InternationalJournal of Coal Geology, 5 (1985) 43?109 43 Elsevier Science Publishers B.V., Amsterdam ? Printed in The Netherlands STRATIGRAPHIC AND INTERREGIONAL CHANGES IN PENNSYLVANIAN COAL-SWAMP VEGETATION: ENVIRONMENTAL INFERENCES TOM L. PHILLIPS', RUSSEL A. PEPPERS^ and WILLIAM A. DIMICHELE^ ^ Department of Plant Biology, University of Illinois, Urbana, IL 61801, U.S.A. ^ Coal Section, Illinois State Geological Survey, Champaign, IL 61820, U.S.A. ^ Department of Botany, University of Washington, Seattle, WA 98195, U.S.A. (Received January 14, 1985) ABSTRACT Phillips, T.L., Peppers, R.A. and DiMichele, W.A,, 1985. Stratigraphic and interregional changes in Pennsylvanian coal-swamp vegetation: Environmental inferences. In: T.L. Phillips and C.B. Cecil (Editors), Paleoclimatic Controls on Coal Resources of the Pennsylvanian System of North America. Int. J. Coal Geol., 5: 43?109. Quantitative analysis of Pennsylvanian coal-swamp vegetation provides a means of in- ferring organization and structure of communities. Distribution of these communities further provides inferences about environmental factors, including paleoclimate. Our observations are based on in situ, structurally preserved peat deposits in coal-ball concre- tions from 32 coal seams in the eastern one-half of the United States and from several seams in western Europe and on spore assemblages from more than 150 seams. There were three times of particularly significant and nearly synchronous vegetational changes in the Midcontinent and Appalachian coal regions during the Pennsylvanian Period. Each was different in kind and magnitude. The first marked changes occurred during the early part of the Middle Pennsylvanian with the fluctuating decline in the high level of lycopod dominance. The abundance of cordaites increased. There was a rise in the occurrences of the lycopod herbs to form intercalated marshlands and an overall increase in floral diversity. Changes ensuing from this time also include shifts in dominant species of lycopod trees and a sustained rise in abundance and diversity of tree-fern spores. The next significant time of change was during the middle part of the Middle Pennsylvanian, representing both a culmination of earlier trends and expansions of cordaites in the Mid- continent where there was a maximum change in species without net loss of diversity. Tree ferns and meduUosan pteridosperms attained subdominant levels of abundance and diverse lycopod species dominated except in the Atokan-Desmoinesian transition of the Midcontinent. The third and sharpest break occurred near the Middle?Late Pennsylvanian boundary when extinctions of the dominant, coal-swamp lycopods allowed development of tree-fern dominance. The Late Pennsylvanian coal swamps apparently were colonized or recolonized mainly by species from outside coal swamps rather than by the survivor populations of the Middle Pennsylvanian swamps. Paralleling the changes in floras through the Pennsylvanian are changes in preserva- tional aspects of the peat. These include a decline in shoot/root ratios from approximate- ly 1 to < 1 during the first time of vegetational change and a rise in this ratio during the second; there was a parallel rise and fall in fusain abundance and a rise in wood/periderm ratios. The stratigraphic distribution of identified coal resources in the United States is 0166-5162/85/$03.30 ? 1985 Elsevier Science Publishers B.V. 44 interpreted as largely dependent on net changes in relative wetness of Pennsylvanian coal swamps, a pattern of drying during the first period of vegetational change, followed by a concomitant increase in continuous wet climate with brackish influence in the Midcon- tinent during the second; this was followed by a time of extreme moisture stress bringing on the third, and most severe, vegetational change. INTRODUCTION The plant ecology of Pennsylvanian tropical coal swamps is reconstructed in part by analyzing forest communities from permineralized peat deposits preserved in situ within coal-ball concretions. Species composition, vegeta- tional structure, and relative dominance (abundance) patterns of species are the community traits in peats that allow us to recognize different and often specific kinds of environments. The growth, habits, and reproductive biology of the plants, primarily the trees, are major bases for reconstructing hypo- thetical habitats and environments of peat accumulation. Since most of the abundant spores in the coals can be related directly to parent species or genera, ecological information also can be derived from palynology. This is especially important for evaluating interregional similarities in vegetational change through time. In this paper we outline the patterns of change in coal-swamp vegetation during the Westphalian and early Stephanian of western Europe and during the Pennsylvanian of the Appalachian and Midcontinent regions of the United States. The cumulative data provide a comparative and quantitative reference base that: (1) allow estimation of vegetational composition of coal swamps; (2) an evaluation of ecological interpretations on a geological time scale; and (3) the partial characterization of the major periods, kinds, and causes of vegetational change in coal swamps. It is our intent here to augment what is known of basinal geology with paleoclimatic and paleogeographic inferences, which also have significant bearing on the patterns of distribution of vegetation. The names Lower (Early), Middle, and Upper (Late) Pennsylvanian are used in this report according to the recommendations of Bradley (1956) and as commonly used by the U.S. Geological Survey. The top of the New River Formation of West Virginia and equivalent strata mark the Lower?Middle Pennsylvanian boundary; the Allegheny?Conemaugh boundary of the north- ern part of the Appalachian region is equivalent to the Middle?Upper Penn- sylvanian boundary. Three major times of change in coal-swamp vegetation, each of different magnitude, are considered. Each was synchronous, or nearly so, across vary- ing extents of the Euramerican paleotropical belt. The two most conspicuous times of departure from abundant moisture in the lowlands of the Eurameri- can coal belt began during the early part of the Middle Pennsylvanian (West- phalian B) and near the Middle?Late Pennsylvanian (Westphalian?Stephani- an) transition and are referred to as the "first drier" and "second drier" 45 intervals, respectively. At these times there were major changes in coal- swamp vegetation, especially during the early part of the Late Pennsylvanian. The significance of the latter was discussed earlier by