population structure and phenotype and preclude establishment of immigrants. These self-regulated, fully incorporated communities then track bio- topes as they migrate through a depositional basin in response to changing sea-level and sedimentary regimes, producing long-term paleoecological stasis. Large-scale perturbations rupture the long- term, community-level constraints on the system, producing rapid changes within the biota.

Competing biofacies recurrence models relegate species interactions to secondary importance and emphasizes the correlated, but individualistic, responses of species adapted to similar environmental conditions (Bambach, 1994). Like the paleoecological stasis model, hierarchical ecosystem processes can also be incorporated into these models as an important process in buffering communities from high frequency disturbances among patches. However, species habitat require- ments, the composition of the species pool, and physically and biotically-mediated dispersal ability control community membership. Thus, as long as species remain adapted to their preferred environments and maintain consistent levels of eurytopy

P.£ Holterhoff/Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996) 47-81 49

(autecological stasis), species populations are maintained within the region, and physical conditions are such that recruits can be dispersed to wherever these environments exist (Miller, 1990, 1993a,b; species pool of Buzas and Culver, 1994), communities will recur through the fossil record.

Building on these brief descriptions, other differences can be highlighted between the two classes of models. In models akin to ecological locking, the regulation of long-term biofacies structure is dominantly at the local, community level, dictated by local disturbances, species autecologies, and species interactions; the community is an emergent entity incorporating these components. Additionally, observations that species invasions or recruitment failures (incursion and "outage" epiboles, respectively [Brett et al., 1990]) are short lived, with the community rapidly returning to its original structure (Brett et al., 1990), fortifies the notion that communities are relatively closed systems. Closed system responses of communities are borne from the self-regulating properties of ecosystems, dominantly controlled by interactions and stereotypical responses to disturbance (Morris et al., 1995). Thus, the composition and diversity of the regional fauna is controlled by these local processes, coupled with the turnover of species among environments, as dictated by levels of eurytopy and the steepness of ecotonal gradients. Long- term, biogeographically large-scale migration of these local habitat patches in response to fluctuating sea-level and sedimentary regimes explains the pattern of community-level faunal tracking in the fossil record.

In the individualistic response models, regulation of long-term biofacies structure is at the regional level, with consistent dispersal of species among habitats during shifting environmental conditions responsible for maintaining paleocommunity identity (Plotnick et al., 1994; Miller, 1990, 1994). Whereas the hierarchical relationship between patch dynamics and community emer- gence may be the same as with ecological locking models, individualistic communities are more open systems. These communities are not self-regulating, as they are dominantly controlled by the flux of species within a region, the physical processes controlling the dispersal and local recruitment of

species, along with local environmental conditions (Underwood and Fairweather, 1989; Roughgarden et al., 1994; Cohen, 1994). Long-term stability and recurrence of paleocommunities is more a reflec- tion of stability of the regional ecosystem, encom- passing the regional species pool, dispersal dynamics, species autecologies, and recurrence of particular habitat types than internal regulation and persistence of local ecosystems tracking preferred habitats (Miller, 1990, 1994).

A vexing problem is that recurrent paleocommunities and biofacies associated with particular depositional environments in the fossil record can be explained by both models. However, the distribution of biofacies within and among cyclic depositional sequences can provide important insights into the mechanisms controlling paleocommunity structure and persistence. Eustatic cycles which have well-developed transgressive and regressive facies have major lithotopes (nearshore clastics, nearshore carbonates, shelf carbonates, offshore ctastics, etc.) that are symmetrically distributed (but not necessarily equal in thickness) about the cycle boundaries and maximum flooding horizons (Fig. 1). Like the major lithotopes, biofacies are also distributed symmetrically within these cycles. The important inference often made from analyses of symmetrical cycles is that analogous depth- controlled depositional systems with their con- tained paleocommunities were continually present, tracking sea-level during transgressions and regressions. Indeed, the recognition of symmetrically- distributed biofacies among the Middle Devonian Hamilton Group cycles (Savarese et al., 1986; Brett and Baird, 1986; Brett et al., 1990; although see Miller et al., 1988 for a counter example) was an important factor leading to the development of the ecological locking model, although the recognition of biofacies symmetry has long been important in interpreting symmetrical cyclic sequences (see Elias, 1937; Heckel, 1977; Boardman et al., 1984 for upper Paleozoic examples). However, recognition of faunal tracking in these stratal packages supports both models, as species could have responded independently to long-term environmental fluctuations (Brett et al., 1990) or as integrated units (Morris et al., 1995), although the vagaries of community assembly may make the

50 P.F. Holterhoff/Palaeogeography, Palaeoclirnatology, Palaeoecology 127 (1996) 47-81

J J .E T,VE SEA- LEVEL ~\0 k~ ~

CB

T R

Symmetrical T - R Cycles

RELATIVE SEA - LEVEL

MFS/CB

M FS/C B

MFS/CB

T N

Asymmetdc , Shal lowing - Upward Cycles

Regressive fine siliciclastics

Regressive offshore carbonates

Regressive nearshore carbonates

Regressive coarse siliciclastics

Transgressive fine siliciclastics

Transgressive offshore carbonates

~ Transgressive nearshore carbonates

Transgressive coarse siliciclastics

Fig. 1. Symmetrical versus asymmetrical depositional cycles and their biofacies sequences. Symmetrical depositional cycles record a series of deepening, then shallowing, environments such that the litho- and biofacies sequences are mirror images about the cycle knick points (Cycle Boundary = CB; Maximum Flooding Surface = MFS); however, the thicknesses of these facies sequences may not be equal. In asymmetrical cycles, the transgressive succession of facies are often condensed into a single, complex bed (Kidwell, 1991a,b) or omitted altogether, merging the CB and the MFS. In this case, the transgressive biofacies successions are telescoped into a single biofacies (A), thus few details can be observed of the ecological dynamics of rising sea-level.

P.F. Holterhoff/Palaeogeography, Palaeoclimatology, Palaeoecology 12 7 (1996) 4 7-81 5 l

pattern of long term faunal tracking unlikely in a purely independent, random system (Drake, 1990; Schopf, 1996).

Biofacies from cyclic sequences which do not demonstrate faunal tracking may be more powerful in addressing these issues. Two patterns may emerge from these systems. In the first, successive cycles may contain different paleocommunities altogether, indicating a high level of instability. This appears to be the pattern following major extinctions (Sheehan, 1991; Schubert and Bottjer, 1995). Lack of coherent faunal tracking and a high level of instability among cycles could support models involving ecological locking, in that self- regulation, as a result of recovery from a turnover event or from repeated environmental stress (i.e. high amplitude/high frequency sea-level change) has yet to buffer the community from biologic and physical disturbances. However, if many of the species are unique to the respective cycles, it could also be argued that the regional species pool is what is truly unstable, as reflected in the unstable composition of local paleocommunities. Similar depositional environments harboring different species assemblages between these cycles may truly be reflecting differences in assembly history (Drake, 1990; Schopf, 1996).

In the second pattern, a different sequence of transgressive and regressive biofacies recurs among successive cycles. Asymmetrical cycles are obvious examples in which the entire succession of transgressive environments is missing or highly condensed and therefore any evidence of the ecological behavior of communities during rising sea-level is lost (Fig. 1 ). In many cases, discordant transgressive-regressive biofacies patterns are probably saying as much about the depositional dynamics of rising and falling sea-level as they are about community structure; hardground paleocommunities on transgressive surfaces are an excellent example (i.e. Fursich et al., 1991). However, the point here is that the recurrence of biofacies unique to deepening or shallowing deposits, without a coun- terpart in the opposing portion of the sea-level curve, among several cycles strongly indicates that depositional systems were not harboring stable paleocommunities tracking sea-level, or even that the depositional systems themselves were persistent

through time. In this case paleocommunities, the recurrent but ephemeral paleocommunities in particular, would have to be composed of taxa differentially recruited from the regional species pool as temporally changing environmental conditions dictate. This pattern is more consistent with predictions made by the individualistic class of models than with models predicting faunal tracking. Recurrence in this situation also indicates that the regional species pool was stable, contrary to the pattern mentioned above. While a variety of factors could potentially affect the recruitment and establishment of species into local communities, the recurrence of conditions unique to transgression or regression, along with the maintenance of, and recruitment from, a stable regional species pool, would produce a pattern of ephemeral biofacies recurrence.

Ideally, a succession of high frequency (less than species durations), high amplitude transgressive-r- egressive depositional cycles should be analyzed for their characteristic patterns of biofacies recurrence. Upper Paleozoic cyclothems of midcontinent North America are just such high frequency (400 ka), high amplitude (approximately + 100 m)(Heckel, 1977, 1986) eustatic cycles well suited for this analysis.

2. Setting and methods

2.1. Stratigraphic framework

This study focuses on the distribution of crinoids within upper Paleozoic cyclothems of midcontinent North America. Brett and Baird (1986) noted many similarities in scale and facies between the Middle Devonian Hamilton Group cycles of New York and the Upper Pennsylvanian Missourian Stage cyclothems of the midcontinent (Heckel, 1977) - - comparisons which are significant given the prominence that the Hamilton Group has played in the coordinated stasis debate (Brett and Baird, in press). In general, stable shelf cycles of both are of comparable thicknesses (5-10 m) and generally can be characterized as symmetrical successions of facies (subequal development of transgressive and regressive deposits). However,

52 P.F Holterhoff/Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996) 47-81

Hamilton cycles are dominated by offshore deposits with the shallowest units represented by thin, shallow subtidal limestones deposited above wave base (Brett and Baird, 1985, 1986) whereas Missourian cycles have thin offshore units and are dominated by subtidal and nearshore carbonates and clastics, including non-marine deposits (Heckel, 1977; Joeckel, 1989); thus, the "polarity" of the two suites of cycles are opposite. Both suites of cycles grade laterally into thick offshore, siliciclastic-rich basinal deposits and eventually grade into very thick, nearshore and non-marine successions into which the marine portions of the cycles pinch out (Brett and Baird, 1986; Bennison, 1984, 1985). Individual cycle periodicities also appear to be the same for the two successions, thus analogous horizons among successive cycles are separated by approximately 400 ka (Brett and Baird, 1986; Heckel, 1986), which corresponds to the Milankovitch secondary eccentricity cycle (Imbrie, 1985).

The primary interval of study is the upper Missourian (Stephanian) Lansing Group of the northern midcontinent and equivalent strata in Oklahoma (Fig. 2). The cyclothems which comprise the Lansing Group (the Plattsburg, Stanton, and South Bend) have been extensively described in field guides and survey reports (Harbaugh et al., 1965; Burchett and Reed, 1967; Burchett, 1971; Cocke and Strimple, 1974; Heckel, 1975a,b, 1978; Heckel et al., 1979; Watney et al., 1985, 1989; Heckel and Pope, 1992; Pabian and Strimple, 1993; see references in Sorensen et al., 1989). General reviews of the depositional processes and tectono- stratigraphic setting of the midcontinent during the Late Pennsylvanian can be found in Heckel (1977, 1980, 1986), Boardman et al. (1984), and Watney et al. (1985, 1989).

All three cycles of the Lansing Group are geographically divided into facies belts that reflect persistent differences in sedimentary regime and subsidence - - these differences producing significant shelf-to-basin topography (Fig. 3; Heckel, 1977; Watney et al., 1989). Numerous other features, such as local highs over anticlines and basement monadnocks (Ireland, 1955; Gay, 1989; Bennison, pers. comm.) or sedimentary wedges (Underwood, 1984; Arvidson, 1990) also had an

important impact on local to subregional sedimentation. However, it is clear that eustatic rise and fall of sea-level was the dominant control on regional sedimentation with these other factors simply controlling how the cycles were lithologi- cally manifest (Heckel, 1978; Boardman et al., 1984; Bennison, 1984, 1985). The localities used in this study appear in Appendix 1; more detailed descriptions of these localities can be found in Holterhoff (1993).

2.2. Materials

2.2.1. Sampling Specimens used in this study were derived from

field sampling and museum collections. Field sam- piing involved surface collecting as well as surface residue and in situ bulk sampling. Residue and bulk samples weighed between 5 and 10 kg and were either oven dried, soaked in light kerosene, and washed through a 500 gm sieve or simply washed with water through the sieve, depending upon the amount and type of matrix. Samples were picked using a low magnification jeweler's visor or binocular microscope and specimens separated into their respective taxa and counted. Material recovered during the course of this inves- tigation is reposited in the University of Cincinnati Geology Museum (UCGM).

Museum samples were generally restricted to "stratigraphic collections", which range from well curated collections to unwashed, unsorted material that had not been cataloged into the collections. If the locality and stratigraphic data could be verified in the field, these samples were used. Museums from which collections were studied include the Springer Collection of the National Museum of Natural History (NMNH), the University of Iowa Geology Museum (SUI), and the University of Nebraska State Museum (UNSM).

2.2.2. Taphonomy and counting The only assumptions concerning the occurrence

of individuals and taxa within a sample are that: (1) each sample represents populations which experienced the same environmental conditions through a limited period of time (within-habitat

P.F. Holterhoff/Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996) 47-81 53

southern

E LL

I "South Bend Cyclothem"

h

= a. "~ "Stanton Cyclothem" 2 E

o .c

c~ Torp~c Sst.

I Birch Creek I

"Plattsburg Cyclothem"

Osage Co., Oklahoma

northern

E LL C

I--

)rpedo

E 14.

m

E LL C

southeastern

Little Kaw Lmst. Mern.

Onion Cr. Sat.

Rock Lake- Eudora Mere.

Tyro Captain Creek Oolite Lmat. Mere.

Was - Lane Shale

"Plattsburg Cyclothem"

Kansas

LL

northeastern

Little Kaw Lmst. Mem.

Rock Lake E

Stoner Lm~.Mem.

Eudora Sh. Mem.

Captain Cree~ Lmat. Mem.

Vilas Sha

Spring Hill Lmst.Mem.

Hickory Creel Sh. Mem. Merriam

_ Lmst. Mem.

southeastern Nebraska

Kitaki Lmst. Mem. "E

Gretna Sh. Mere. m

Little Kaw Lmst, Mem. u~

Rock Lake Shale

Kiewitz Stoner Sh. Bed Lmst. Mere.

Dyson Hollow Lmst. Bed

Eudora Sh. Mere.

Captain Creek Lmst. Mem.

Vilas Shale

E Spring Hill Lmst.Mem. u.

Hickory Creek Sh. Mem.

Merriam Lmst. Mere.

o

Fig. 2. Lithologic units of the study interval. Name changes reflect lateral facies changes of the cycles from the northern midcontinent into Oklahoma. Names in quotes are informal stratigraphic units as correlated into the basinal facies belt.

time averaged sensu Kidwell and Flessa, 1995) and (2) individuals are sampled (both taphonomi- cally and during collection) such that the relative abundances of species in the sample are propor- tional to their relative abundances in that environment during deposition. Both of these conditions may actually be common for marine fossil assemblages (Kidwell and Flessa, 1995).