comparing palynolog- ical and peat data (Peppers and Phillips, 1973; Phillips et al., 1974). The vegetational change during the Late Pennsylvanian, which was by far the most wide-reaching, was also detected in earlier studies of both clastic sedi- ments and coal in the United States (White and Thiessen, 1913; Kosanke, 1947; Schemel, 1957; Winslow, 1959) and in Europe (Davies, 1929;Stsche- golev, 1975). White, in particular (White and Thiessen, 1913, pp. 76?77), noted the same pattern appEirently from field observations of compression floras, underclays, and coal partings long before extensive data from spore and coal-ball peat studies were available: "The Conemaugh (of lower Stepha- nian age) time witnessed several changes in the floras which may be of cli- matic cause. Most prominent among these are a rapid decrease, approaching extinction, of the colossal lycopods (Lepidodendreae), and the rapid devel- opment of the group of gigantic tree ferns, such as Psaronius, whose sup- posed fronds, Pecopteris, became highly varied, very large, and more or less distinctly villous in most species. The evidence therefore, points to the oc- currence of short dry seasons." The absence of stigmarian seat-earths in the Upper Pennsylvanian has been noted many times (Huddle and Patterson, 1961). Similar observations by Feys (1964, p. 67) in the Stephanian of the Massif Central were attributed to the limnic nature of the basins in France. A much more subtle series of vegetational changes occurred during the "first drier interval" than during the Middle to Late Pennsylvanian transi- tion. These were documented first palynologically in the Illinois Basin Coal Field by Peppers (1979), subsequently compared with northeastern Tennessee (Phillips and Peppers, 1984) and now with eastern Kentucky by means of spore floras and peat. The drier intervals may be interpreted as two progres- sively-more-severe, pulse-like changes that most significantly altered the trop- ical coal-swamp vegetation of the Pennsylvanian. The Early Permian may be regarded as the beginning of the third and driest interval, essentially termi- nating the age of Euramerican coal swamps. Essential to our inference of "drier intervals" in the Pennsylvanian are interregional similarities of ensuing changes in the coal-swamp vegetation patterns that suggest widespread, near- ly contemporaneous kinds of events. The changes in coal-swamp vegetation between the "first" and "second drier intervals", in the middle part of the Middle Pennsylvanian, have been difficult to compare between the Appalachian region and the Midcontinent because of the dearth of permineralized peat deposits in coal balls in the Appalachian region. Differences in regional dominance patterns occurred during this time. This is due probably in ptirt to regional geological controls that may overshadow the role of paleoclimate during this wet interval. 46 SOURCES OF DATA AND VARIABLE FACTORS IN RELATION TO METHODS Information from in situ peat in coal-ball concretions is derived from 32 bituminous coal seams in the eastern one-half of the United States (Fig. 1) and from coals of Europe (Fig. 2). These provide a basis for the quantitative analyses of coal-swamp vegetation. Localities and pertinent geological infor- mation are given on the maps and in the text or can be found in the compila- tion by Phillips (1980). Determinations of major bituminous coal deposits in the Pennsylvanian System of the United States (Fig. 1) are based on identified bituminous coal resources compiled from reports by state geolog- ical surveys by Phillips et al. (1980). The quantitative data in the United States are derived mostly from the Paleobotanical Collections of the Department of Plant Biology, University of Illinois at Urbana. The collections from the Shuler and Urbandale Mines of Iowa were loaned by the Biological Laboratories of Harvard University; coal-ball peels were made available for the Weldon Mine of Iowa by the Department of Botany, University of Iowa, for the "High Splint Coal" of Tennessee by the Department of Geological Sciences, University of Tennes- see, and for the coal above the Middle Kittanning of Pennsylvania by the Department of Geology, Michigan State University. The data from western Europe were obtained from coal-ball peels from collections of the British Museum of Natural History (London) and Uni- versity of Montpellier (coal balls from the Union Seam of Lancashire, Eng- land) and from the University of Liege (coal balls from the Bouxharmont Seam of Belgium). Slide collections utilized for quantitative analyses include the following: the Kidston Collection, Hunterian Museum at Glasgow, and the Scott and Oliver Collections, Palaeontology Department, British Museum (N.H.) (coal balls from the Upper Foot Seam at Shore-Littleborought); the Felix Collection, Humboldt-Universitat zu Berlin, D.D.R. (coal balls from the Katharina Seam at Langendreer, Ruhr Coal Basin of North-Rhine West- phalia). Spore analyses from more than 150 coals (Figs. 1, 4) in the eastern United States provide information on sequential coal-swamp floras including coals from which coal balls are not available. Coal samples were, in part, provided by the state geological surveys of Oklahoma, Kansas and Kentucky. Spore data from coals outside the Illinois Basin Coal Field are based usually on one channel sample from each coal. Slide preparations and residues are housed in the Coal Section, Illinois State Geological Survey at Urbana. Techniques of analysis The techniques for study and vegetational analysis of peat deposits are those given by Phillips et al. (1976, 1977) and by Phillips and DiMichele (1981). The coal-ball peats provide means of establishing whole plant assem- blages including the plant sources of dispersed spores. Community analyses pp. 47?48 i 11 H 1 V 3 a a 33 1 BIIIASllOd ij? ^& 1 9*5 j o 9 "^ i y 3Aia ? 3 N 1 i 0 a a \? NVXOIV N V M 0 a y 0 33X0a3H 3 3TIIA/k3SVD ASNN3d 3naaiw NVnVHdlS3M a NVnVHdlS3M NVINVAHASNNSd y3M0n V NVnVHdlS3/V\ 0-9 NVIdnwVN n T3 S B 3 0 -K X! >, 0) w a s o n! T1 a C > i^ >, s s -^ a 1) IV 01 Oi .M 0 J3 O ta 4) j3 0 T3 a. 3 * ^ 0 w 1 CMC 1^1 2 M T3 O* 0) ? a, ? > 1^ 1- '^ i* 0 ? (U '^ IS o> CM " 3 c 0 a 0 ?> S^^ ?-3 M c .gi i^ -? 'S -9 8-^1 o 3 g S o c c o o 0 -a 6 O O c o o 0) > A ?> ^ ?a tw '3 C 0 a "5 '-' s D o o o .2-y S JD ^ OH '? 1 1 o o C ? Crt JZ a> i e ii o i: V .2 ?" T3 .- UJ -a j3 -o 5 _o S o % an 5 ^ 1 ? ? o ao 0) .2 fa t3 AN C WESTPHALI^N D STEPHANIAN AUTUNIAN S"TT^ YLVAN1 AN UPPER PENNSYLVANIAN LOWER :PERMIAN U.S,A. SERIES : 1 MARMATON ^TONI'^CITYI'-ANSINGDOUGLASISH/WNEEIWABAUNSE KANSAS GROUPS I 1 ? g S S *S S?3Ss S ? 