These time averaged samples obscure many short term spatial and temporal, within community (i.e. patch) dynamics. Thus, collected samples do not necessarily represent "community snapshots", but rather the aggregate biotic composition of a depositional environment during a geologically limited period of time. These individual samples are here considered local paleocommunities in that each is "unique" in its exact biotic composition, paleogeographic position, and paleoenvironmental setting (Miller, 1990). Groups of local paleocommunities with similar biotic composition are here

referred to as biofacies. Biofacies actually represent different classes of local paleocommunities (Miller, 1990), united in this case by similarity of taxonomic composition and abundance as identified by multivariate analyses.

Regarding the second assumption, the muscular and ligamentary articulations of the multiplated skeletons of echinoderms lead to rapid disarticulation, potentially obscuring relative abundance patterns (Lewis, 1980; Lane and Webster, 1980; Meyer and Meyer, 1986). Even given this universally rapid rate of disarticulation, styles and rates of disintegration among subtaxa of the major echinoderm groups can be significantly different (Greenstein, 1991; Meyer et al., 1989). Indeed, as pointed out by Meyer et al. (1989), some crinoid morphotypes may be more prone to disarticulation than others, regardless of taphonomic setting. I have observed in this study that, although the isolated plates of some taxa were relatively


11 10 9

MFS1

MFS2

MFS3

A

B

' l ~ - ' ~ NEBRASKA "~ ,

i \ i ~,

/ KANSAS \',.h ! I Bourbon ~ 3 / _ . i Arch i

Cherokee t I ' ~ " - - OKLAHOMA Basin 9 ~" . . . . . . . ] 10

i Anadarko 11 , Basin Arkoma '

Wichita- I . Basin , ~l A A! Arbuckle - I,,..,, A A A A A

Ouachita Mtns. -" '"- ' A A J

t N

Forest City

Basin

I Chautauqua Arch

P.F. Holterhoff/Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996) 4~81 55

common, their articulated remains were rarely, if ever, encountered. This systematic preservational bias requires that disarticulated remains be used to assess the composition and diversity of crinoids, as recommended by Lane and Webster (1980).

Most disarticulated remains of theca can be confidently identified to the generic level; however some groups, generally composed of two to four closely related genera, could only be grouped into generic complexes, which are essentially subsets of families. In any event, the criteria used for counting individuals within each taxa were as follows: (1) articulated crowns, dorsal cups, or portions of cups which include a unique element (e.g. infrabasal circlet or anal series) are counted as one individual; (2) disarticulated unique elements (e.g. fused infrabasal cone of Paragassizocrinus or any of the plates comprising the anal series of cromyocrinids) are counted as one individual; (3) diag- nostic radials and primibrachials were counted in multiples of five (e.g. the first five plates equals one individual, 6-10 plates two individuals, 11-15 plates three individuals, etc.). For the Lansing Group, this conservative procedure yielded 11,465 individuals from 27 different genera or generic complexes tallied among the 34 samples used in the quantitative analyses (Appendix 2).

2.2.3. Quantitative analyses The abundances of taxa for each sample were

percent transformed to reduce spurious results produced by differences in sample size. Pairwise Pearson correlation coefficients were calculated among the samples and variables to assess the degree of covariation of taxon abundances between samples (Q-mode) and abundances among samples between taxa (R-mode) (Sokal and Rohlf, 1981). The correlations among samples and taxa were used as measures of similarity in the multivariate analyses.

A variety of multivariate techniques were used to identify patterns in the distribution of crinoids within and among the cycles of the Lansing Group. Cluster analysis is a classification technique that divides objects into groups or clusters which have not been pre-defined (Kovach, 1988; Gnanadesikan et al., 1989). It is an excellent exploratory technique to ascertain which subsets of samples and/or variables are similar. Agglomerative, hierarchical clustering ties together objects with the highest levels of similarity and continues to tie together objects and clusters at lower levels until all objects and clusters are united into a single cluster at the lowest level of similarity. Q-mode and R-mode clusters can be combined into a "two-way" cluster diagram with the abundance matrix highlighting the structure of the two dendrograms (Springer and Bambach, 1985; Miller, 1988). For this study clustering was performed using the "unweighted pair-group method with arithmetic averaging" (UPGMA) of Sneath and Sokal (1973), run on NTSYS - - PC version 1.50 of Rohlf (1989).

One drawback to cluster analysis is that it is a divisive technique which produces discrete groups whether they exist in the data set or not. Thus, inappropriate groupings can be forced upon data which are fundamentally gradational. An ordination was performed to help elucidate the gradational nature of the samples. Unlike cluster analysis, these techniques simply attempt to locate samples within ordination space defined by interobject distances (as defined by pair-wise dis- similarities [1-similarities]) rather than assigning membership into groups, although, if the data are organized into discrete groups, this would be reflected in the ordination (Springer and Bambach, 1985).

Non-metric multidimensional scaling (MDS), executed using SYSTAT (1986), version 3 was

Fig. 3. Study area and regional cross section. Lithologic symbols as in Fig. 1. Numbered stations on map correspond to stations along cross section; horizontal distances along cross section not to scale. A. Due to lateral pinch out and facies changes, datums vary along the cross section. Although awkward, this scheme more accurately reflects the actual topography during Lansing deposition. Datum for stations 1-6 is maximum flooding surface (MFS1) of the South Bend Formation; datum for stations 6 10 is MFS3 of Plattsburg Formation equivalent in the upper Warm Formation; datum for station 11 is base of Captain Creek-equivalent in the Barnsdall Formation. Note the thin, ramp-like geometries to the north expanding southward to the shelf margins at stations 5 and 6. Note also the well defined bathymetric basin at stations 7 and 8, pinching out southward against prograding terrigenous wedges. B. Regional setting, cross section location, and major paleogeographic features of midcontinent North America.


used to generate a two-dimensional scatter plot of the samples. This technique assumes a monotonic relationship between dissimilarity scores and ordination distances and thus uses the rank order of the pair wise dissimilarity scores, not their magni- tudes, to generate the ordination distances (Fasham, 1977). The difference between the dissimilarity scores and the ordination distances is called stress and is a measure of the distortion generated by the ordination process of the original similarity relationships among the samples (Fasham, 1977; Gnanadesikan, 1977; Rohlf, 1989). Different levels of stress have been given informal "goodness of fit" terms, ranging from perfect (0.00) to poor (>0.40), to succinctly convey how severely the original dissimilarity matrix has been distorted (Rohlf, 1989).

To help identify the nature of the distortion in the MDS plot, a minimum spanning tree (MST) can be plotted on the ordination (Rohlf, 1989). A MST is a nearest neighbor plot that connects objects with the greatest similarity, links all objects to at least one other object, and minimizes the sum of the distances between all objects (Warshauer and Smosna, 1981; Bjerstedt, 1988). The MST for this study was generated using NTSYS - - pc version 1.5 of Rohlf (1989). The MST can also identify the major gradient upon which the data may be organized by the plot of samples along the main axis of the MST (Bjerstedt, 1988).

To assess the significance of paleocommunity membership into these biofacies, canonical variate analysis (CVA) was performed. CVA, like principal components analysis, seeks to represent more clearly the data set by reducing it to a few composite dimensions, the canonical variates. However, the goals are quite different in that CVA attempts to discriminate between previously defined groups by minimizing within-group variance while maxi- mizing the between-group variation (Albrecht, 1980). Mahalanobis distances (squared distances between group centroids) are calculated and pair- wise and pooled multivariate analysis of variance are performed; if the group centroids are close (similar composition of different biofacies) or the within-group variation approaches the among- group variance (highly variable within-biofacies

composition), the groups will not be recognized as significantly different. Univariate analysis of variance was also performed to assess the contribution each variable made to differentiating samples into biofacies. CVA was executed using the CANDISC procedure of SAS (1989), version 6.

An important additional question is: how trophically distinctive are the taxonomically-defined biofacies? Although a detailed analysis of the feeding paleoecology of these crinoids is beyond the goal of this paper, an analysis of guild distribution patterns can distinguish structurally different biofacies from those produced merely by minor taxonomic reordering (Table 1; Fig. 4; see Holterhoff, submitted). Here I used CVA to test if the taxonomically-defined biofacies are distinguishable based upon their guild composition.

3. Results

3.1. Plattsburg Cyclothem

Q-mode analysis of the ten samples from the Plattsburg Cyclothem identified three clusters (Fig. 5). Q-cluster 1 is composed of samples from nearshore deposits (Appendix 1). Sampled lithologies include even to wavy bedded wackestones/packstones, calcareous shale with thin packstones/grainstones, and muddy shales. All of these samples are from more shoreward horizons than the other Plattsburg samples, based upon their paleogeographic and stratigraphic position.

Q-cluster 1 is dominated by taxa from R-clusters 1, 2, and 3, with very little contribution from taxa of cluster 4. Delocrinus is particularly abundant along with members of the cymbiocrinids, ampelocrinids, cromyocrinids, and pirasocrinids. Erisocrinus, along with Stellarocrinus and stenopecrinids also make important contributions to Q-cluster 1.

Q-cluster 2 includes samples from deposits composed of calcareous shales and argillaceous, nodu- lar, and/or thin bedded packstones and grainstones, often with gutter casts. This Q-cluster is dominated by Erisocrinus of R-cluster 3 and Apographiocrinus of R-cluster 4. Delocrinus, of R-cluster 2, and Cibolocrinus, of R-cluster 4, also make important contributions to Q-cluster 2.

P.E Holterhoff/Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996) 47--81

Table 1 Morphological criteria for Late Paleozoic crinoid guilds

57

Pinnules No. of a r m s Brachials Body size Fan type Ecology

Guild 1 yes > 10 biserial large very dense high E, nearshore Guild 2 yes 5-10 biserial large dense Guild 3 yes > 10 uniserial l a r g e dense-intermediate Guild 4 yes 5-10 uniserial l a r g e intermediate-open Guild 5 yes > 10 uniseral s m a l l intermediate-open Guild 6 yes 5-10 uniserial small open Guild 7 no variable uniserial variable very open Parag. (8) stalkless low E, offshore

Q-cluster 3 is composed of one sample from a late transgressive deposit in the basin and one from the maximum flooding horizon at the distal shelf edge. Lithologies are dominated by calcareous shales with minor, thin argillaceous packstones. This Q-cluster is dominated by Apographiocrinus, Contocrinus, and Cibolocrinus of R-cluster 4.

Whereas the Q-clusters are discrete, in that within cluster associations are generally at a higher level of similarity compared to between cluster linkages, it is clear from the abundance matrix that most of the taxa are wide ranging. It is the pattern of abundance distributions among the taxa that is controlling the structure of the clusters. Examination of this abundance matrix clearly shows the gradation of one Q-cluster into the next, with most of the abundant taxa displaying some form of unimodal, possibly Gaussian distribution of abundances across environments.

3.2. Stanton Cyclothem

A detailed analysis of the paleoecology of crinoids from the Stanton Cyclothem has been presented elsewhere (Holterhoff, submitted). However, a review of the cluster analysis is presented here to facilitate discussion. The 20 Stanton samples can be grouped into five Q-mode clusters and the 27 taxa grouped into five R-mode clusters (Fig. 6). While the R-mode clusters are easily recognized, the Q-mode clusters are more difficult to differentiate, particularly clusters 1 and 2. The rationale for dividing these localities into their respective clusters will be discussed later; however,

this does point out the previously discussed problem of forced grouping in cluster analysis.

Q-cluster 1 is composed of samples from generally nearshore deposits. Associated lithologies include wavy bedded packstones/wackestones, burrow mottled argillaceous packstones, and silty to muddy shales. Like Q-cluster 1 of the Plattsburg Cyclothem, these localities represent the most shoreward deposits of the Stanton samples.

Q-cluster 1 is generally characterized by taxa from R-cluster 1, specifically by the abundance of ampelocrinids, Stellarocrinus, pirasocrinids, Delocrinus, and Graffhamicrinus, with negligible contributions from taxa of the other R-mode clusters, except Erisocrinus of R-cluster 2.

Q-cluster 2 contains samples from many different settings, although all of them could be considered "moderate depth" settings. Associated lithologies include muddy shale, calcareous shale with thin bedded packstones with gutter casts, and "whole fossil" packstones and grainstones. Q-cluster 2 includes many taxa of R-clusters 1-3, but is most abundant in Erisocrinus and Apographiocrinus, as well as Delocrinus, pirasocrinids, and stenopecrinids.

Q-cluster 3 samples are all from offshore, aerobic deposits which include argillaceous wackestones, massive sponge-rich wackestones, argillaceous packstones, and calcareous shale with minor, thin packstones. Q-cluster 3 is clearly defined by the abundance of taxa from R-cluster 3, specifically Apographiocrinus, Cibolocrinus, and Kallimorphocrinus, as well as Erisocrinus of R-cluster 2 and Delocrinus of R-cluster 1.

Samples from Q-cluster 4 are from offshore


Guilds 2 3 4 5 6 7 8

Dlchocrlnus • CAMERATES Dlnoa'ocdnua •

cromyocdnlde • 8teIlarocrlnus •

Delocdnue • Erlsocdnus • Graflhamlcrlnus • Ulocdnua •

DISPARIDS

FLEXlBLES

Exodocrir.m • Glaukoeocdnus • laudonocdnids • plrasocflnlds • Plummedcrlnus • scytalocdnlde • ='enopecdnide • Terpnocrlnus •

ampelocflnlds • cymbiocdnlds • Blbntocdnua • Paragasslzo~nus *

tu Exocdnu8 • Z G=latescdnus • 0 0 Apographlocrtnus • l gmphlocrlnlde •

I.ecythlocrlrttm •

Kalllmorphocdnw •

Clbolocdnus • Euonychocdnus • Paramphlcdnue •

Fig. 4. Guild membership of crinoid taxa examined for this study. Generic identifications conform to Moore et al. (1978); generic complexes (lower case) are discussed below. Although camerates were encountered, they are extremely rare and were not included in the analyses. Cromyocrinids include those Missourian taxa identified by Webster (1981) as belonging to the Cromyocrinidae, except Ulocrinus. Laudonocrinids include Laudonocrinus, Bathronocrinus, Athlocrinus, and Schistocrinus. Pirasocrinids include those Missourian taxa identified as members of the Pirasocrinidae by Moore et al. (1978) except Stenopecrinus, Perimestocrinus, and Triceracrinus, which form the stenopecrinid complex. Scytalocrinids include Haeratocrinus and Hydriocrinus. Ampelocrinids include most species of Aesiocrinus and Polusocrinus. Cymbiocrinids include Aesiocrinus detrusus, Allosocrinus, Moundocrinus, and Oklahomacrinus. Graphiocrinids include Contocrinus and Euerisocrinus.

deposits characterized by shales with abundant pyrite or siderite/limonite concretions. These samples are vertically transitional with dysaerobic molluscan faunas of the maximum flooding horizon of the Stanton Cyclothem (Boardman et al., 1984). Q-cluster 4 is easily differentiated as a distinct group based upon the abundance of Exocrinus, the sole component of R-cluster 4,

along with the abundance of Apographiocrinus of R-cluster 3.