0 '-<<- I : o S ; > m O m r- ^ 2 '^ i o m Z 3) O 33 M U LBER R Y S U M M IT* M U LKY B E V tE R ? CRO W EBURG FLE M IN G ? M IN E R A L * S 'W E IR -P I RO W E E LM O NO DAW AY TH A Y E R > {/) > to M U LBER R Y LE X IN G TO N SU M M IT M U LKY JEVIER/W HEELER* CRO W EBURG ? FLE M IN G M IN E R A L TE BO ffl ? rTSBU R G ffi O RYW O O D ? RO W E (/) o <= 3J M YSTIC M U LKY BEVIER/W HEELER W H ITE B R E AST W ILE Y M U N TE R V ILLE No 5 LUCAS CO .* L.C O AL FORO CO. LAD O SD ALE ? ? M ANBECK ? M ASTIE PLUS CLIFFLAND,HASTIE NO DAW AY O V ID * o > DESMOINESIAN :MISSOURIAN VIRGILIAN WOLF- CAMPIAN MIDCONTI- NENT SERIES SPOON JCARBONDALE |MODESTOJBOND[ MATTOON W?f!N's 3MS SlSigSfSSSg'SSSfc ?s SHU M W AY* ? ? C ALH O U N ? ? O P D Y K E * ? ? FR IE N D SV1LLE ? ? B R ISTO L H IL L *? ? C H APEL {N o,ei*? A TH E N SVILLE * ffl Z o m > O c/5 O -1 > m r 33 z X ? m z o H ? m O 3) 2 ? O 3) - -n moCQOm-<0 ?e? ?*i ? ?e? ?????? ? 3J ^^^^^^ 1 z No, 14 ? No, 13 5 No. 12 ? No. 1 1 5 No. 10 No, 9+ ffl No. 8 ffl No. 7 5 No, 6 ? BANCRO FT No, 4 ? ELM LIC K ? EM PIR E ffl ICE HOUSE ffl f h -< 1 ALLEGHENY CONEMAUGH MONONGAHELA APWLACHliN GROUPS a FORMATIONS CONEMAUGH MONONGAHELA DUNKARD GLENSHAWiCASSELMAN M AH O N IN G U. FR E E PO R T ? L. FR E E PO R T ? U, K IT T A N N IN G ? M , K ITTA N N IN G ffl L, K ITTA N N IN G ? C LA R IO N ? ? U.M ERCER (STOCKTON "A") STOCKTON (L.MERCERfS CO ALBURG W ASH IN G TO N ffl W AYNESBURG ? UNIO NTO W N SEW IC KLEY R ED STO N E ? P ITTS B U R G H * ffl E LK LIC K H A R LE M ffl U. BAKER STO W N ? L. BAKERSTO W N ffl P IN E C H EEK BR USH C R EEK ? m en -\ < o z > > r > 3) > m o o -c i> z i m ' z , C o H ii ill 1 i ii ? >? ?^rf**? ?*?? * * W AYNESBURG SEW IC KLEY R EDSTO NE P IT T S B U R G H ** TH O M AS PR IN C ESS No. 10 PENNSYLVANIA E. KEN TUCKY PRINCESS CO ALS 1 '> M AH O N IN G Q U. FR EEPO R T L. F R E E P O R T* M , K ITTA N N IN G L.K ITTA N N IN G *^ C LA R IO N B R O O K V ILLE TI0N ESTA(No.3B) B E D FO R D UM ERCER(No,3A} L. M ERCER (NO. 31 W AYNESfflJRG "A" W AYNESBURG U N IO NTO W N M EIG S C R EEK FISH PO T R ED STO N E ? P ITTS B U R G H ? AN D ER SO N W ILG U S O X o m z z m w M AH O N IN G U .FR EEPO R T ? L.FR E E PO R T U .KITTAN N IN G M ,a L, K ITTA N N IN G C O AL G RO UPS W AYNESBURG SEW IC KLEY PITTSBU R G H BAR TO N H AR LEM U. BAKERSTOW N L, BAKERSTOW N BRUSH CREEK > -< l- 1> z o pp. 49?50 t- t? C' C" t- u- V 1 (-1) |o 1 (X) I 3 (I) lO (H) |3 O) lo (d) ^3 1 33SO W NviaixHSva snoa3JiNoaavo 31 aa 1 lAi 'iS3 SNomis xis iasnaninod vnaa raaoj 3y3i Nn3w NIZNV fie- on 3N9I0DIA Hsvazins laaasNi is -^^o:3ldVW|S39M0301S ssNiij aw sninaa a3N3isaoa a3isaoH 3a3i3oa33N |N3Mn3>y s:sa. ^3- 9a3aN3XI3 ? * ? O ^ IS 4.1 < >i33avr :ddn :H3wnH30S 111^ H3SV oZZZ o n3H1l*pidVV8: S3i8nsv3w nvo3 3iaaiw VIIODSWON; SNi99or- annHS^ ONvnussMno S3ansv3w : iido yo sn IVOD MSMOliBNOisniwI d3ddn INIOd pMM 3nvasa3Ma: 0 NvnvHdiS3M a NVnVHdlSBM -vHdi^SBM ^-a NvianiwN yvo a3ddn \iASNN3d 3naam NVINVA1ASNN3d a3*01 a o u jp "3 o o o SP 3 -2 w ^ T3 ca O -a -2 .2 ?= a 3 0) ^ ? ^ to ?S'S o-g !- 3 a o t .2 ? * O ji J3 ^ w .& J3 fc( jO 51 determine dominance patterns, community structure and, in relation to evidence from many environmental indicators, relative ecological amplitudes (Phillips and Peppers, 1984) for dominant trees. The major factors that appear to have controlled plant distribution in coal sw^amps were relative wetness, brackish influences and nutrients. One of the most significant limitations to community analysis, besides the difficulty of identifying dismembered plant parts, is reflected in the shoot/ root ratio of peat. In many of the Lower and Middle Pennsylvanian coal swamps, lycopod trees form 60?95% of the peat biomass. It is estimated that 20?30% of their original biomass was root system and 70?80% was aerial plant parts or shoot systems. If a whole tree were preserved as peat, it would result in a shoot/root biomass of about 4/1. The amount of aerial plant material is critical for community analyses that may be based solely on that. Shoot/root ratios of peats are commonly close to 1/1 indicating enormous loss of aerial plant material, which greatly reduces the accuracy of reconstructing standing vegetation from peat deposits preserved in coal balls and may significantly alter peat-spore relationships since spores are more likely to be preserved. However, most of the community analyses are carried out for coal seams with high shoot/root ratios, such as the Springfield (No. 5) and Herrin (No. 6) Coal Members of the Illinois Basin Coal Field (Fig. 1). Shoot/root ratios reported below were derived from large random coal-ball samples or profiles of the permineralized peat. Quantitative relationships of spore floras and peat deposits The relative numerical abundance of palynomorphs (homospores, micro- spores, prepollen and pollen less than about 210 iim in diameter) from chan- nel samples of coal is the main means of estimating regional, interregional and stratigraphic patterns of coal-swamp vegetation. Coal-ball assemblages are a base of reference for these palynological data and provide the most direct evidence of the vegetation and of the botanical constituents of the coals. Both kinds of data, which have different kinds and degrees of resolu- tion, are amenable to making estimates of the vegetation and environments. Each data set provides a basis to evaluate the other and together they provide a much better resolution of the geologic times and kinds of botanical change. Coal-ball peat deposits from a given locality may not be generally representa- tive of the broader composition of the vegetation. In some cases this can be explained by taking into account the paleogeography and environments of deposition, which are important in the interpretation of regional gradients in the vegetation. In other cases we have to rely on data from the spore assemblages for the regional picture. Conversely, the dispersal of spores into coal swamps from surrounding vegetation, particularly acute in those swamps with large perimeter to small areas, may drastically bias estimates of vegeta- tion without comparative data from peat deposits. Both sources of data have preservational biases that need not be presented here, but are nevertheless 52 important and noted elsewhere. The extensive use of coal-spore floras is predicated on the premise that there are relative, predictable relationships (including biases) between the dominant components of spore assemblages and those of peat deposits. Such quantitative relationships need to be out- lined briefly in order to emphasize their particular utility and their differ- ences among the main kinds of trees. Reproductive biology in relation to estimates of vegetation Quantitative estimates of biomass based on miospores and peat may vary significantly because the dominant trees had different reproductive biologies. There are sharp contrasts in resolution levels between the lower vascular plants, which dominated the coal swamps, and the gymnosperms, which were generally subdominant. Habits and habitats are also closely linked with the differences in reproduction. With the exception of calamitean (sphenop- sid) trees, which are omitted here, the principal trees reproduced only sexu- ally. Lower vascular plants. Psaronius tree ferns and the dominant Lycospora- producing lycopods {Lepidophloios, true Lepidodendron and Paralycopo- dites) can be ranked as the first and second, respectively, in so-called over representation in almost any spore flora. The tree ferns were homosporous and produced small spores that were dispersed broadly by wind and water; palynology samples this output. All the lycopod trees were heterosporous; the Lycospora producers released large numbers of moderately small, wind and water-borne microspores. Palynomorph studies measure only the micro- spore fraction from the lycopods. The large reproductive outputs of the tree ferns and Lycospora producers are consistent with extensive geographic dominance of these trees in the Pennsylvanian coal swamps. Palynomorph floras are particularly sensitive to the mutual relative abundances of these two groups and to the regional appearance and extinction of the source plants. The level of resolution is so good that spores from these plants occur in most spore assemblages even if the source plants were growing at some distance from the sample site. It is this representation of regional floras that adds precision to miospore biostratigraphy; at the same time, it may bias an assessment of the local coal-swamp vegetation (Peppers, 1982). Spore floras provide a clear documentation of the relative importance of Lepidophloios harcourtii, L. hallii and Paralycopodites to the extent that Lycospora species can be related to major species of trees. Many patterns determined by palynology have been confirmed from coal-ball peat sources. However, the abundance of Lycospora granulata in the upper Pottsville and Allegheny Formations in the Appalachian region may, in part, represent Lepidodendron hickii. The abundance of Lycospora granulata in the Des- moinesian of the Midcontinent parallels that of the source plant, which is largely Lepidophloios hallii. This places in doubt the interpretation of Lycospora granulata beyond the peat confirmation range of Lepidophloios 53 hallii. Some very tenuous ecological evidence suggests that Lepidodendron hickii and Lepidophloios may both be important sources of Lycospora granulata in the Lower Pennsylvanian and in the middle and upper parts of the Middle Pennsylvanian in the Appalachians. The morphology of their spores may not be significantly distinct to allow differentiation when making spore analyses by use of standard light microscope. It is noteworthy that the two fundamentally most similar lycopods in the peat assemblages are Lepi- dophloios and Lepidodendron hickii. Where the significantly different leaf cushions are missing or the stelar structure is unpreserved, their tissues are generally indistinguishable. There are two generically distinct groups currently included in Lepidoden- dron. In this paper we refer to one group as "coal-swamp" Lepidodendron (L. vasculare, L. scleroticum, L. phillipsii, L. dicentricum) and the other as true Lepidodendron (L. hickii, L. serratum) (DiMichele, 1983). Peat-based studies (Leclercq, 1930) document the importance of coal-swamp Lepjdode/T- dron species in lower Westphalian A coal swamps from western Europe and also from the Middle Pennsylvanian of the Midcontinent (DiMichele, 1979; Eggert and Phillips, 1982). Nevertheless, the record of Cappasporites in the spore flora is almost always less than the relative abundance of the parent plant in the peat. The microspore output was significantly less (based on cone size and number) than that of Lycospora producers; Cappasporites is typically twice as large in diameter and was commonly dispersed as tetrads. In Europe these microspores have been variously reported as Granisporites or Apiculatisporis (verified by R. Ravn, Amoco Production Company, pers. commun., 1983). Gymnosperms. In contrast to the over-representation of tree ferns and Lyco- spora-bearing lycopods in palynomorph assemblages, the two groups of seed plants are possibly under represented (cordaites) or often not represented at all (Medullosa). The lack of true representation of Schopfipollenites (large prepollen of Medullosa) in the spore flora is not a problem in the Lower and lower Middle Pennsylvanian when Medullosa was not very abundant in coal swamps. However, in younger coals the absence of such data exaggerates the percentage of the other spores, especially Lycospora and tree-fern spores. The study of Winslow (1959) provides substantial information (see Fig. 12) on the abundance of Schopfipollenites from Medullosa in the Illinois Basin Coal Field. As long as Medullosa ranked third or less in the coal swamps and did not significantly change in abundance, this absence of data can be taken into account. However, the importance of Medullosa in coal swamps has to be determined directly from the peat. Where we lack such information on peat deposits, it leaves open the possibility that Medullosa did attain greater im- portance than we presently recognize. It was never dominant on a whole- seam basis at any of the sites we have sampled extensively. The cordaites were the only gymnosperms to dominate any of the Penn- sylvanian coal swamps that we have studied, and this occurred in both limnic 54 (Galtier and Phillips, in press) and paralic coal swamps (Raymond and Phillips, 1983). The disparity in abundance of Florinites, the most common pollen of the cordaites, and the actual importance of cordaites in some coal swamps has been noted (Neves, 1958; Peppers, 1982). Most cordaitean-rich (subdominant) coal-swamp floras register small percentages of pollen be- cause, in part, they are usually coupled with Lycospora-dominated spore floras. The great abundance of cordaites in some lower Desmoinesian coal swamps may not be as widespread as inferred from coal-ball peats. Preserva- tional factors are also very important in relating peat and spore data as are the preparation techniques for spore analysis. RESULTS Information from coal-ball peat and palynomorph studies is presented separately for three stratigraphic intervals, each of which includes one of the major