Samples from Q-cluster 5 are associated with the dysaerobic deposits of the "core shale" of the Stanton Cyclothem. This clayey shale contains abundant limonite concretions and is dominated by a mature molluscan fauna of Boardman et al. (1984). Q-cluster 5 is basically a monospecific


0 . 0 -

A : >43.0% - 54.0%

• : >27.0% - 43.0%

• : >16 .0%- 27.0% 0.5

0 : >5.0% - 16.0%

- - : > 0 . 0 % - 5 . 0 %

1.0

0.3 0.5 1.0

I , I Q " m^~^ CCQN

CBMO

TMBO . . . . .

O75K

K96K

K47K N75K

BRQK

NFDK . . . . .

HPTO

R - mode

I 3

%'..% ~.'~ %%%%%

• .,.,~ %%

, , .%

%,, , *,.%

• - - - - . ~ . " . 3 o . . r - ,

4

- • 0 0 0 - 0 O 0 0 0 0 0 0

0 - 0 0 - 0

- 0 i i 0 i i i i i

I 1 0

0

I i B

0 0 0 0 0 0 0

0 0 0 -

0

m m o r n

. . . . 0 0 - - - 0 0 -

. . . . . o O ' - - e o , _ _ - O - 0 0 - - 0 @ 0

- 0 . . . . 0 0 0 - -

Fig. 5. Two-way cluster diagram of localities and crinoid taxa from the Plattsburg Formation. Dendrograms link individual objects and clusters at differing levels as indicated by the similarity scales adjacent to the dendrograms. The relative abundances of the different taxa within each sample are indicated; note shifting abundance peaks from upper left to lower right. This analysis identifies four taxonomic associations (R-mode clusters 1-4, left to right) and three sample associations (Q-mode clusters 1-3, top to bottom).

assemblage dominated by Paragassizocrinus tarri of R-cluster 5.

Like the Plattsburg Cyclothem, the Stanton Cyclothem abundance matrix shows that most of the taxa are wide ranging, with the pattern of abundance distributions among the taxa controlling the structure of the clusters, again with most of the abundant taxa displaying some form of unimodal distribution. Whereas Q-cluster 5 is unique, the other clusters grade from one into the next, such that some of the marginal samples might be thought of simply as samples that do not easily group with members of the other Q-clusters; this seems especially true for clusters 1 and 2.

3.3. South Bend Cyclothem

The South Bend Cyclothem is a thin cycle compared to the thicker Stanton and Plattsburg cyclothems; thus it is not as well exposed and was

not as extensively sampled as the underlying units. However, the four samples here (Fig. 7) represent a wide range of facies and provide important comparisons to the Stanton and Plattsburg cycles.

Sample LXQN is from the middle shale horizon of the Kitaki Limestone of Pabian and Strimple (1993). This unit is a calcareous shale containing fusulinids, bound above and below by burrow mottled, osagid-grain and fusulinid-rich packstones/grainstones. The crinoid fauna is dominated by Delocrinus and cromyocrinids, along with pirasocrinids, scytalocrinids, and Erisocrinus.

Sample ERQK is from argillaceous wackestones of the upper Little Kaw Limestone at the South Bend shelf edge (Heckel, 1975b; Pabian and Strimple, 1993). The crinoid fauna from this unit is dominated by Erisoerinus and Apographiocrinus, with several other taxa following in abundance.

Sample ARCK is from sponge-rich calcareous shale and thin bedded packstones of the upper

60 P.F.. Holterhoff/Palaeogeography, Palaeoclirnatology, Palaeoecology 127 (1996) 47-81

O : >60.0% 0 .0

• : >40.0% - 60.0%

• : > 25.0% - 40.0%

• : >16.0"/, - 25.0% 0 .5

O : >8.0% - 16.0%

- - : >0.5% - 8.0%

1.0

0.0 0.5 1.0

I I I SLWO

1 ~ BCQO Q - m o d e

UTDO .__j ...- .<,.- ...- .+,.-,

-

i

2 ~ NTSO

~ WLQK

LCWO

r , TOQK

- - ~ PHFK

..1:;, ,tt, ,5.K

• ? ; - ; - ; ' u K c K

t O~ -0 ~ m ' o w o

• ' ~ .x- . " w m .-1 .--

o '<+ O@ O O 0 0 - - - 0

- • - 0 - i O m i mm

i

i

i

m

O

0 - O

O O O O - - "

I

i

I

m

I

I I

. . = . o - o .

@

O

O O

O

O

, , UCWO, . . . . . ,

4 I~, LTDO

IL' ETQK

5 i RFPK I

I PGSK

i

i

i

i

0 0 m

- 0 0 0 0 •

0 0

| i i i ~ ,

- 0 i i

0 0 i

i i i

i i

I

o

m m

i

+15 i

m

- O i

0 - 0

0 - 0

O 0 e l l

- 0 0 - O 0

- 0 • - • 0 - 0

- - I Fig. 6. Two-way cluster diagram of localities and crinoid taxa from the Stanton Formation. Dendrograms link individual objects and clusters at differing levels as indicated by the similarity scales adjacent to the dendrograms. The relative abundances of the different taxa within each sample are indicated; note shifting abundance peaks from upper left to lower right. This analysis identifies five taxonomic associations (R-mode clusters 1-5, left to right) and five sample associations (Q-mode clusters 1-5, top to bottom). Q-cluster 2 assignment of PRQN and TMQK as diagnosed from forthcoming MDS analysis of pooled Lansing Group data.

Little Kaw Limestone. This unit lies on a rise generated by the Chautauqua Arch and accentu- ated by the underlying Onion Creek Sandstone (Heckel, 1975a; Bennison, written comm.). This fauna is dominated by Apographiocrinus, with sub- ordinate Stellarocrinus.

Sample RRWK is from massive, sponge-rich wackestone of the upper Little Kaw Limestone. This locality lies between the South Bend shelf edge to the north and the Chautauqua Arch to the south. This fauna is dominated by

Apographiocrinus and Er&ocrinus with important contributions from Cibolocrinus and pirasocrinids.

The small number of samples from the South Bend Cyclothem make it difficult to describe the overall structure of the data matrix. However, it does suggest that taxa are wide ranging, with the pattern of abundance distributions among the taxa controlling the structure of the clusters much like the underlying Stanton and Plattsburg cycles.

Individual cluster analyses of the three cyclo-


• :>35.0% * 50.0%

• :>25.0% - 35.0%

• : >15.0% - 25.0*/.

O : >5.0% - 15.0%

- - : > 0 % - 5.0%

0 . 0 ~

R- mode

0.5

1.0

0.0 0.5 1.0 "- I I I - . ~ . - - 0 ' " = " " " ~-*=

LXQN • • O O o - I I 0 - - 0 0 0 - 0 - - - - • ERQK O

I '1 ARCK O O O O O • O O • O Q-mode RRWK 0 • ------ • -- -- O

Fig. 7. Two-way cluster diagram of localities and crinoid taxa from the South Bend Formation. Dendrograms link individual objects and clusters at differing levels as indicated by the similarity scales adjacent to the dendrograms. The relative abundances of the different taxa within each sample are indicated; note shifting abundance peaks from upper left to lower right.

thems of the Lansing Group indicate that: (1) groups of samples, or local paleocommunities, are generally united at moderate to high levels of similarity, defining biofacies within each cyclothem; (2) these biofacies are consistently associated with depth-controlled paleoenvironments; (3) although cluster diagnoses between neighboring biofacies can be problematic, most are discrete, occurring at low levels of similarity; (4) the lack of cluster "chaining" and numerous unique outliers indicates that the data sets are highly struc- tured, as opposed to some random set of taxonomic abundances; (5) this structure is imparted not by turnover of taxa between the biofacies, but rather by the waxing and waning of the abundances of key taxa; thus the biofacies are actually gradational.

3.4. Lansing Group

To assess the relationships and potential persistence of these biofacies among the three cyclothems, the entire data set was pooled and analyzed. The fundamental question this pooled analysis asks is: do the biofacies identified within each cyclothem remain grouped together, indicating unique nearshore-offshore biofacies within each cyclothem and lack of persistence, or do biofacies from similar paleoenvironments among all cycles

group together, indicating persistence and recurrence of biofacies? Cluster analysis of the 34 samples and 27 taxa of the Lansing Group shows that they can be divided into five Q-mode clusters and five R-mode clusters (Fig. 8). Q-cluster 1 is composed of members of Q-clusters 1 from the Plattsburg and Stanton cyclothems, plus SBLXQ from the South Bend cyclothem. Again, all of these samples are from shoreward deposits. This Q-mode defined biofacies is dominated by members of R-clusters 1 and 2, with Delocrinus, pirasocrinids, cromyocrinids, Erisocrinus, and ampelocrinids most abundant.

Q-cluster 2 is composed of members of Q-clusters 2 from the Plattsburg and Stanton cyclothems, plus SBERQ and SBARC from the South Bend cyclothem. These samples all represent "mid- depth" paleoenvironments in a variety of paleogeographic and depositional settings. This biofacies is characterized by a high diversity of taxa from R-clusters 1-4, although it is dominated by Erisocrinus, Apographiocrinus, and Delocrinus.

Q-cluster 3 is composed of members of Q-clusters 3 from the Plattsburg and Stanton cyclothems, plus SBRRW from the South Bend cyclothem. These samples are all from offshore, aerobic horizons. This biofacies is dominated by taxa of R-cluster 3, especially Apographiocrinus and Cibolocrinus, with important contributions from Erisocrinus of R-cluster 2.


0 . 0 m

• : >80.0%

• : >50.0% - 80.0O/.

• : >30.0°/* " 50.0`=/0 0.5

0 : >20.0% - 30.0%

O : >10.0% " 20.0°/0

: > 0 % - 10 .0% 1 . 0 -

.0 0.5 1.0

I , I ~ PLCCQ

Q - mode STSLW STBCQ

1 1,1,1, ~ . . . . . ' ......... , PLCBM ~ S ~ I N ,,*%,,- ,~

,,, ,, / ,, PLTMB . . . . . $TUTD

,,',,",," " STLKC ',"~," " " SBLXQ

PL075 $TPRQ S'I-I'MQ

~ $BERQ $'I'WLQ PLKg6

., " ~ STLGW " PLK47

~ PLN75 ,.,,,..,, ~ ,, PLBRQ

SBARC $TNTS

STBNF ST75B SBRRW STTOQ STPHF STUKC STUGW PLHPT

4 r~ $TLTD STETQ

5 j STRFP $TPGS

R - mode j |

[ . ' - - . . . . . . • . . . -

: O 0 - 0 0 0

- o 0 0 o - - 0 0 - 0 0 -

- - 0 . . . . . . 0 0 0 0 - - - - 0

0 - - - 0 0 -

. . . . 0 0 - - - - - 0 0 0 - -

9 - • I l l I ~ l l l I I I ~ I i l I I O I l m l

- - 0 - 0 0 - - - -

0 - - 0 - - O 0 . . . . @ - - - -

0 0 - -

" 0 Q - -

- - 0 - - 0 - - - - O - - C

m

m

- - 0 - -

- - 0 - -

m

m

0 m J

0

~ 0

m

- - 0 - - - - - - 0 - - - - 0

0 0 - - • - - O - - l )

0 0 - - • . . . . . . . 0 . . . . IP 0 . . . . . - - - - - O - - O - O - - 0 - - 0 - - 0 • 0 0 - - - 0 - - 0 - - - -

. . - - - - Q - - Q - - -

O - - - • - - - - 0 - - - - 0 - - 0 - - 0

- - O - - O - - - - O - -

. . . . . . - - I - = - I - - ' _ - - -

Fig. 8. Two-way cluster diagram of localities and crinoid taxa from the Lansing Group. Dendrograms link individual objects and clusters at differing levels as indicated by the similarity scales adjacent to the dendrograms. The relative abundances of the different taxa within each sample are indicated; note shifting abundance peaks from upper left to lower right. This analysis identifies five taxonomic associations (R-mode clusters 1-5, left to right) and five sample associations (Q-mode clusters 1-5, top to bottom). Samples from the three formations are interspersed among Q-clusters 1-3, indicating that these faunal associations recur from cycle to cycle. STPRQ and STTMQ assignments to Q-cluster 2 diagnosed from MDS analysis.

Q-clusters 4 and 5 are the same as Q-clusters 4 and 5 f rom the S tan ton Cyclothem. These biofacies represent offshore, dysaerob ic env i ronments developed near m a x i m u m f looding o f the S tan ton cycle. The S tan ton cyc lo them was the last m a j o r f looding

event o f the Missour i an ( B o a r d m a n and Heckel , 1989) and has wel l -developed dysae rob ic / anae ro - bic, s tarved deposi ts . The under ly ing P la t t sburg and over lying South Bend cycles are compara t ive ly m i n o r cycles and do no t have well deve loped

P.F. Holterhoff/Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996) 4~81 63

anaerobic horizons at outcrop. Thus, these offshore biofacies are well developed in the Stanton cyclothem whereas they are rare and feebly developed in the other cycles.

Like the individual cycles, the pooled Lansing Group abundance matrix re-emphasizes that most of the taxa are wide ranging, with the pattern of abundance distributions among the taxa controlling the structure of the biofacies, again with most of the abundant taxa displaying some form of unimodal distribution. The large sample size hopefully captures the natural range of within- facies variation such that the biofacies "boundaries" are accurately portrayed; however, the transition between biofacies 1 and 2 is still problematic and the other biofacies boundaries may be artifi- cially sharp due to the clustering method itself.

To help clarify these problems, Q-mode MDS ordinations were performed (Fig. 9). The final stress of this plot was 0.125, which is considered a "good" fit to the original similarity matrix (Rohlf, 1989). Cross cutting linkages, or "folding over" of the MST branches, indicate where distortion in collapsing the multidimensional similarity matrix onto the two MDS axes occurred. Placement of the two questionable samples from the cluster analysis, STTMQ and STPRQ, is resolved by these plots. The sole linkage of STTMQ is with STWLQ, a member of Q-cluster 2; the unique abundance of Ulocrinus explains its difficult classification. STPRQ clearly plots within the field of other members of Q-cluster 2 and has linkages to members of Q-clusters 1 and 3. This transitional linkage reflects the high diversity of this sample and the abundance of Delocrinus, Erisocrinus, and Apographiocrinus - - major components of Q-clusters 1, 2, and 3, respectively. Thus, it was included within Q-cluster 2 of Figs. 5 and 7.