times of vegetational change in the Pennsylvanian. The geographic scope of each interval differs because of the distribution of coal balls and relevant palynomorph data. The intent is to utilize vegetational patterns from the Illinois Basin Coal Field as a basis for interregional comparisons. The three intervals are the following: Early and early Middle Pennsylva- nian, which shows an interregional change in composition of the vegetation at the onset of the "first drier interval"; the middle part of the Middle Penn- sylvanian, during which regional differences in dominance patterns become evident; and, the late Middle to Late Pennsylvanian transition, which in- cludes the "second drier interval" ? the most conspicuous change in coal- swamp vegetation. COMPOSITION OF COAL SWAMPS AND EVIDENCE OF INTERREGIONAL CHANGE IN THE EARLY AND EARLY MIDDLE PENNSYLVANIAN Peat deposits in western Europe ? lower Westphalian A and Westphalian A/B boundary Coal balls of early Westphalian or equivalent age have been found only in Europe. The most abundant of these occur in approximately equivalent seams of the lower Westphalian A and extend from the Lancashire Coal Fields to the Ruhr in what Schopf (1941) called the "Great Coal-ball hori- zon". In the Ruhr and The Netherlands, the coal is the Finefrau-Nebenbank (Kukuk, 1909; Hirmer, 1928; Koopmans, 1928); in Belgium, it is the Boux- harmont (Leclercq, 1925). In the Lancashire Coal Fields coal balls occur in the Union Seam and in its upper split, the Upper Foot Seam, and the equiv- alent in Yorkshire, the Halifax Hard Seam (Stopes and Watson, 1909). Stu- dies of the coal-swamp plants from these peat deposits began over a century ago, and the floras are among the best known in the Upper Carboniferous. Quantitative peat data are presented here from the Union Seam and its upper split at Lancashire, and from the Bouxharmont Seam of Belgium. 55 Many occurrences of coal balls in the slightly younger Katharina Seam of the Ruhr, at the Westphalian A/B boundary, have been reported (Kukuk, 1909). However, only permanent slide collections from the earliest studies (Felix, 1886) were available for quantitative assessments. The Katharina Seiun is important in providing evidence of the onset of changes in vegeta- tion that are also indicated by palynomorph studies in the United States (Peppers, 1979). Early coal-swamp composition and differences in the distribution of peat components Distribution of coal balls in the Union Seam varied, but they were locally scattered throughout the seam profile, which averaged about 1.5 m in thick- ?l , I r- 85 XBANKHALL I ''?~^^"~~-^ROWLEY 1^ 90 95 w V HflPTON VALLEY ' A FUSAIN = 7 57, S/R = 086,069 FUSAIN = 102,103% U/V/O/V SEAM 30- 53''45 N~ YORKSHIRE LANCASHIRE 25- lOO 60- OLD MEADOWS V BACUP TODMORDEN Wk LYCOPODS ES3 PTERID0SPERM5 ED FERNS ^ SPHENOPSIDS El3 COROalTES S/R=SHOOT/ROOT PERCENT j BIOMASS RATIOS OF PEAT COMPOSITION \, -20 r^o- UPPER FOOT SEAM 53''40 N^ from Ordinance Survey Sheets 103 and 109 Blocl t' O t^ rH O^ a t-' 05 X CO t' o O 0 B a o ?a o o c^ CO t^ in N o d t- t- in O O O rH rH rH iH -^ in O rH rH d d to CO o o N in CO CO (N r~ oi o CO CO o in CO CO in iH o 05 05 05 E35 X to CO CO CO ' M (N Oi rH o o o o c^ a 13 ffi +J ffi . _ Z 0 Z o .2 S ^. S s. ^. pj D ? & D 35 _H tn to ?a " Pi OS > oi ^ I g o I o o d ? ?) p! BS a a ^ C C 0 c ?g o o 0 o g c 'S 'S c o D & D S pa ?3 1 3 00 ?J in 2 ^ pg C XJ O* ?-' 0 ? M CO .3 B .S-a fl o c V Id bG ?a 45 a 3 u ?t-l 0 Si ? H 2'i 05 ^ c O' .2 c w pa < !N in w 00 CO in Oi ifl Tj* in lO t- o in in CO CO to Tf o in TJ! T}* iri t- -^ in CO in TJJ 01 r-_ 00 iri t~ iH ^ -^ -^^ S '^ o o ^ '^. w tn O 3 m S P3 3 B 3 35 cn 3 s 3 a g !3 SO) C s s u ? CO 5 59 with 8.1 and 6.0% of the biomass, respectively. Cordaites and pteridosperms, with 2.9% and 2.5% of the biomass, are relatively rare. While the diversity of the Bouxharmont rivals that of the Upper Foot, seed plants account for only 5.4% of the total biomass, and the paucity of Lyginopteris is an obvious quantitative difference. Peat constituents and preservation The shoot/root ratios from the Union and Bouxharmont Seams are 0.69 to 1.16 with a simple average of 0.96 (Table 1). Variation in shoot/root ratios is 0.69?0.86 at Rowley. The highest shoot/root ratio (1.16) occurs at the Hapton Valley site at the line of union; this is the only place where Lepidodendron vasculare was probably not the most abundant lycopod and also where the largest portion of peat was identifiable (unidentifiable mate- rial in the Union Seam ranges from 7.3 to 17.8%). Based on the hypothetical 4/1 shoot/root ratio of a completely preserved lycopod tree, the peat com- position indicates loss of at least 71?83% of the aerial biomass. The peat deposits from the Hapton Valley Mine, which have the highest shoot/root ratios, consist of 36.1% periderm (mostly from aerial plant parts) and 4.3% wood (about equally from stems and roots). The periderm/wood ratio is approximately 8.5/1, a ratio similar to that found in peat deposits of Westphalian age that are highly lycopod-dominated and that have small amounts of cordaites. These durable secondary tissues account for about one-half the total fusain. Peat from the Union and Bouxharmont Seams consists of 5.2?10.3% fusinized material (Table 1). Fusain is highest in the Union Seam (Rowley: 10.2?10.3%), decreasing netir or at the line of union. The lowest fusain occurs in peat from the Upper Foot (Shore-Littleborough: 2.9?6.5%) (Tables 1,2). The amount of fusain has to be considered relative to shoot/root ratios because aerial portions of plants are thought to be the main fusinized com- ponents. Thus, 7.5 to 10.3% fusain is a relatively high amount for a peat de- posit with shoot/root ratios of about one; in the Union Seam even some stig- marian root systems were partially fusinized. Composition of peat deposits in the Katharina Seam Coal-ball distribution in the Katharina Seam of the Ruhr varied consider- ably from site to site (Kukuk, 1909, 1938) but apparently was limited to scattered parts of the upper or lower benches, which were commonly sep- arated by a zone of many thin, closely spaced shale partings. On the basis of the Felix Collection, representing 81 coal balls (2,297 cm^), the major taxa were lycopods (64.9% biomass), with almost equal amounts of cordaites (9.7%), pteridosperms (9.8%), and sphenopsids (10.8%). Small ferns account- ed for 4.8%. The lycopods are the same taxa as those in the Union Seam, but Paralycopodites and its cones are more abundant in the Katharina. Lyginop- 60 teris accounted for 91% of the pteridosperm peat. This genus became extinct just above the Westphalian A/B boundary; the Katharina Seam is the young- est peat source of Lyginopteris. The collection is no doubt biased for aerial portions of plants, and the shoot/root ratio is 2.14. Fusain content is comparably very low at 0.8%. Lycopods and cordaites were the main sources of roots, accounting for 30.2% (total of all roots is 31.8%). Cordaitean assemblages were represented largely by roots, which exhibit eccentric growth rings in the wood as in de- posits from the Upper Foot and Bouxharmont Seams and in some strati- graphically next higher peats in eastern Kentucky. The almost equal abundance of cordaites, seed ferns and calamites (sphe- nopsids) represents the earliest indication, based on coal-ball peats, of a significant change in the vegetation of early Westphalian coal swamps. The concurrent rise of cordaites and calamites as well as the diminution in lyco- pods is consistent with the palynomorph data from coals of comparable age in Europe and the United States, but the changes in the spore floras are much more complex as will be noted in the next section. Peat deposits and spore floras in the eastern United States ? Central Appalachians and Illinois Basin Coal Field Paleobotanical data on composition of Early to early Middle Pennsyl- vanian coal swamps of the United States are restricted currently to the Central Appalachians and the Illinois Basin Coal Field (Fig. 4). Peat deposits occur in eastern Kentucky and Tennessee (Fig. 1) in the upper Pottsville and equivalent coals and, hence, provide the next younger stratigraphic links with data from older coal-ball peats in western Europe. Spore floras provide the stratigraphic basis for comparisons between coals of the Illinois Basin Coal Field, the Central Appalachians and western Europe from as low as the lower Westphalian A up to the Pottsville?Allegheny (Appalachian) and Atokan?Desmoinesian (Illinois Basin) transitions dealt with in the next section. Composition of peat deposits in the Central Appalachians The Upper Path Fork coal bed of the lower Middle Pennsylvanian (ap- proximately equivalent to the Lower Elkhorne) (Fig. 1) is the oldest Penn- sylvanian seam to yield abundant coal balls. The dominant and subdominant plants are lycopods and cordaites, which is a recurrent pattern in those peat data that we have from the upper Pottsville (Table 3). Despite the dearth of data from coal-ball peat (Fig. 5) in the lower Middle Pennsylvanian, the fol- lowing observations can be made: (1) The increased abundance of cordaitean trees noted in the Katharina Seam continues into the lower Middle Pennsylvanian as does the rarity of Medullosa and Psaronius. 61 Fig. 4. Principal Pennsylvanian coal basins in the eastern half of the United States. The Western Interior Coal Region includes the Forest City, Cherokee and Arkoma Basins. The Eastern Interior Coal Region includes the Illinois and Michigan Basins. Modified from Wanless(1969). (2) Dominance of lycopod trees continues with variability, including Lepidophloios harcourtii (Upper Path Fork), Paralycopodites (Hamlin) and Lepidodendron hickii ("High Splint"). (3) Shoot/root ratios are generally even lower than those in western Europe (Union and Bouxharmont Seams) and correlate with the relative abundance of cordaites (the more cordaites, the lower the shoot/root ratio) including the lowest ratio (0.47) and the most cordaites in the Upper Path Fork (Table 3). (4) Fusain levels are high in relation to shoot/root ratios except in the Hamlin coal bed, which has significant variability in shoot/root ratios but low fusain levels. The peat deposits in the lower Middle Pennsylvanian, which had variable depositional environments, would be largely uninterpretable without the spore floras from these and many other coals in eastern Kentucky and Ten- nessee. However, the peat deposits are a reference base of in situ vegetation. In terms of the average swamp moisture and the environments of peat ac- cumulation, the key data from peats are: (1) the record low shoot/root 62 o m a o J y m +* < 0) H Oi o N lO ?* to r-l CO ?J" t- ?0 0 a 0 ai >. o a lO 00 .0 a) ? ?o to* a ^ ttf o -^ "cS t3 tS Qj 0 S B ? JM O M O _ O - Ot CO PI ' (U o aJ ? Oi O P^ 63 PAR&LYCOPODITES 85% SIC ILL ARIA 4% S/R-13 PARALYCOPODITES &B'/^ FUSAIN = 0,8% LEPIDOPHLOIOS W. S/R = 0 87 FUSAIN=2.6% UPPER PATH FORK (Cranks Creek) HAMLIN (Lewis Creek) E3 CORDAITES ESS) FERNS R^ LYCOPODS E^ PTERIDOSPERMS 1^;^ SPHENOPSIDS Fig. 5. Map of a portion of eastern Kentucky (Central Appalachians) with peat constitu- ents determined from random samples of coal balls; coals and localities are indicated. Summary data are given in Table 3. The Shock Branch (1) and Lewis Creek coal-ball deposits (2) were reported as being in the Copland coal bed by Schopf (1961), but ac- cording to those who mapped the Hyden West GQ (Lewis, 1978) and Cutshin GQ (Ping, 1977) they are in the Hamhn coal zone. The coal balls from the Bear Branch coal-ball locality (3) according to Schopf (1961) are probably slightly older than the Copland coal bed (?) and may also be from the Hamlin coal zone. The Crank's Creek coal balls occur in the Upper Path Fork Coal 427 m below the Copland coal bed and about 323 m below the Smith coal zone (the equivalent of the Hamlin coal zone). ratios (and high fusain content) for a lycopod-dominated coal swamp in the Upper Path Fork; this is consistent with Peppers' (1979) "drier interval"; (2) an interval of Para/ycopodjYes-dominated coal swamps, which is rare in the Pennsylvanian; and then (3) a shift to abundance of trees that might be considered generally more indicative of clastic swamps than coal swamps ? true Lepidodendron, a different cordaites (bearing Cardiocarpus magnicel- lularis ovules), and calamites. The Upper Path Fork and "High Splint" coal beds are exceptional because their vegetation combined abundant wood- and periderm-producing trees (cordaites and lycopods). The peat composition of the Upper Path Fork in- cludes 29.6% wood and 20.2% periderm for a ratio of 1.47. Cordaites con- tribute a total of 25.6% wood, 95% of which is derived from roots. Lyco- pods contribute a total of 19.1% periderm, more than 75% of which is de- rived from the aerial portions of the trees. The "High Splint" has one of the highest wood/bark ratios, 2.0 (32%/16%). With the exception of cordaitean- rich coals, Lower and Middle Pennsylvanian peat deposits are composed generally of about one-third or more periderm and relatively little (4?7%) wood. The relative increase of cordaites resulted in one of the most important changes in composition of the peat in the lower Middle Pennsylvanian, if 64 the limited sampling were generally