While the MDS/MST plot emphasizes the gradational relationship among the samples, the MDS/Q-cluster plot clearly shows the grouping of samples into discrete units with no intercalation of samples among the groups. This is an indication that the clusters are discrete, although the proxim- ity of clusters 1-4 suggests that they are closely related, as opposed to cluster 5, which is isolated and unique. The ordering of the clusters along the

MST suggests that the samples are arrayed along a depth gradient, from shoreward horizons of cluster 1 to offshore, dysaerobic horizons of clusters 4 and 5.

A CVA was performed to more rigorously test these biofacies (Fig. 10). To limit the distorting effects of outliers, Biofacies 5 and Para- gassizocrinus tarri were not included in this analysis because of their obvious uniqueness. In the first plot of the first two canonical variables (Fig. 10A), the samples, as indicated by their biofacies designations, fall into distinct regions on the plot. The individual groups are composed entirely of samples from a single biofacies, demonstrating that the taxonomic structure of these local paleocommunities and biofacies is distinct. Tables 2 and 3 show that while most of the pair-wise comparisons of these groups are not convincingly significant, the pooled comparisons are all significant. This indicates that the location of biofacies centroids, or means, are not generally significantly different in taxonomic, or variable space as delineated by the composition of only two biofacies, except when the comparisons are between the extreme biofacies. However, in the pooled analysis the multidimensional space, as defined by the taxa among all of the biofacies, is much greater and the centroids occupy significantly different locations within this large variable space.

The second plot (Fig. 10B) attempts to show graphically the contribution of each original variable (as the level of correlation of each original variable to the canonical variables) to the total canonical structure of the two canonical variables; these are somewhat like variable scores in principle components analysis. Delocrinus, Glaukosocrinus, ampelocrinids, cromyocrinids, and pirasocrinids are strongly positively correlated to both canonical variates (CV) 1 and 2; Erisocrinus is strongly negatively correlated to CV 2 and positively correlated to CV 1; Apographiocrinus is strongly negatively correlated to CV 1 and 2; and Exocrinus is extremely negatively correlated to CV 1. Samples with similar CV loadings are dominantly controlled by these respective taxa. Review of the univariate tests also indicate that these taxa, along with Ciboloerinus, are the only taxa in the data set which are significantly non-randomly distributed

64 P.F. Holterhoff/Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996} 47-81

D i m e n s i o n 2 1.5

1

0 .5

0

(0.5)

(1)

(1.5)

(2)

STTMQ

" ~ !..~,~4~ ~ P ~ 7 S ~ ' _ . - - ST'/SB

PLCBM PLTMB PLO75 ~ ~ ( ~ W ~PLNFD STIJCW

. . . . ~ r p . ~ , / ~ ~ ~ ~ . . ~ . ~ . H.F . . . . . .

STSLW7 **'~:kc _~__ !s.,~,c ==--" \ e / ~IUTU : "~STLTD

PLCCQ i ~ S T E T Q

STPGS

5 ) ( , i , i , ,

(2. 2) (1.5) (1) (0.5) 0 0.5 1 1.5 M D S s t r e s s = 0 .125 D i m e n s i o n 1

A

D i m e n s i o n 2

1.5 . Cluster 2 1 [.SrrMQ : 1

I . . : / Cluster 3 0.5 Cluster I . I P'I'~-~..i-PLN75pLBRQ/ eST75e . . . . . . .

• PLTMBpLO7S. J I e~4, : ~ I .PLNFD a ' ~ ' " , ,c . , .,~w,, I I "w'° :....o % 1 ~ b ; ~ % , . ,

0 --. ;~:Z~..s~.K~ . . . . . . . . 1"'" I""~YI~R~'" e r . p . . 1~,~ ~Y0Xd" , .... I o SBLXQ / / seZ=r. I "o~ PLHPT I STSLW STUTD ° I / - - - I T •

(05) I I (1) STi~TQ

Cluster 4

(1.5) Cluster 5

(2) "

5) (2 ) ' , , , , , (2. (1,5) (1) (0.5) 0 0,5 1 1.5 M D S s t r e s s = 0 .125 D i m e n s i o n 1

B

Fig. 9. Q-mode Multidimensional Scaling (MDS) ordination of the Lansing Group samples. A. MDS with samples linked by a minimum spanning tree (MST). Note the scatter of samples about the ordination, indicating a fair amount of gradation between samples. B. MDS overlain by Q-clusters. Although the spread of samples suggests significant overlap in cluster identity, each cluster is generally confined to a particular "field" within the ordination, indicating that the clusters are indeed readily discernable.

P.F. Holterhoff/Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996) 47 81 65

CAN 2 15

10

5

0

% O

20 (10 /4

A

@

I I I I (30) (20) (10) 0 10

CAN 1

CAN 2 0.8

0 .6

0 .4

0 .2

(0.2)

(0.4)

(O.e)

(0.8) (1)

B

Sxocr.

Apogr.

Eu~ny. Galat. g~ph. :

Clboh • • • Kalll. ~ .

Param.

ampel.

• • Glauk. Deloc.

eecymbh Plumm.

~St Graft" Exorl. ell,

• ~. Terpn.

EIIba. Ulocr.

st•no.

laudo,

Erlso.

0.6 I I I I I I

(0.8) (0.S) (0.4) (0.2) 0 0 .2 0 .4

CAN 1

or;my

piras.

Fig. 10. Canonical variate analysis (CVA) of samples grouped by taxonomically-defined biofacies. A. Plot of the samples on the first two canonical variables; over 85% of the cumulative variance in the data are explained by these two axes. The biofacies are well defined and clearly separated. B. Plot of the level of correlation of each original variable to the total canonical structure of the two canonical variables. This provides a sense of how the taxa are structuring the CVA plot. See text for discussion.

66 P. F Holterhoff/Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996) 47-81

Table 2 Pair-wise Mahalanobis distances with (probabilities of >distance) between biofacies

Biofacies 1 Biofacies 2 Biofacies 3 Biofacies 4

Biofacies 1 0 ( 1.000) Biofacies 2 194.303 (0.1313) 0 (1.000) Biofacies 3 378.941 (0.0618) 305.119 (0.0782) 0 (1.000) Biofacies 4 2092 (0.0245) a 1901 (0.0273) a 1187 (0.0555) 0 (1.000)

ap < 0.05; significant difference between biofacies pairs.

Table 3 MANOVA tests of equality of biofacies means

F Pr>1 ~

Wilks' lambda 6.9302 0.0012 Pillai's trace 8.5106 0.0001 Hotelling-Lawley trace 4.5433 0.0471 Roy's greatest root 28.2742 0.0007

ap <0.05; biofacies means are significantly different within the Lansing Group.

among the biofacies (Table 4). One other taxon is marginally significant (Euonychocrinus) while the remaining 16 are essentially randomly distributed. That the majori ty o f the taxa are not significantly distributed with regard to the biofacies m a y account for the non-significant pair-wise comparisons among biofacies ment ioned earlier.

I t is clear f rom the previous analyses that the 34 samples, or local paleocommunit ies , f rom the Lansing G r o u p can be grouped into five discrete biofacies based upon the similarity o f their taxonomic-abundance composit ion. Nearshore biofacies 1 is characterized by Delocrinus, Glauko- socrinus, ampelocrinids, cromyocrinids, and pirasocrinids. Midshelf biofacies 2 is dominated by Erisocrinus and Apographiocrinus, along with a variety o f other taxa. Offshore biofacies 3 is dominated by Apographiocrinus and Cibolocrinus. Dysaerobic biofacies 4 is dominated by Exocrinus and Apographiocrinus and biofacies 5 is monospecific, dominated by Paragassizocrinus tarri.

To assess the congruence o f taxonomic and morphological distribution patterns, as well as to assess how different in ecological structure the biofacies are, the individual and pooled sample guild abundances were plotted on the Q-mode clusters (Fig. 11). These plots show that a l though

Table 4 Multiple comparison ANOVA tests for random distribution of taxa among biofacies 1-4

F Pr>F

Exocrinus 133.2138 0.0001 b Apographiocrinus 23.8191 0.0001 b Cibolocrinus 17.4551 0.0001 b cromyocrinids 12.6781 0.000 lb Erisocrinus 9.6864 0.0001 b pirasocrinids 9.5411 0.0002 b Delocrinus 6.7675 0.0014 b ampelocrinids 6.4400 0.0019 b Glaukosocrinus 5.1290 0.0059 a Euonychocrinus 3.6578 0.0242 a stenopecrinids 2.3428 0.0946 Graffhamicrinus 2.0314 0.1322 Stellarocrinus 2.0287 0.1326 Plummericrinus 1.8807 0.1558 Laudonocrinus 1.8534 0.1605 Exoriocrinus 1.6371 0.2032 Lecythiocrinus 1.5378 0.2266 Galateacrinus 1.1766 0.3364 Scytalocrinus 1.1237 0.3563 Elibatocrinus 1.0825 0.3726 graphiocrinids 0.8909 0.4580 Ulocrinus 0.7456 0.5340 Terpnocrinus 0.2778 0.8409

ap<0.05; taxa are non-randomly distributed among biofacies 1-4. bp < 0.002 (Bonferroni corrected c~ = 0.05).

there is some variat ion a m o n g samples within biofacies, each biofacies does have a characteristic guild structure. The paleoecological significance o f this structure is not the focus o f this paper (see Holterhoff, submitted); however, guild structure does provide an important , trophically-based reference for describing these biofacies. Biofacies 1 is dominated by guilds 1 through 4, with negligible contr ibutions f rom the other guilds. Biofacies 2 is

P.F. Hoherhoff/Palaeogeography, Palaeoclimatology, Palaeoe¢ology 127 (1996) 4~81 67

PLCCQ STSLW STBCQ PLCBM STWlN PLTMB STUTD STLKC SBLXQ PLO75

STPRQ STTMQ SBERQ STWLQ

PLK96 STLCW

PLK47 PLN75

PLBRQ SBARC STNTS PLNFD STBNF ST75B

SBRRW SI-rOQ STPHF STUKC STUCW PLHPT STLTD STETQ STRFP STPGS

A

Biofacies 1

0 10 20 30 40 50 60 70 80 90

Percent Guild Abundance within Samples 100

BIOFACIES

2

3

4

5

Biofacies 2

Biofacies 3

Biofacies 4

Biofacies 5

[ ] Guild 1

[ ] Guild 2

[ ] Guild 3

[ ] Guild 4

[ ] Guild 5

[ ] Guild 6

[ ] Guild 7

[ ] Parag.

0 10 20 30 40 50 60 70 80 90 100

B Percent Guild Abundance within Biofacies

Fig. 11. Guild abundances among samples and biofacies. The distribution of guilds among samples (A) and pooled within biofacies (B) clearly indicate that each biofacies has a characteristic guild composition, although the ordering of samples within biofacies is simply an artifact of cluster ordering from the CA. See Holterhoff (submitted) for a detailed discussion of these guild distribution patterns.


dominated by guilds 2 and 3, although most of the guilds are well represented. Biofacies 3 is clearly dominated by guilds 6 and 7 with an important contribution from guild 2. Biofacies 4 is dominated by guilds 5 and 6, while Biofacies 5 is basically unique, containing the sole occurrences of Paragassizocrinus tarri of Guild 8.

The CVA more rigorously confirms these observations (Fig. 12A). Like the taxonomic analysis, Biofacies 5 and Paragassizocrinus tarri were not included in this analysis. In this plot of the first two canonical variables, the samples, as indicated by their biofacies designations, generally fall into distinct regions on the plot. Most of these groups are composed entirely of samples from a single biofacies, demonstrating that the guild structure of these paleocommunities and biofacies are as distinct as their taxonomic structure. However, two samples from Biofacies 2, STTMQ and PLK96, plot with members of Biofacies 1. Like many of the paleocommunities of Biofacies 1, these samples are dominated by Guild 2. However, while this guild is represented by Delocrinus in many of the Biofacies 1 faunas, these Biofacies 2 samples are dominated by Ulocrinus and Erisocrinus. This suggests that although the taxonomic composition of these two paleocommunities is similar to members of Biofacies 2, their trophic structure is more similar to those of Biofacies 1, indicating that there was differential habitat selection among taxa within guilds. Pair-wise and pooled tests are all highly significant (p<< 0.05; not presented for brevity), even with the misclassification of two samples, indicating the distinctness of guild composition among the biofacies.

A plot of the degree of correlation of each original variable to the canonical variables on the canonical variates (Fig. 12B) shows that guilds 1-4 are strongly positively correlated to canonical variable 1, whereas Guilds 5, 6 and 7 are strongly negatively correlated to this variable. Guild 5 is strongly correlated to the second variable while Guilds 6 and 7 are strongly negatively correlated to this variable; guilds 1-4 show only a slight correlation to this axis. These two plots together support the observations made in the guild abundance plots: Biofacies 1 is dominated by Guilds 1-4; Biofacies 2 is composed of a combination of

Guilds 1-4 and 6 and 7; Biofacies 3 is dominated by Guilds 6 and 7; and Biofacies 4 is dominated by Guilds 5 and 6. Univariate analysis of these guilds confirms that all are significantly non- randomly distributed among the biofacies at the p <<0.05 level (not presented for brevity).

3.5. Summary of results

The multivariate statistical analyses performed on this data set reveal that the 34 crinoid samples from the three cycles of the Lansing Group can be divided into five recurrent crinoid biofacies, each with characteristic taxonomic-abundance and guild structure. The only "ecostratigraphic" signa- ture (distinct biofacies associated with a single cycle) are the two extreme offshore, dysaerobic biofacies of the Stanton cyclothem. Examination of crinoids from the Iola cyclothem (unpublished data), which is the next underlying major cyclothem comparable to the Stanton, reveals that both of these biofacies are present. However, the taxa that are abundant and characteristic of these biofacies (Exocrinus and P. tarri) do have rare, isolated occurrences in the offshore horizons of the other cyclothems (unpublished data), intimating their continued presence within the region. Thus, these dysaerobic biofacies are simply assemblages from paleoenvironments that were well developed at the outcrop belt only during the most extensive flooding events.

These biofacies are not randomly distributed among the cyclothem facies, but rather are controlled by position within transgressive or regressive hemicycles, regional paleogeographic position, and local topography. These sedimentological factors would have controlled the local paleohy- drographic regime, which is considered paramount in crinoid ecology (Meyer, 1973, 1979, 1982; Kammer, 1985; Kammer and Ausich, 1987; Baumiller, 1990, 1993; Holterhoff, submitted).