representative of the vegetation. Spore floras are not supportive of such a generalization. Nevertheless, there is some basis to compare the prevalent wood source (cordaitean roots) with changes in climate inferred from other data. Most specimens of Amye/on, the roots of cordaites, in the Lower and lower Middle Pennsylvanian peat deposits exhibit eccentric growth rings except in the Upper Path Fork coal bed. This phenomenon extends stratigraphically up to and including the Hamlin coal. The only intervening coal not in Table 3 from which coal balls have been examined (Hignite coal bed of eastern Kentucky) exhibits the same pattern of growth rings and lycopod-cordaites dominance. Widespread insect burrowing is particularly evident in woods of the Upper Path Fork, Hignite, and Hamlin coal beds (Taylor and Scott, 1983). Comparison between miospore floras in the Illinois Basin Coal Field and the Central Appalachians The correlation of miospore floras in coals in the lower Middle Pennsyl- vanian of the Illinois Basin Coal Field and eastern Tennessee have been re- ported by Phillips and Peppers (1984), and data from eastern Kentucky coals have been added herein to the comparison for the obvious reason that most of the peat data are derived from that area. Despite the potential biases that are involved in using essentially random coal samples (and only one channel sample per coal), there are a number of parallel interregional trends in the abundance patterns of dominant Lycospora (Fig. 6), the intercalated rises in spores of herbaceous lycopods (Fig. 7) and the general increase in abundance and diversity of tree-fern spores (Figs. 7, 8). There are also a number of quantitative interregional differences in the lycopods, sphenopsids and cor- daites. The patterns of change in abundance of lycopod spores are the most diag- nostic of the changes in vegetation. There is a stratigraphic decline in domi- nance of Lycospora pellucida {Lepidophloios harcourtii) at about the time of the Alma and Upper Elkhorn coals in the Appalachians. Lycospora pel- lucida remains subdominant in the Central Appalachians longer than in the Illinois Basin Coal Field (Fig. 6), and L. granulata is much more abundant from the outset in the Illinois Basin Coal Field than in the Appalachians. The maximum abundance of Lycospora micropapillata (Paralycopodites) occurs between the Lower Whitesburg and Copland coals beds of eastern Kentucky; peat deposits in the Hamlin coal bed, which is within this interval, contain abundant Paralycopodites. While the major stratigraphic pattern of change is the shift in dominance of Lycospora pellucida {Lepidophloios harcourtii) to Lycospora granulata {Lepidophloios hallii and/or Lepidoden- dron hickii) across the Lycospora micropapillata zone, we are uncertain of the relative importance of the tree species {Lepidophloios or Lepidoden- dron hickii) high in the Pottsville and on into the Allegheny of the Appalach- ians. Although at relative low levels, Cappasporites (coal-swamp Lepidoden- 65 EASTERN EASTERN ILLINOIS BASIN KENTUCKY TENNESSEE trt III tn (/) -I -I -1 1- HADDIX UNNAMEDCOAL o o UNNAMED COAL SMITH > 7=* X > o S REYNOLDSBURG UNNAMEDCOAL UJ MASON Oat HOOPER MORGAN SPS, u ILLINOIS EASTERN BASIN KENTUCKY W^ Lycospora granulata ^ L pellucido 10 PERCENT ? L.micropapilloto ? ? Lycospora spp. EASTERN TENESSEE ~M 33l." m 32 31 F ?? \m 30 29 28 26 m s^m^im^mm wmt mmmmm\ mm<' ?SWi wim Iran w/zmmm %w//;, m^ Fig. 6. Interregional stratigraphic comparison of the relative abundance of Lycospora species (percent number of miospores in channel samples) in lower Middle Pennsylvanian coals of the Illinois Basin Coal Field, eastern Kentucky and northeastern Tennessee (Cen- tral Appalachians). Data derived, in part, from Phillips and Peppers (1984) with additions from the following coals and sources in eastern Kentucky: 2 = Mason Coal (MAC. 2783), 838 m S. line, 701 m E. line, Varilla Quad., 25-C-72, Bell Co.; 4 = River Gem Coal (MAC. 991 A-B), Saxton Mine, 4.8 km N. of Jellico, just N. of Kentucky?Tennessee line, Whit- ley Co.; 7 = Pond Creek Coal (MAC. 2700), 1524 m S. line, 152 m E. line, Nangatuck Quad., 13-P-86, Martin Co.; 8 = Pond Creek Rider (MAC.2699), 1005 m S. line, 1387 m E, line, 2-L-86, Lick Creek Quad., Pike Co.; 9 = Alma Coal (MAC. 2678 A-D), W. side Tug Fork, directly across from Borderland (West Virginia), Pike Co.; 10 = Upper Elkhorn Coal (MAC. 2712), 37?40'15"N, 82?44'32"W, at junction of Lewis Fork and Brandykeg Creek at Lancer, Lancer Quad., Floyd Co.; 11 = Amburgy Coal (MAC. 2760), 805 m E. of Hyden, Leslie Co.; 21 = Hamlin Coal (MAC. 2669), E. side Hwy. 15, SE corner, Hyden 7.5' Quad., Perry Co.; IS = Little Fireclay Coal (MAC. 2675a) and 22 Copland Coal (MAC. 2675b), Greater Branch along U.S. 460 (Geol. Soc. Ky., 1953, stop 2); 23 = Had- dix Coal (MAC. 2711), 823 m S. line, 1.4 km E. line, ll-G-71, Creekville Quad., Clay Co.; 14 = Lower Whitesburg (MAC. 2666 A-B), 15 = Upper Whitesburg (MAC. 2666 CD), 18 = Hazard No. 4 (MAC. 2666 EG), 19 = No. 4 Rider (MAC. 2666H), 30 = Hazard No. 6 (MAC. 2666 J-M), and 32 = Hazard No. 7 (MAC. 2666 N-P), Messer Branch Section, 1.6 km SW Hazard, 2-1-76, 3-1-76, Hazard South Quad., Perry Co.; 33 = Hazard No. 8 (MAC. 2671 AD), Four Seam Coal Company at head of Buffalo Creek, Perry Co.; 34 = Hazard No. 9 (MAC. 2671 A-C), along Buffalo Creek, below Four Seam Coal Company, Perry Co.; 36 = Princess No. 4 (MAC. 2737 A-C), 1.5 km S. hne, 884 m E. line, 25-X-80, Oldtown Quad., Greenup Co.; and 37 = Princess No. 5 (MAC. 1987 A-E), NW corner intersection Hwy. 60 and 64,8 km SW Ashland, Boyd Co. 66 LYCOPODS ARBORESCENT HERBACEOUS CAPPASPORITES DENSOSPORITES, CRISTATISPORITES, RADIIZONATES. CIRRATRIRADITES TREE FERNS PUNC TA TI5P0RITES PUNCTA TOSPORITES Ml NUT US MINUTUS w Esa pn LAEVIGA TOSPORI TES GLOBOSUS wm [IS t- EH P ? 