To compare the paleoecological models set forth earlier, biofacies patterns within and among the cycles need to be assessed. To facilitate this, a composite stratigraphic column of the samples was constructed and the distribution of the biofacies plotted (Fig. 13). Some cyclothems display biofacies omissions within certain portions of trans-


CAN 2 12

10

8

6

4

2

0

(2)

(4)

(e) (15)

A

O

O

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I I I (10) (5) 0

CAN 1

CAN 2 0.8

0.6

0.4

0.2

0

(0.2)

(0.4)

(o.s)

(0.8)

(1) (1)

B

Guild 5

Guild 1 Guild 4 • •

• Guild 3

Guild 2

Guild 6

Guild 7

I I (0.5) 0 0.5

C A N 1

Fig. 12. Canonical variate analysis (CVA) of samples grouped by taxonomically-defined biofacies using guilds as the test criterion. A. Plot of the samples on the first two canonical variables; over 97% of the cumulative variance in the data are explained by these two axes. See text for discussion. B. Plot of the level of correlation of each original variable to the total canonical structure of the two canonical variables. This provides a sense of how the guilds are structuring the CVA plot.

70 P.F. Holterhoff/Palaeogeography, Palaeoclirnatology, Palaeoecology 12 7 (1996) 4 ~81

South

Bend Fm.

LL t - O t -

O3

LL

.g) t,D

f l -

BIOFACIES I .]-'~' RHC ,-~- T ~ C MFS

=2 CB I I

121RHC

PGSK - ETQK -

UCWO- UKCK- BOLT- TOOK - LCWO- TMQK- LKCK -

BCQO - SLWO - ~ I ~ M t") -

5_1-- MFs

3[THC

21

L___

RHC

MFS

THC

0 10 20 30 40 50 60 70 80 90100

SAMPLE PERCENT ABUNDANCE OF

GUILDS 6 & 7

Fig. 13. Guild and biofacies distribution within a composite stratigraphic section of the Lansing Group. The samples have been combined into a single stratigraphic section to facilitate paleoecological interpretations; Stanton Formation samples PHFK, 75BK, and BNFK have been combined into the BOLT sample here. The samples have been ordered according to their position within hemicycles (transgressive = THC; regressive = RHC), position relative to significant cyclic surfaces (cycle boundaries = CB; maximum flooding surface= MFS), and depth relative to paleogeographic position and local topography. For example, two samples from the same relative position within a cycle but from different settings (e.g. one shelf and the other basin) would be ordered such that the deeper of the two would be placed closer to the MFS. Note that the biofacies, particularly within the Stanton Formation, are fundamentally symmetrically positioned around the MFH with the exception of Biofacies 3, as highlighted by the > 50% abundance of guilds 6 and 7.

gressive or regressive hemicycles, p roduc ing asymmetr ica l biofacies sequences. However , these omiss ions are no t consis tent a m o n g all o f the cycles, indica t ing incomple te sampl ing ra ther than some pa leoecologica l phenomenon . E x a m i n a t i o n o f add i t iona l col lect ions and s t ra t ig raphic sections

no t inc luded in the quant i ta t ive analyses s t rongly suggests tha t these missing biofacies are present a long the ou tc rop belt. Fo r example , while no local pa leocommuni t i e s o f biofacies 1 are here identif ied f rom the t ransgressive hemicycle o f the South Bend cyclothem, this biofacies does a p p e a r

P.F Holterhoff/Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996) 47-81 71

to be present in nearshore, calcareous sandstones at the base of the Little Kaw Limestone Member from localities in southern Kansas.

Biofacies 3 is a noteworthy exception. As emphasized earlier, biofacies 3 represents a trophically distinctive assemblage fundamentally different in composition from other biofacies and therefore is not the product of minor sampling and counting errors. It is consistently found in the upper transgressive deposits of all three cycles and has yet to be identified from any portion of the regressive units. The potential does exist that this biofacies remains intact as local paleocommunities during regressions in isolated refuges (sensu Petuch, 1976, 1981, 1982) that have not been sampled. In this scenario biofacies would not only track favored environments, but also expand and contract in response to basin-wide environmental fluctuations (i.e. expand during transgressions, contract during regressions). Although the statistical significance of a recurrent biofacies pattern based upon only three cycles could be questioned, the quantitative data presented here, along with subsidiary analysis of numerous additional collections and stratigraphic sections within the three cycles of the Lansing Group, provide a dense enough sampling net to strongly suggest that biofacies 3 simply does not occur in these regressive facies. In most cases local paleocommunities of biofacies 3 are distinctly below maximum transgressive horizons and thus do not represent assemblages at the maximum flooding "core" of these symmetrical cycles, as do local paleocommunities of biofacies 5 of the Stanton cyclothem. Thus, the biofacies pattern within the Lansing Group is one of recurrent biofacies asymmetry, even though the depositional cycles appear to be generally symmetrical (Fig. 14).

4. Discussion

As noted in the introduction, biofacies recurrence patterns which do not conform to faunal tracking more strongly conform to models of communities of independent species under the control of regional processes than to models akin to ecological locking. The identification and recur-

rence of a trophically unique assemblage of taxa only in later transgressive deposits among the three cyclothems of the Lansing Group does not conform to faunal tracking, and therefore supports independent responses-regional processes models. Although this biofacies is ephemeral, this is not to imply that the individual taxa which comprise this biofacies are also not present in other environments/biofacies. Indeed, all of the taxa which are characteristic of biofacies 3, particularly members of guilds 6 and 7, do occur in local paleocommunities of regressive facies but in highly variable abundances (Fig. 8; Appendix 2). For example, Apographiocrinus and Ciboloerinus are the dominant taxa of local paleocommunities of biofacies 3, yet among local paleocommunities of midshelf biofacies 2, Apographiocrinus is still generally abundant and widespread while Cibolocrinus is rare and sporadically distributed. This pattern suggests that taxa are incorporated into local paleocommunities to some variable extent, but not in a coordinated fashion as might be predicted across adjacent facies transitions. Thus, taxa appear to be responding independently to paleoenvironmental factors along a depth gradient.

In contrast to biofacies 3, the other biofacies do appear to be present in both transgressive and regressive hemicycles, particularly nearshore biofacies 1 at cycle tops and bottoms and midshelf biofacies 2, generally in mid hemicycle positions (Figs. 13 and 14). The same depth zone occupied by biofacies 3 during transgressions may indeed be occupied by local paleocommunities characteristic of other biofacies during regressions, incorporating at very different abundances the taxa characteristic of biofacies 3. It appears that only during later transgression did a suite of paleoenvironmental conditions come together that did not recur during subsequent regressions. These unique conditions provided the setting for populations of the open fan crinoids of guilds 6 and 7 to expand and proliferate into a trophically unique biofacies.

In retrospect, these results may not be surprising. Transgressions are generally accompanied by many other environmental changes besides rising base- level. The formation of estuaries and shoreward

72 P.F Holterhoff/Palaeogeography, Palaeoclirnatology, Palaeoecology 127 (1996) 47-81

t,.)

'0

o

¢o

o

BIOFACIES

2 Regressive Hemicycle

4

4

iitiil ~ i ; 3

2

1

2

4

5

4 ~ ; ~ '

2

1

2

4

5

@

4

2

Transgressive Hemicycle

Regressive Hemicycle


Regressive Hemicycle


Cycle Boundary

Max. Flooding

Cycle Boundary

Max. Flooding

Cycle Boundary

Max. Flooding

Cycle Boundary

Fig. 14. Idealized sequence of crinoid biofacies distribution based upon data from the Lansing Group. Cycles 1, 2, and 3 contain biofacies 1-5. Biofacies 3 only occurs within the transgressive hemicycles, as indicated by stippling; it does not occur in an analogous position in the regressive hemicycles, as indicated by the arrows, as is predicted by faunal tracking.

entrapment of siliciclastics can produce major shifts in substrate types and significantly impact sedimentation-sensitive biota. Terrestrial nutrient increases due to flooding of continental interiors, as well as incursions of nutrient-enriched deep marine water, can significantly increase produc- tivity (Neumann and Mclntyre, 1985; Hallock and

Schlanger, 1986). Changes in the geographic extent of different habitats can increase or decrease the regional abundances of different taxa. Extensive, flooded continental interiors can potentially increase hinterland precipitation, producing changes in basinal water mass structure. Climatically-induced changes associated with a

P.E Holterhoff/Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996) 47-81 73

deepening event include varying temperature, sea- sonality, and evaporative potential, to name a few. |ndeed, systematic differences in disturbance regime (storm intensity and frequency) as controlled by temperature and atmospheric and ocean circulation patterns, coupled with changes in pro- ductivity, should have a strong influence on the biota (Huston, 1994) within a given depth zone between transgression and regression. One could envision that these effects would be most severe during later transgression, when rapid flooding would produce maximum depositional systems dis- equilibrium. Clearly, the dynamics of transgression and regression, and their influence upon marine biofacies, need to be investigated further.

In order for local paleocommunities of the various biofacies to become established with the recurrence of the proper environmental conditions, the pool of species for potential recruitment into these paleocommunities needs to be maintained (Buzas and Culver, 1994). Two related mechanisms can be envisioned to maintain species within a region or depositional basin in the face of major environmental perturbations: metapopulation dynamics and refuges.

Species which experience high rates of local extinction can maintain stability within a region by dispersal and establishment of new local population patches, effectively spreading the risk of regional extinction among numerous, stochasti- cally fluctuating populations. The sum of these highly fluctuating local populations is the metapopulation (Hanski, 1989, 1991; Hanski and Gilpin, 1991; Miller, 1993a,b). Metapopulations can not only maintain taxa within habitats and established biofacies, but can provide founders for local communities of different biofacies across many habitats (Miller, 1990). In this way, during ecological bottlenecks (i.e. lowstands), taxa within the region can be maintained in marginal habitats as low abundance and/or highly dispersed population patches until favorable conditions return, allowing the proliferation of large, numerous, and/or high abundance population patches.

Review of the Lansing Group data matrix (Fig. 8; Appendix 2) shows that /3 diversity, or turnover of taxa across facies transitions, is very low; most taxa are found in all biofacies (except

5). This indicates that even the most stenotopic species maintained populations in at least one other biofacies, in most cases many biofacies. The widespread occurrence of taxa was probably facili- tated by dispersal across habitats, linking abundant and rare local populations within and among habitats. These large metapopulations would have been extremely important in the recurrence of these biofacies among the three cycles by maintaining even stenotopic taxa in suboptimal habitats.

Refuges during environmental bottlenecks may be important in maintaining viable populations for later re-expansion. This is different from the biofacies refuges mentioned earlier in that any group of taxa can populate the refuges and abundances need not resemble established biofacies structure. Refuges which harbor taxa for long intervals of time for later expansion across the benthic "landscape" might be thought of as spatio- temporally large scale source-sink systems (Pulliam, 1988; Miller, 1993a,b). During Lansing Group lowstands, the margins of the deep Anadarko Basin (Galloway et al., 1977; Kumar and Slatt, 1984; Watney et al., 1989) would have provided refuge for offshore taxa which could later migrate across the northern shelves during flooding. In this case the basinal lowstand populations serve as the sources during flooding, with offshore taxa dispersing out across the shelves. Shelf populations of offshore taxa experience extinctions during the lowstand bottlenecks. Thus, shelf populations of offshore taxa are sink populations in that they cannot maintain themselves beyond a single depositional cycle.

! recently discussed the importance of offshore refuges for maintaining Upper Pennsylvanian and Lower Permian crinoid diversity in the North American midcontinent (Holterhoff, 1995) and these preliminary results have important implications for faunal stability. Infilling of the Anadarko Basin through this interval progressively eliminated offshore environments in the midcontinent, as well as closed deep water connections to the Texas basins to the south (Rascoe, 1962; Handford and Dutton, 1980). This infilling is closely correlated with decreases in ~/diversity, or the cumulative diversity of all facies for a single cyclothem, through the Lower Permian. This is not surprising

74 P.F.. Holterhoff/Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996) 47-81

given that offshore facies are progressively omitted from Permian cycles in the midcontinent (Elias, 1937), thus there are fewer facies to sample. However, midshelf and nearshore ~ (within biofacies) and point (single sample) diversities decrease as well. If communities are closed local systems, diversity trends should not be correlated across biofacies; however, correlated diversity trends across biofacies is more consistent with a regionally integrated system. Additionally, as noted earlier, [3 diversity is quite low for the Lansing Group fauna; a pattern which appears to be characteristic of upper Paleozoic crinoids of the midcontinent as a whole. Again, this indicates that while taxa have preferred habitats, they are still widely distributed across many environments. These marginal local populations are probably ultimately maintained by populations within the species' preferred habitats, the entire network acting as a metapopulation. The loss of offshore habitats through basin filling and the closure of immigration pathways to the deep basins in Texas eventually eliminated many taxa that were rare but significant contributors to more shoreward facies. Thus, the correlated decrease in regional (7) diversity, produced largely by offshore habitat loss, with nearshore and midshelf Qt and point diversities, strongly suggests that local paleocommunities are open and that "regional processes", such as extinctions of source populations following the loss of refuges during environmental bottlenecks and the resultant break- down in the metapopulation mosaic, ultimately controls biofacies structure. Given the potential level of openness of this system, it is intriguing that ephemeral biofacies should recur at all, con- sidering the potential alternatives, namely that different taxa within guilds adapted to particular conditions should not occasionally become abundant (Drake, 1990; Schopf, 1996). This is what invariablly leads to questions of whether local paleocommunities are "similar or the same" (Bennington and Bambach, 1996).

While I have stressed the independent aspects of taxa in their paleoecological distributions, this is not to undermine the general importance of interactions altogether. When taxa occupy the same patch, either permanently or in passing, they will often interact to some degree. Clearly in some

settings interactions are the dominant processes controlling community structure (Huston, 1994). The key issue is: to what degree do specific interactions within a local paleocommunity control long term biofacies persistence? Studies emphasiz- ing the idiosyncrasies of community history, such as the order of species assembly into the community, past disturbance events, recruitment failures, and biotic invasions, point to the potential frustration of developing deterministic interaction models for communities without knowledge of these events (Sutherland, 1974; Barkai and McQuaid, 1988; Underwood and Fairweather, 1989; Drake, 1990). Thompson (1994) discusses in detail how within - - and among - - community variation in species interactions produces a dynamic geographic mosaic of coevolution. He notes that most taxa occur across a range of environmental and community settings and that species interactions often change in intensity, polarity, and number and identity of interactors across these settings. Thus, coevolution and community organization models based on studies of interactions within local communities are missing the true dynamics of how species interact. This geographically explicit view of species interactions and coevolution is a potentially powerful alternative model for describing fauna-wide stabilization, as opposed to models involving within community ecological locking. This shift in scale of interaction models could expand ecosystem analysis much as models of local disturbance-mediated ecosystem processes describe long term, large scale stability (O'Neill et al., 1986; Smith and Urban, 1988; Morris et al., in press). Indeed, a regime of high frequency (less than average species durations), large amplitude perturbations, and concomitant migration of local populations and biofacies via metapopulation processes, may serve to increase gene flow among populations, dampening special- izations for both environmental and biotic interactions (Coope, 1979). This helps to explain observations of relative evolutionary stability in the face of cyclic perturbations (Sheldon, 1993).

It would be difficult to ascertain exactly how Upper Paleozoic crinoids, as described here, would have interacted with other organisms such that a positive feedback system would develop, buffering

P. F. Holterhoff / Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996) 47-81 75

the entire crinoid community from change (see Meyer and Ausich, 1983 for general discussion of crinoid interactions). Macurda and Meyer (1983) have noted multi-species feeding assemblages of modern comatulid crinoids and the potential for mutualistic feeding associations does exist, although it would be difficult to disentangle mutu- alisms from co-occurrence at a favorable feeding site. Also, many individual specimens from this study display regeneration, ostensibly as a result of failed predation, and parasitic borings and overgrowths. More detailed analysis of these nega- tive interactions among biofacies and among taxa may provide important insights into biotic processes which affected the distribution of taxa and biofacies. These "within fauna" patterns are different than the "evolutionary escalation" ascribed to the intensified interactions between crinoids and teleost fish which generated large scale paleoecological and evolutionary patterns (Meyer and Macurda, 1977; Meyer, 1985; Schneider, 1988; Vermeij, 1987).

A fundamental problem in this debate is recon- ciling contradictory observations made by neoecol- ogists, Quaternary paleoecologists, and deeper Phanerozoic paleoecologists (DiMichele, 1994). These contradictions may be resolved by dealing expressly with issues of scale. Brown (1995) and Huston (1994) both note that "Gleasonian" and "Clementsian" arguments can often be resolved by expressly stating the spatio-temporal scale of the ecological processes being observed. Landscape ecologists have long noted that spatio-temporally small scale patterns appear highly variable while at increasing scales this complex mosaic becomes stable and predictable, indeed that large scale stability is dependent on small scale complexity (Watt, 1947; O'Neill et al., 1986; Smith and Urban, 1988). Quaternary terrestrial (Coope, 1979; Huntley and Webb, 1989) and marine (Taylor, 1978; Crame, 1986; Pauly, 1990, 1991; Valentine and Jablonski, 1993) paleoecologists overwhelm- ingly support individualistic and idiosyncratic responses of species to climatic fluctuations. However, an extended succession of these cycles could look strikingly similar to the patterns in the Upper Carboniferous, as noted here. Thus, taxa may be responding independently, or at least not

as integrated functional units, but environmental forcing would produce highly deterministic responses from the fauna. This large scale, recur- ring, deterministic faunal response may indeed be what is recorded during intervals of faunal stability.

Acknowledgements

This paper is a highly modified portion of a Ph.D. dissertation completed at the University of Cincinnati. I thank David Meyer, Arnold Miller, and Wayne Pryor for their thoughtful input into that earlier version. Roger Pabian, Tomasz Baumiller, John Alroy, Andrew Cohen, Karl Flessa, and Michal Kowalewski also provided stimulating discussion on crinoids and paleocommunities. Special thanks go to Dan Mosher, Ted Huscher, John Pope, and Matt Joeckel for their hospitality at home and hard work in the field. John French, Darwin Boardman, and Phil Heckel provided helpful information on paleoecologic, lithostratigraphic, and biostratigraphic work in progress. The thoughtful reviews of Steve Holland and Tom Kammer greatly improved the manu- script. Thanks also to Ken Schopf and Linda Ivany for their stimulating review and invitation to participate in the coordinated stasis symposium. I also thank the following individuals and institu- tions for allowing access to their collections: Julia Golden of the Department of Geology, University of Iowa (SUI); Roger Pabian of the University of Nebraska State Museum (UNSM); Fred Collier (formerly) of the Smithsonian Institution (NMNH); and Dan Mosher (formerly) of Bartlesville Wesleyan College (BWC). I also thank the following organizations for their financial support of this research: Sigma Xi; The Paleontological Society; The Kansas Geological Survey; The Theo. Roosevelt Fund of the American Museum of Natural History; The Caster Fund of the Department of Geology, the Department of Geology, and the University Research Council of the University of Cincinnati; and the National Science Foundation (EAR 9304789). Many of the concepts stated here matured while supported by the University of


Arizona Research Training Group on the Analysis of Biological Diversification.

Appendix 1

Sampled localities from the Lansing Group South Bend Formation (SB-) LXQN. L-X Quarry, Cass County, Nebraska: NW, Sec. 18,

T.12N, R.12E. Crinoids collected from a prominent shale unit in upper half of the regressive Kitaki Limestone Member of the northern shoreward facies belt (Pabian and Strimple, 1993).

ERQK. Elk River Quarry, Montgomery County, Kansas: NE, SW, Sec. 17, T.32S, R.14E. Crinoids collected from southern exposures of upper surface of transitional algal mound facies of transgressive Little Kaw Limestone Member.

RRWK. Railroad Cut, Wayside, Montgomery County, Kansas: NW, NW, Sec. 3, T.34S, R.14E. Crinoids collected from the upper surface of sponge-dominated dense wackestone facies of transgressive Little Kaw Limestone Member.

ARCK. Anderson Ranch, Caney, Montgomery County, Kansas: SW, NW, Sec. 6, T.35S, R.14E. Crinoids collected in gullies behind house from the upper portion of highly argillaceous sponge-dominated packstones and wackestones between basal conglomerate and overlying calcareous sponge-rich shales of the transgressive Little Kaw Limestone. This locality lies on the Chautauqua Arch.

Stanton Formation (ST-) PRQN. Platte River Quarries, Louisville area, Cass and

Sarpy counties, Nebraska: (Rock Lake Quarry; SE, NW, Sec. 3, T.12N, R.10E) (Derby Quarry; NE, NW, Sec. 16, T.12N, R.11E)(Ash Grove Quarry; NW, See. 18, T.12N, R.12E). All crinoids collected from the Kiewitz Shale Bed exposed in quarries along the Platte River Valley near Louisville, Nebraska. This unit is a light gray, slightly to strongly calcareous shale at the base of the regressive Stoner Limestone Member.

LKCK, UKCK. Lower and Upper Kill Creek, Kansas Route 10, DeSoto, Johnson County, Kansas: SE, SE, Sec. 33, T.12S, R.22E. Crinoids were collected from two horizons within the transgressive Captain Creek Limestone. The Stanton Formation at this locality in the Open Marine Facies Belt rests upon a thickened interval of Vilas Formation siliciclastics (P. Heckel, pers. comm.; J. French, pers. comm.). LKCK is from a shale parting and bedding plane at the top of the middle third of the transgressive Captain Creek Limestone. UKCK is from the uppermost 2-3 cm of the Captain Creek.

WLQK. Wilson County Lake Quarry, Buffalo, Wilson County, Kansas: SE, SE, Sec. 17, T.27S, R.16E. Crinoids collected from whole fossil grainstones and shales at the base of the regressive Stoner Limestone Member algal mound complex.

TMQK. Table Mound Quarry, Elk City Lake State Park,

Montgomery County, Kansas: NE, Sec. 16, T.32S, R.15E. Crinoids collected from intermound argillaceous packstones and grainstones in the middle to upper portions of the transgressive Captain Creek Limestone Member algal mound complex.

WINK. West Independence, U.S. Rt. 160, Montgomery County, Kansas: NW, NW, Sec. 36, T.32S, R14E. Crinoids collected from highest fossiliferous shales beneath the Timber Hill Siltstone Bed (Heckel, 1975a). This thick shale unit is classified as the Eudora Shale; however it is equivalent to the upper regressive Stoner Limestone Member, which pinches out between here and the Elk City State Lake dam.

PHFK, BNFK, 75BK (=BOLT of Fig. 13). (Patterson Hog Farm pond: NE, NE, Sec.23; Bolton North Farm pond: SW, SW, Sec. 25; U.S. Rt. 75: NW, NW, Sec. 36), T.33S, R.14E; Bolton, Montgomery County, Kansas. Crinoids from these localities are all from the sponge-rich dense wackestone facies of the transgressive Captain Creek Limestone Member.

PGSK, RFPK. (Patterson Ground Silo: NW, NW, Sec. 24; Rollins Farm Pond: NW, NW, Sec. 25), T.33S, R.14E; Montgomery County, Kansas. Crinoids collected from the upper gray shale facies of the maximum transgressive Eudora Shale Member.

TOQK, ETQK. Tyro Quarry, Montgomery County, Kansas: SW, SE, Sec. 30, T.34S, R.15E. Crinoids were collected from two horizons at the quarry north of Tyro, Kansas. This locality lies on the crest of the Chautauqua Arch. This rise developed a localized oolite shoal during transgression, the Tyro Oolite (Heckel, 1975b; Bennison, pers. comm.), TOQK is from a thin wackestone horizon at the top of the Tyro Oolite, transitional with the overlying gray, phosphatic shales of the Eudora Shale Member. ETQK is from the lower gray shale facies of the maximum transgressive Eudora Shale Member.

LTDO, UTDO. Lower and Upper Tank Dike, Copan, Washington County, Oklahoma: NW, SW, Sec. 10, T.28N, R. 13E. Crinoids collected from the Copan tank dike excavation are from the middle portion of the Barnsdall Formation, which is equivalent to the Stanton Formation. LTDO is apparently transitional with an underlying dysaerobic molluscan fauna equivalent to the maximum transgressive Eudora Shale Member. UTDO occurs several meters above LTDO directly below a nearshore molluscan limestone stringer, above which nonfossiliferous shales and sandstones cap the cycle.

LCWO, UCWO. Lower and Upper Copan West, Oklahoma Route 10, Copan Lake, Washington County, Oklahoma: NW, Sec. 18, T.28N, R.13E. Crinoids collected from two horizons within the transgressive deposits of the Stanton Formation- equivalent Barnsdall Formation. Crinoids from LCWO are directly above the Torpedo Sandstone from early transgressive calcareous shales. UCWO is approximately 7 m above LCWO from late transgressive calcareous shales.

NTSO. North Torpedo Switch, Glade along Cottonwood Lane, Osage County, Oklahoma: NE, SE, Sec. 7, T.26N, R.12E. Crinoids collected from the upper portion of early regressive shale of Stanton-equivalent Barnsdall Formation. This is an attenuated sequence which lies atop the Backius Anticline (Rascoe, 1975).

BCQO, SLWO. (Bull Creek Quarry: SE, SE, Sec. 4; Skiatook

P.F Holterhoff/Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996) 47 81 77

Lake West, Oklahoma Route 20: NE, NW, Sec. 21), T.22N, R.10E: Osage County, Oklahoma. Crinoids collected from transgressive argillaceous packstones of the Stanton-equivalent Barnsdall Formation.

Plattsburg Formation ( P L - )

CCQN. Cedar Creek Quarry, Cass County, Nebraska: SW, Sec. 7, T.12N, R.12E. Crinoids were collected from the top of the middle third of the transgressive Merriam Limestone Member.

075K. Old U.S. Rt. 75, north Altoona, Wilson County, Kansas: NE, NW, Sec. 28, T.28S, R.16E. Crinoids from this locality were collected from the middle to upper wackestones of the regressive Spring Hill Limestone Member.

N75K, K47K, K96K, NFDK, BRQK. (New U.S. Rt. 75: SE, NE, Sec. 28, T.28S, R.16E; Kansas Rt. 47: NW, NW, Sec. 17, T.29S, R.16E; Kansas Rt. 96 and Neodesha Farm Ditch: SE, SW, Sec. 23, T.30S, R.15E; Brickton Quarry: NE, NE, See. 1, T.31S, R.15E), Wilson and Montgomery counties, Kansas. Crinoids from all of these localities were collected from the Hickory Creek Shale Member. NFDK is from the lowest, near- maximum flooding horizon of the Hickory Creek Shale and represents the deepest conditions of these samples. BRQK, K47K and N75K are from higher in the Hickory Creek Shale and represent early regressive conditions. K96K comes from the transition of the Hickory Creek Shale into the algal mound facies of the Spring Hill Limestone and therefore higher in the regressive hemicycle.

TMBO. Top of The Mound, Borrow Pit, Bartlesville, Osage County, Oklahoma: SE, Sec 3, T.26N, R.12E. Crinoids were collected from the lower portion of the transgressive facies of the Plattsburg-equivalent upper Warm Formation. This locality is on the Dewey Faulted Flexure, a basement related structure within the basinal facies belt.

HPTO. Hairpin Turn on county road, Osage County, Oklahoma: SW, Sec. 15, T.25N, R.12E. Crinoids collected from the upper portions of the transgressive unit of the Plattsburg-equivalent upper Warm Formation.

CBMO. Crystal Bay Marina, Parking lot exposure, Skiatook Lake, Osage County, Oklahoma: NE, Sec. 21, T.22N, R.11E. Crinoids collected from the medial shale of a thin limestone and shale horizon within the upper Wann Formation, equivalent to the Plattsburg Formation.

References

Albrecht, G.H., 1980. Multivariate analysis and the study of form, with special reference to canonical variate analysis. Am. Zool., 20: 679-693.

Arvidson, R.S., 1990. Stratigraphy, carbonate petrology, dia- genesis, and trace element cement Geochemistry of the Wyandotte Limestone (Upper Pennsylvanian), Miami County, Kansas. Thesis. Univ. Iowa, 205 pp.

Bambach, R.K., 1994. The null hypothesis for community sta-

bility and recurrence-environmental selection of adapted organisms. Geol. Soc. Am. Abstr. Programs, 26(7): A-519.

Barkai, A. and McQuaid, C., 1988. Predator-prey role reversal in a marine benthic ecosystem. Science, 242: 62-64.

Baumiller, T.K., 1990. Crinoid functional morphology and the energetics of passive suspension feeding: implications to the evolutionary history of Paleozoic Crinoidea. Theis. Univ. Chicago, 150 pp. (unpublished).

Baumiller, T.K., 1993. Survivorship analysis of Paleozoic Cri- noidea: effect of filter morphology on evolutionary rates. Paleobiology, 19(3): 304-321.

Bennington, J.B. and Bambach, R.K., 1996. Statistical testing for paleocommunity recurrence: are similar fossil assemblages ever the same? In: L.C. Ivany and K.M. Schopf (Edi- tors), New Perspectives on Faunal Stability in the Fossil Record. Palaeogeogr. Palaeoclimatol. Palaeoecol., xx:xx-xx.

Bennison, A.P., 1984. Shelf to trough correlations of Late Des- moinesian and Early Missourian carbonate banks and related strata, Northeast Oklahoma. In: N.J. Hyne (Editor), Limestones of the Mid-Continent. Tulsa Geol. Soc. Spec. Publ., 2: 93-126.

Bennison, A.P., 1985. Trough-to-shelf sequence of the early Missourian Skiatook Group, Oklahoma and Kansas. In: W.L. Watney et al. (Editors), Recent Interpretations of Late Paleozoic Cyclothems. SEPM Mid-Cont. Sect. Proc. 3rd Annu. Meet. Field Conf., pp. 219 245.

Bjerstedt, T.W., 1988. Multivariate analyses of trace fossil distribution from an Early Mississippian oxygen-deficient basin, central Appalachians. Palaios, 3: 53-68.

Boardman, D.R. and Heckel, P.H., 1989. Glacial-eustatic sea- level curve for early Late Pennsylvanian sequence in north- central Texas and biostratigraphic correlation with curve for midcontinent North America. Geology, 17: 802-805.

Boardman, D.R., Mapes, R.H., Yancey, T.E. and Malinky, J.M., 1984. A new model for the depth-related allogenic community succession within North American Pennsylvanian cyclothems and implications on the black shale problem. In: N.J. Hyne (Editor), Limestones of the Midcontinent. Tulsa Geol. Soc. Spec. Publ., 2: 141-182.

Boucot, A.J., 1983. Does evolution take place in an ecological vacuum? II. J. Paleontol., 57 (1): 1-30.

Boucot, A.J., 1986. Ecostratigraphic criteria for evaluating the magnitude, character and duration of bioevents. In: O.H. Walliser (Editor), Global Bio-Events. Lecture Notes Earth Sci., 8: 25-45.

Boucot, A.J., 1990. Community evolution: its evolutionary and biostratigraphic significance. In: W. Miller (Editor), Paleo- community Temporal Dynamics: The Long-Term Develop- ment of Multispecies Assemblages. Paleontol. Soc. Spec. Publ., 5: 48-70.

Brett, C.E. and Baird, G.C., 1985. Carbonate-shale cycles in the Middle Devonian of New York: an evaluation of models for the origin of limestones in terrigenous shelf sequences. Geology, 13: 324-327.

Brett, C.E. and Baird, G.C., 1986. Symmetrical and upward shallowing cycles in the Middle Devonian of New York State

App

endi

x 2-

-Cri

noid

abu

ndan

ce a

mon

g L

ansi

ng G

roup

loca

litie

s

Sou

th B

end

Fm.

Sta

nton

Fm

. P

latts

burg

Fm

. LX

QN

ER

QK

AR

CK

RR

WK

S

LW

O B

CQ

O W

INK

UT

DO

LK

CK

PR

QN

TM

QK

WLQ

K L

CW

O N

TS

O B

NF

K

75B

K T

OQ

K P

HF

K U

KC

K U

CW

O L

TO

O E

TQ

K R

FP

K P

GS

K

CC

QN

CB

MO

TM

BO

O75

K

KO

6K K

47K

N

75K

BR

QK

NF

DK

HP

TO

cro

my

21

1 1

3 17

2

12

3 2

32

0 4

12

11

0 0

0 5

1 5

2 0

0 0

13

2 11

8

5 1

2 O

0

4

Ste

ll.

1 5

3 7

46

7 6

6 0

8 1

9 11

16

2

0 1

0 10

4

40

0 0

0 0

1 7

1 8

1 6

0 4

13

.w~

Del

oc.

44

7

1 3

49

6 24

5

22

361

4 77

15

14

29

79

69

32

15

10

11

17

2

0 14

2

16

190

5 10

9

4 9

33

Gra

ft.

0 0

0 0

1 0

23

0 4

82

0 8

5 4

0 0

0 0

0 13

0

0 0

0 1

0 11

14

5

2 6

1 0

3 ~

"

Er}

so.

6 16

2

43

10

2 14

7

3 14

1 7

180

149

36

21

143

65

35

74

52

60

7

0 0

3 1

12

105

40

28

61

44

20

22

~

'

Ulo

cr.

3 3

0 1

1 1

2 1

0 0

0 0

1 0

0 0

0 1

1 0

0 0

2 0

0 3

~"'

: 0

25

9 8

1 0

0 0

Ter

pn.

3 0

0 1

0 0

0 0

0 15

0

2 0

1 0

0 0

0 0

0 0

O

0 0

0 0

0 0

0 0

0 0

0 3

~"

G[a

uk.

3 1

0 0

8 3

0 2

0 5

0 1

1 2

1 0

0 0

2 0

0 0

0 0

0 1

5 0

0 0

1 0

1 0

C~

cb

laud

o.

0 1

2 4

1 0

0 1

0 40

1

45

2 1

1 5

6 1

0 3

0 O

0

0 3

0 3

11

2 0

1 0

1 5

Plu

mm

0

0 0

0 4

2 1

0 0

11

0 1

0 4

0 0

0 0

0 0

1 0

0 0

0 0

6 0

0 0

1 0

1 0

Exo

ri.

0 1

0 2

7 2

1 3

1 19

1

2 1

12

1 14

0

7 2

6 3

0 0

0 0

0 5

1 0

0 4

0 1

10

plra

$.

15

3 2

15

33

5 25

23

6

93

1 79

38

34

3

0 5

1 7

12

20

0 0

0 12

1

12

13

10

2 5

1 2

27

scyt

a.

8 4

1 3

2 1

3 2

1 38

0

8 1

1 3

2 1

3 1

3 13

0

0 0

0 0

2 1

1 0

1 0

1 10

$ten

o.

0 0

1 0

3 3

0 0

1 32

3

29

16

75

2 2

2 0

8 22

29

0

0 0

1 1

12

16

14

1 5

7 1

30

~7

ampe

l.

1 0

0 0

62

5 5

4 1

21

0 3

2 10

2

0 1

0 1

0 11

0

0 0

17

1 6

0 1

0 2

0 0

3

cym

bi.

3

1 1

1 7

2 9

3 0

26

1 7

1 6

1 0

1 4

4 1

10

0 0

0 24

0

B

1 0

0 3

0 1

4 ~,

~

Elib

a.

3 8

0 3

2 2

1 1

0 74

3

15

12

1 0

4 0

10

3 19

12

2

0 0

1 1

1 0

0 0

2 0

1 1

Exo

cr.

1 4

0 1

1 0

2 0

0 32

0

21

2 1

7 2

5 0

11

0 37

5 21

3 1

0 0

0 1

0 0

2 2

1 4

19

~'

Gal

at.

0 0

0 1

0 0

1 0

0 0

0 15

3

0 5

7 0

6 1

8 21

2

0 0

0 0

2 1

1 3

0 0

12

12

Apo

gr.

3 12

5

102

2 0

1 5

0 27

7 0

73

111

80

73

159

272

382

261

150

331

99

2 0

0 0

1 6

6 12

26

27

3

0

398

gra

ph

. 0

1 0

10

0 0

0 0

0 12

0

3 0

0 0

1 0

0 0

0 0

0 0

0 1

0 G

13

1

3 4

13

9 39

3 "~

Kal

li.

0 3

0 10

0

0 0

0 1

39

0 5

66

11

3 30

0

25

122

273

45

12

0 0

0 0

0 0

0 1

0 0

12

13

~

Lecy

t. 0

0 0

7 0

0 0

0 0

0 1

31

0 0

7 25

0

14

0 0

6 0

0 0

0 0

0 3

1 0

0 0

1 0

Par

am

0 O

0

0 0

0 0

0 0

0 0

1 21

8

4 32

2

7 2

41

16

0 0

0 0

0 0

2 0

2 2

0 1

22

Cib

~.

0 0

0 30

1

0 0

1 0

8 1

5 1

3 7?

28

3 10

1 24

1 6

196

8 5

0 0

0 0

5 23

4

7 20

2

27

127

Euo

ny.

0 0

0 O

0

0 0

0 0

0 0

0 1

3 5

25

2 21

0

3 1

3 0

0 0

0 0

0 0

0 O

0

0 2

~.

Par

ag.

0 0

0 0

0 0

0 0

0 0

0 0

0 0

0 0

0 0

0 0

0 1

80

33

0 0

0 0

0 0

0 0

0 0

Tot

al

115

71

19

247

257

43

130

67

42

1391

33

63

2 47

2 33

4 24

7 81

3 53

3 79

4 53

1 82

1 10

16

361

85

33

90

12

12

7 40

9 10

4 75

16

5 10

0 13

9 11

57

* Ta

xono

mic

abb

revi

atio

ns a

fter

Fig

. 4.

P.E Holterhoff/Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996) 47-81 79

and their implications for the punctuated aggradational cycle hypothesis. Paleoceanography, 1: 431-445.

Brett, C.E. and Baird, G.C., 1995. Coordinated stasis and evolutionary ecology .of Silurian to middle Devonian faunas in the Appalachian Basin. In: D.H. Erwin and R.L. Anstey (Editors), New Approaches to Speciation in the Fossil Record. Columbia Univ. Press, New York, pp. 285-315.

Brett, C.E., Miller, K.B. and Baird, G.C., 1990. A temporal hierarchy of paleoecological processes within a Middle Dev- onian epeiric sea. In: W. Miller (Editor), Paleocommunity Temporal Dynamics: The Long-Term Development of Multispecies Assemblages. Paleontol. Soc. Spec. Publ., 5: 178-209.

Brown, J.H., 1995. Macroecology. Univ. Chicago Press, Chi- cago, 269 pp.

Burchett, R.R., 1971. Guidebook to the Geology Along Por- tions of the Lower Platte River Valley and Weeping Water Valley of Eastern Nebraska. Conserv. Surv. Div. Nebraska Geol. Surv., Univ. Nebraska, 39 pp.

Burchett, R.R. and Reed, E.C., 1967. Centennial Guidebook to the Geology of Southeastern Nebraska. Conservation and Survey Division, Nebraska Geological Survey, University of Nebraska, Lincoln, 83 pp.

Buzas, M.A. and Culver, S.J., 1994. Species pool and dynamics of marine paleocommunities. Science, 264: 1439-1441.

Cocke, J.M. and Strimple, H.L., 1974. Distribution of Algae and Corals in Upper Pennsylvanian Missourian Rocks in Northeastern Oklahoma. In: Field Trip Guidebook, South- Central Sect. Geol. Soc. Am., 42 pp.

Cohen, A.S., 1994. Extinction in ancient lakes: biodiversity crises and conservation 40 years after J.L. Brooks. In: K. Martens et al. (Editors), Speciation in Ancient Lakes. Arch. Hydrobiol. Beih. Ergebn. Limnol., 44: 451--479.

Coope, G.R., 1979. Late Cenozoic fossil Coleoptera: Evolution, Biogeography, and Ecology. Annu. Rev. Ecol. Syst., 10: 247-267.

Crame, J.A., 1986. Late Pleistocene molluscan assemblages from the coral reefs of the Kenya coast. Coral Reefs, 4: 183-196.

DiMichele, W.A., 1994. Ecological patterns in time and space. Paleobiology, 20(2): 89 92.

Drake, J.A., 1990. Communities as assembled structures: do rules govern pattern? TREE, 5: 159-164.

Elias, M.K., 1937. Depth of deposition of the Big Blue (Late Paleozoic) sediments in Kansas. Geol. Soc. Am. Bull., 48: 403 432.

Fasham, M.J.R., 1977. A comparison of nonmetric multidimensional scaling, principal components and reciprocal averaging for the ordination of simulated coenoclines and coenoplanes. Ecology, 58: 551-561.

Fursich, F.T., Oschmann, W., Jaitly, A.K. and Singh, I.B., 1991. Faunal response to transgressive regressive cycles: example from the Jurassic of Western India. Palaeogeogr. Palaeoclimatol. Palaeoecol., 85: 149-159.

Galloway, W.E., Yancey, M.S. and Whipple, A.P., 1977. Seismic stratigraphic model of depositional platform margin, eastern Anadarko Basin, Oklahoma. AAPG Bull., 61: 1437 1447.

Gay, S.P., 1989. Gravitational compaction, a neglected mecha- nism in structural and stratigraphic studies: new evidence from mid-continent, USA. AAPG Bull., 73: 641-657.

Gnanadesikan, R., 1977. Methods for Statistical Data Analysis of Multivariate Observations. Wiley, New York, 311 pp.

Gnanadesikan, R. and 11 others, 1989. Discriminant analysis and clustering. Statist. Sci., 4: 34-69.

Greenstein, B.J., 1991. An integrated study of echinoid taphonomy: predictions for the fossil record of four echinoid families. Palaios, 6: 519-540.

Hallock, P. and Schlanger, W., 1986. Nutrient excess and the demise of coral reefs and carbonate platforms. Palaios, 1: 389-398.

Handford, C.R. and Dutton, S.P., 1980. Pennsylvanian-Early Permian depositional systems and shelf-margin evolution, Palo Duro Basin, Texas. AAPG Bull., 64:88 106.

Hanski, I., 1989. Metapopulation dynamics: does it help to have more of the same? TREE, 4:113-114.

Hanski, I., 1991. Single-species metapopulation dynamics: concepts, models and observations. Biol. J. Linnean Soc., 42: 17-38.

Hanski, I. and Gilpin, M., 1991. Metapopulation dynamics: brief history and conceptual domain. Biol. J. Linnean Soc., 42: 3-16.

Harbaugh, J.W., Merriam, D.F., Wray, J.L. and Jacques, T.E., 1965. Pennsylvanian Marine Banks in Southeastern Kansas. Geol. Soc. Am. Field Conf. Guide Book Annu. Meet., 66 pp.

Heckel, P.H., 1975a. Stratigraphy and Depositional Framework of the Stanton Formation in Southeastern Kansas. Kans. Geol. Surv. Bull., 210, 45 pp.

Heckel, P.H., 1975b. Upper Pennsylvanian Limestone Facies in Southeastern Kansas. In: Karts. Geol. Soc. 31st Annu. Field Conf. Guidebook, 66 pp.

Heckel, P.H., 1977. Origin of phosphatic black shale facies in Pennsylvanian cyclothems of imdcontinent North America. AAPG Bull., 61:1045 1068.

Heckel, P.H., 1978. Upper Pennsylvanian cyclotheimc limestone facies in Eastern Kansas. Kans. Geol. Surv. Guide- book Ser., 2, pp. 79.

Heckel, P.H., 1980. Paleogeography of eustatic model for deposition of midcontinent Upper Pennsylvanian cyclothems. In: T.D. Fouch and E.R. Magathan (Editors), Paleo- zoic Paleogeography of West Central United States. Rocky Mount. Sect. SEPM Spec. Publ., Denver, pp. 197-215.

Heckel, P.H., 1986. Sea-level curve for Pennsylvanian eustatic marine transgressive-regressive depositional cycles along midcontinent outcrop belt, North America. Geology, 14: 330-334.

Heckel, P.H. and Pope, J.P., 1992. Stratigraphy and cyclic sedimentation of Middle and Upper Pennsylvanian Strata around Winterset, Iowa. Iowa Geol. Surv. Guidebook Ser., 14, 53 pp.

Heckel, P.H., Brady, L.L., Ebanks, W.J. and Pabian, R.K., 1979. Pennsylvanian Cyclic Platform Deposits of Kansas and Nebraska (Guidebook Ser., 4). Kans. Geol. Surv., 79 pp.

Hoffman, A., 1981. The ecostratigraphic paradigm. Lethaia, 14: 1-7.

80 P. F Holterhoff/Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996) 47-81

Holterhoff, P.F., 1993. The application of filtration theory to the paleoecology of crinoids from the Lansing Group (Mis- sourian, Upper Pennsylvanian), Midcontinent, North Amer- ica. Thesis. Univ. Cincinnati, 307 pp.

Holterhoff, P.F., 1995. Diversity structure of Permo-Pennsylva- nian crinoid faunas from midcontinent North America: relationships between alpha and gamma. Geol. Soc. Am. Abstr. Programs, 27(3), p. 60.

Holterhoff, P.F., submitted. Filtration models, guilds, and biofacies: crinoid paleoecology of the Stanton Formation (Upper Pennsylvanian), midcontinent, North America. Palaeogeogr. Palaeoclimatol. Palaeoecol.

Huntley, B. and Webb, T., 1989. Migration: species' response to climatic variations caused by changes in the earth's orbit. J. Biogeogr., 16: 5-19.

Huston, M.A., 1994. Biological Diversity. Cambridge Univ. Press, Cambridge, 681 pp.

Imbrie, W.S., 1985. A theoretical framework for the Pleistocene ice ages. J. Geol. Soc. London, 142: 417-432.

Ireland, H.A., 1955. Pre-cambrian surface in northeastern Okla- homa and parts of adjacent states. AAPG Bull., 39: 468-483.

Joeckel, R.M., 1989. Geomorphology of a Pennsylvanian land surface: pedogenesis in the Rock Lake Shale Member, southeastern Nebraska. J. Sediment. Petrol., 59: 469-481.

Kammer, T.W., 1985. Aerosol filtration theory applied to Mis- sissippian deltaic crinoids. J. Paleontol., 59: 551-560.

Kammer, T.W. and Ausich, W.I., 1987. Aerosol suspension feeding and current velocities: distributional controls for Late Osagean crinoids. Paleobiology, 13: 379-395.

Kidwell, S.M., 1991a. The stratigraphy of shell concentrations. In: P.A. Allison and D.E.G. Briggs (Editors), Taphonomy: Releasing the Data Locked in the Fossil Record. Plenum, New York, pp. 212-290.

K.idwell, S.M., 1991b. Condensed deposits in siliciclastic sequences: expected and observed features. In: G. Einsele et al. (Editors), Cycles and Events in Stratigraphy. Springer, Berlin, pp. 682-695.

KidweU, S.M. and Flessa, K., 1995. The Quality of the Fossil Record: Populations, Species, and Communities. Annu. Rev. Ecol. Syst., 26: 269-299.

Kovach, W.L., 1988. Multivariate methods of analyzing paleoecological data. In: W.A. DiMichele and S.L. Wing (Edi- tors), Methods and Applications of Plant Paleoecology. Paleontol. Soc. Spec. Publ., 3: 72-104.

Kumar, N. and Slatt, R.M., 1984. Submarine-fan and slope facies of Tonkawa (Missourian-Virgilian) Sandstone in deep Anadarko Basin. AAPG Bull.,, 68: 1839-1856.

Lane, N.G. and Webster, G.D., 1980. Crinoidea. In: T.W. Broadhead and J.A. Waters (Editors), Echinoderms. Univ. Tenn. Stud. Geol., 3: 144-157.

Lewis, R., 1980. Taphonomy. In: T.W. Broadhead and J.A. Waters (Editors), Echinoderms. Univ. Tenn. Stud. Geol., 3: 27-39.

Macurda, D.B. and Meyer, D.L., 1983. Sea lillies and feather stars. Am. Sci., 71: 354-365.

Meyer, D.L., 1973. Feeding behavior and ecology of shallow- water unstalked crinoids (Echinodermata) in the Caribbean Sea. Mar. Biol., 22: 105-129.

Meyer, D.L., 1979. Length and spacing of the tube feet in crinoids (Echinodermata) and their role in suspension-feeding. Mar. Biol., 51: 361-369.

Meyer, D.L., 1982. Food and feeding mechanisms: Crinozoa. In: M. Jangoux and J.M. Lawrence (Editors), Echinoderm Nutrition. Balkema, Rotterdam, pp. 25-42.

Meyer, D.L., 1985. Evolutionary implications of predation on Recent comatulid crinoids from the Great Barrier Reef. Paleobiology, 11: 154-164.

Meyer, D.L. and Macurda, D.B., 1977. Adaptive radiation of the comatulid crinoids. Paleobiology, 3: 74-82.

Meyer, D.L. and Ausich, W.I., 1983. Biotic interactions among Recent and among fossil crinoids. In: M.J.S. Tevesz and P.L. McCall (Editors), Biotic Interactions in Recent and Fossil Benthic Communities. Plenum, New York, pp. 377-427.

Meyer, D.L. and Meyer, K., 1986. Biostratinomy of Recent crinoids (Echinodermata) at Lizard Island, Great Barrier Reef, Australia. Palaios, 1: 294-302.

Meyer, D.L., Ausich, W.I. and Terry, R.E., 1989. Comparative taphonomy of echinoderms in carbonate facies: Fort Payne Formation (Lower Mississippian) of Kentucky and Tennes- see. Palaios, 4: 533-552.

Miller, A.I., 1988. Spatial resolution in subfossil molluscan remains: implications for paleobiological analysis. Paleobiol- ogy, 14: 91-103.

Miller, K.B., Brett, C.E. and Parsons, K.M., 1988. The paleoecologic significance of storm-generated disturbance within a Middle Devonian muddy epeiric sea. Palaios, 3: 35-52.

Miller, W., 1990. Hierarchy, individuality, and paleoeco- systems. In: W. Miller (Editor), Paleocommunity Temporal Dynamics: The Long-Term Development of Multispecies Assemblages. Paleontol. Soc. Spec. Publ., 5: 31-47.

Miller, W., 1993a. Models of recurrent fossil assemblages. Leth- aia, 26: 182-183.

Miller, W., 1993b. Benthic community replacement and population response. Neues Jahrb. Geol. Palaontol. Abh., 188:133 146.

Miller, W., 1994. Ecology of coordinated stasis. Geol. Soc. Am. Abstr. with Programs, 26(7): A-520.

Moore, R.C. and 8 others, 1978. Systematic descriptions. In: R.C. Moore and C. Teichert (Editors), Treatise on Inverte- brate Paleontology, Part T, Echinodermata 2, Vol. 2. Geol. Soc. Am., Boulder, Univ. Kansas Press, Lawrence, pp. 405-812.

Morris, P.J., Ivany, L.C. and Schopf, K.M., 1992. Paleo- ecological stasis in evolutionary theory. Geol. Soc. Am. Abstr. with Programs, 24(7): A313.

Morris, P.J., Ivany, L.C., Schopf, K.M. and Brett, C.E., 1995. The challenge of paleoecological stasis: reassessing sources of evolutionary stability. Proc. Natl. Acad. Sci. USA, 92: 11,269-11,273.

Neumann, A.C. and McIntyre, I., 1985. Reef response to sea- level rise: keep-up, catch-up, give-up. Proc. 5th Int. Coral Reef Congr., 3, pp. 105-110.

O'Neill, R.V., DeAngelis, D.L., Waide, J.B. and Allen, T.F.H., 1986. A Hierarchical Concept of Ecosystems (Monogr. Popul. Biol., 23). Princeton Univ. Press, 253 pp.

P.F. Holterhoff/Palaeogeography, Palaeoclimatology, Palaeoecology 127 (1996) 47 81 81

Pabian, R.K. and Strimple, H.L., 1993. Taxonomy, Paleoecol- ogy and Biostratigraphy of the Crinoids of the South Bend Limestone (Late Pennsylvanian-Missourian, ?Virgilian) in Southeastern Nebraska and Southeastern Kansas. Nebraska Geol. Surv. Prof. Publ., 1, 55 pp.

Pauly, G., 1990. Effects of late Cenozoic sea-level fluctuations on the bivalve faunas of tropical oceanic islands. Paleobiol- ogy, 16:415 434.

Pauly, G., 1991. Late Cenozoic sea level fluctuations and the diversity and species composition of insular shallow water marine faunas. In: E.C. Dudley (Editor), The Unity of Evo- lutionary Biology. Dioscorides Press, Portland, pp. 184 193.

Petuch, E.J., 1976. An unusual molluscan assemblage from Ven- ezuela. Veliger, 18: 322-325.

Petuch, E.J., 1981. A relict Neogene caenogastropod fauna from northern South America. Malacologia, 20: 307-347.

Petuch, E.J., 1982. Geographical heterochrony: contemporane- ous coexistence of Neogene and Recent molluscan faunas in the Americas. Palaeogeogr. Palaeoclimatol. Palaeoecol., 37: 277 312.

Plotnick, R.E., McKinney, M. and Gardner, R., 1994. Multiscale ecosystem stasis and multiscale disturbances: the thousand natural shocks that flesh is heir to. Geol. Soc. Am. Abstr. with Programs, 26(7): A-520.

Pulliam, H.R., 1988. Sources, sinks, and population regulation. Am. Nat., 132: 652-661.

Rascoe, B., 1962. Regional stratigraphic analysis of Pennsylva- nian and Permian rocks in western mid-continent, Colorado, Kansas, Oklahoma, Texas. AAPG Bull., 46:1345 1370.

Rascoe, B., 1975. Tectonic origin of preconsolidation deforma- tion in Upper Pennsylvanian rocks near Bartlesville, Okla- homa. AAPG Bull., 59:1626 1638.

Rohlf, F.J., 1989. NTSYS-pc, version 1.50. Exeter Publ., Setauket.

Roughgarden, J., Pennington, T., and Alexander, S., 1994. Dynamics of the rocky intertidal zone with remarks on gen- eralization in ecology. Philos. Trans. R. Soc. Lond. B, 343: 79-85.

SAS Institute Inc., 1989. SAS/STAT User's Guide, Version 6, 4th Ed., Vol. 1. SAS Institute Inc., Cary, NC, 943 pp.

Savarese, M., Gray, L.M. and Brett, C.E., 1986. Faunal and lithologic cyclicity in the Centerfield Member (Middle Dev- onian, Hamilton Group) of Western New York: a reinter- pretation of depositional history. In: C.E. Brett (Editor), Dynamic Stratigraphy and Depositional Environments of the Hamilton Group (Middle Devonian) in New York State, Part I. N.Y. State Mus. Bull., 457: 32-56.

Schneider, J.A., 1988. Frequency of arm regeneration of comatulid crinoids in relation to life habit. In: R.D. Burke et al. (Editors), Echinoderm Biology. Balkema, Rotterdam, pp. 531-538.

Schopf, K.M., 1996. Coordinated stasis: biofacies revisited and the conceptual modeling of whole-fauna dynamics. In: L.C. Ivany and K.M. Schopf (Editors), New Perspectives on Faunal Stability in the Fossil Record. Palaeogeogr. Palaeocli- matol. Palaeoecol., 127: 157-175.

Schopf, K.M., Ivany, L.C. and Morris, P.J., 1992. Onshore--

offshore trends in light of ecological locking. Geol. Soc. Am. Abstr. with Programs, 24 (7):A313.

Schubert, J.K. and Bottjer, D.J., 1995. Aftermath of the Per- mian-Triassic mass extinction event: paleoecology of Lower Triassic carbonates in the western USA. Palaeogeogr. Palaeoclimatol. Palaeoecol., 116: 1-39.

Sheehan, P.M., 1991. Patterns of synecology during the Phaner- ozoic. In: E.C. Dudley (Editor), The Unity of Evolutionary Biology. Dioscorides Press, Portland, pp. 103-118.

Sheldon, P.R., 1993. Making sense of microevolutionary patterns. In: D.R. Lees and D. Edwards (Editors), Evolutionary Patterns and Processes. Academic Press, London, pp. 19-31.

Smith, T.M. and Urban, D.L., 1988. Scale and resolution of forest structural pattern. Vegetatio, 74: 143-150.

Sneath, P.H.A. and Sokal, R.R., 1973. Numerical Taxonomy. Freeman, San Francisco, 573 pp.

Sokal, R.R. and Rohlf, F.J., 1981. Biometry. Freeman, San Francisco, 859 pp.

Sorensen, J.H., Johnsgard, S.J. and Wozencraft, C., 1989. Bibli- ography of Kansas Geology, 1823-1984. Kans. Geol. Surv. Bull., 221,416 pp.

Springer, D.A. and Bambach, R.K., 1985. Gradient versus cluster analysis of fossil assemblages: a comparison from the Ordovician of southwestern Virginia. Lethaia, 18: 181-198.

Sutherland, J.P., 1974. Multiple stable points in natural communities. Am. Nat., 108: 859-873.

SYSTAT, 1986. Version 3. Evanston, IL, 417 pp. Taylor, J.D., 1978. Faunal response to the instability of reef

habitats: Pleistocene molluscan assemblages of Aldabra Atoll. Palaeontology, 21: 1-30.

Thompson, J.N., 1994. The Coevolutionary Process. Univ. Chi- cago Press, 376 pp.

Underwood, A.J. and Fairweather, P.G., 1989. Supply-side ecology and benthic marine assemblages. TREE, 4: 16-20.

Underwood, M.R., 1984. Stratigraphy and depositional history of the Bourbon Flags (Upper Pennsylvanian), an enigmatic unit in Linn and Bourbon Counties in east-Central Kansas. Thesis. Univ. Iowa, Iowa City, 111 pp. (unpublished).

Valentine, J.W. and Jablonski, D., 1993. Fossil communities: compositional variation at many time scales. In: R.E. Ricklefs and D. Schluter (Editors), Species Diversity in Eco- logical Communities. Univ. Chicago Press, pp. 341-349.

Vermeij, G.J., 1987. Evolution and Escalation: An Ecological History of Life. Princeton Univ. Press, 527 pp,

Warshauer, S.M. and Smosna, R., 1981. On minimum spanning trees and the intergradation of clusters. Math. Geol., 13: 225-235.

Watney, W.L., Kaesler, R.L. and Newell, K.D., 1985. Recent interpretations of Late Paleozoic cyclothems. In: Mid-continent Sect. SEPM 3rd Annu. Field Conf. Guidebook, 271 pp.

Watney, W.L., French, J.A. and Franseen, E.K., 1989. Sequence stratigraphic interpretations and modeling of cyclothems. In: Kans. Geol. Soc. 41st Annu. Field Conf. Guidebook, 211 pp.

Watt, A.S., 1947. Pattern and process in the plant community. J. Ecol., 35: 1-22.

Webster, G.D., 1981. New crinoids from the Naco Formation (Middle Pennsylvanian) of Arizona and a revision of the Family Cromyocrinidae. J. Paleontol., 55:1176- l 199.

Crinoid biofacies in Upper Carboniferous cyclothems, midcontinent North America: faunal tracking and the role of regional processes in biofacies recurrence

Documents