10 PERCENT Fig. 7. Interregional stratigraphic comparison of the relative abundance of lycopod mio- spores other than Lycospora and tree-fern spores (based on channel samples) in lower Middle Pennsylvanian coals of the Illinois Basin Coal Field, eastern Kentucky and north- eastern Tennessee (Central Appalachians). Numbered plots of data refer to those coals listed in Fig. 6. dron) is generally more abundant in the Illinois Basin Coal Field than in the Appalachians. This may provide a clue to the differences in abundant lyco- pods of the two regions since Lepidophloios tends to occur in assemblages with either coal-swamp Lepidodendron or Lepidodendron hickii but not both. The herbaceous lycopods sporadically exhibited marked increases in abundance, namely near the transition between Lower and Middle Pennsyl- vanian (near the Westphalitm A/B boundary), in the Lycospora micropapil- lata zone and high in the Breathitt (upper Pottsville) Formation and equiva- lent strata. Miospore floras indicate that northeastern Tennessee usually has a greater abundance of cordaites and sphenopsids than the other sampled areas (Fig. 8); this is consistent with the peat of the "High Splint" coal bed. The upward increasing abundance and diversity of tree-fern spore floras does not agree with any peat data we have. Near the top of the Breathitt (uppermost Pottsville) tree-fern spores dominate in several of the Hazard Coals of eastern Kentucky. TOTAL TREE FERNS PUNCTATISPORITES UINUTUS PUNCTATOSPORITES MINUTUS LAEVIGATOSPORITES GLOBOSUS 1 37 36 35 34 35 32 31 30 29 28 7' 27 < 26 T' 25 < 24 > V ?J1 21 -z. 20 ^ 19 17 16 15 14 13 12 UJ 10 ?? _J n ^ 6 5 4 3 2 [ mm I li ? STriWiSiiisssai 'i m'Wmm H EH ii 1 n 10 PERCENT OTHER FERNS GRANULATISPORITES LEIOTRILETES LOPHOTRILETES OTHERS SPHENOPSIDS CALAUOSPORA LAEVIGATOSPORITES I B e 67 CORDAITES FLORINITES s I I Fig. 8. Interregional stratigraphic comparison of the relative abundance of miospores of total Psaronius, other ferns, sphenopsids and cordaites (based on channel samples) in lower Middle Pennsylvanian coals of the Illinois Basin Coal Field, eastern Kentucky and northeastern Tennessee (Central Appalachians). Numbered plots of data refer to those coals listed in Fig. 6. COMPOSITION OF COAL SWAMPS AND EVIDENCE OF INTERREGIONAL CHANGE IN THE LATE MIDDLE PENNSYLVANIAN Lower Desmoinesian spore floras of selected coals in the Western Interior Coal Region The relative distribution of major sporomorphs in 17 Desmoinesian coal seams in Oklahoma were summarized by WOson (1976). We macerated sam- ples from several of the coals in the Western Interior Coal Region, principally in order to compare their spore content with the vegetational composition of the peat. The results obtained from our study (Fig. 9) are difficult to compare with Wilson's (1976) summary for the following reasons: (1) only some of the coals he included were studied for the present report; (2) some of our coal samples are from Kansas and Missouri, rather than from Oklaho- ma; and (3) only a few samples were analyzed from each of the individual coals that contain coal balls. Differences in reporting the data are also a problem since Wilson reported relative abundances of major genera, whereas we have divided some genera into species and have combined others accord- ing to the major plant groups that produced the spores. For example, Wilson 68 showed that Laevigatosporites is the most abundant genus in many of the coals, but we have divided the genus into the tree-fern spores, L. globosus and Punctatosporites (Laevigatosporites) minutus, and the sphenopsid spores, L. desmoinesensis and L. vulgaris. The stratigraphic change in lower Desmoinesian vegetation, based on the sporomorphs of selected coals (Fig. 9), is most extensive near the basalmost coal (approximately equivalent to the Rock Island (No. 1) Coal Member of Illinois). The unnamed coal of Iowa (lower Desmoinesian) is the oldest coal that contains a dominance of Florinites (cordaites). This is associated with a ^ LYCOPODS FERNS ISPHENOPSIDS CORDAITES [ TREES HERBSi TREES OTHER L Y CO SP OR A 5 1 LA EV IG AT O SP O RI TE S GL OB OS US T^ 1 ,? ~"^ S :t , ij 25% COALS i I i PUN CT AT OS PO RI T M IN U TU S G RA NU LA TI SP O R LO PH O TR IL ET ES TR IO UI TR IT ES CA LA M O SP O RA LA R G E LA EV IG AT O SP O RI FL O R IN IT ES ?C,-V\\' ??, ^V ? 1^" ?.-. = IRON POST 1 OKLAHOMA ~^: ;-^:- e - p BEVIER MISSOURI ?, ^\ f 2: 1 CROWEBURG < KANSAS C/) FLEMING KANSAS ,.,, 1 1 2 UJ Q MINERAL KANSAS L Q: UJ 5 TEBO MISSOURI V 1.1 WEIR- PITTSBURG KANSAS SECOR OKLAHOMA .- f UNNAMED IOWA ^ ?"?;"' 1 ??1 Fig. 9. Stratigraphic pattern of percent abundance of major miospore taxa from channel samples of selected coals in the lower to middle Desmoinesian of the Western Interior Coal Region (see Fig. 4 for coal basins). Sources of coal samples with ISGS maceration numbers are stratigraphically the following: Unnamed Iowa coal from Urbandale Mine (MAC. 823) with location in Raymond and Phillips (1983); Secor Coal (MAC. 2713), Blevins, Burdette and Vogt Mine No. 1. SE of Checotal, Oklahoma; Weir-Pittsburgh Coal (MAC. 2691), NE SE NE Sec. 30, SOS, 25E, Crawford Co., Kansas; Tebo Coal (MAC. 635), Hudson-Frazer Mine, Sec. 36, T. 44 N., R. 25 W., Henry Co., Missouri; Mineral Coal (MAC. 2687), Bill's Coal Company, Chetopa Mine, Sec. 28, 34S, 21E, Kansas; Fleming Coal (MAC. 2690), Pittsburg and Midway Hole No. 4, NW SE SE, Sec. 29, 32S, 21E, Cherokee, Kansas; Croweburg Coal (MAC. 2686), Mackie-Clemens Mine No. 22, Sec. 4, 28S, 25E, Kansas; Bevier Coal (MAC. H71-B), strip mine, 11 miles S of Bevier, NE NW Sec. 31, 56N, 14W, Macon Co., Missouri; and Iron Post Coal (MAC. 2684), Peabody Rogers County Mine No. 2, W of Vinita, Rogers Co., Oklahoma. 69 decline in the abundance of Lycospora to the lowest sustained level observed until the Middle-Upper Pennsylvanian transition. The main miospore pattern in the lower Desmoinesian sequence of coal swamps is the concomitant rise and return to dominance of Lycospora-producing lycopods and a decline in cordaites. The tree ferns are quite abundant in the lowermost coals, decline somewhat in the middle part, and begin increasing again toward the Bevier and Iron Post coals, associated with a shift from Laevigatosporites globosus to Thymospora. Lower Desmoinesian peat composition in the Western Interior Coal Region Coal-ball data from seven of the coals in Oklahoma, Kansas, and Iowa sug- gest stratigraphic trends in peat composition that are generally consistent with those in sporomorph floras. Differences in vegetation from the Arkoma Basin and northeastern Oklahoma shelf to the northern part of the Forest City Basin are also apparent from the peat and spore data (Fig. 10), but it should be noted that the oldest coals with coal balls from each of the basins are not stratigraphically equivalent (Fig. 1). Those available data indicate an increasing amount of cordaites from Oklahoma to Iowa and a decreasing amount of lycopods. The coal-ball peats in the oldest Desmoinesian coals of Iowa (Table 4) indicate greater abundance and diversity of cordaites than in Kansas where cordaites were also quite significant. The time-transgressive PARaLVCOPODITES 63% LEPIDOPHLOIOS 9% 5/R = 0 98 FUSA1N = 44% PARALYCOPODITES 73% LEPIDOPHLOIOS 5% LEPIDODENDRON DICENTRICUM 3% S/R = I 24 FUSAIN= 2 6% LEPIDOPHLOIOS LEPIDODENDRON SERRATUM 24 4 % 8 2% S/R= 2 65 FUSAIN= 20 I % PERCENT VOLUME 100 90 80 70 60 50 40 30 20 10 0 VT^ CORDAITES ESna FERNS K3 LYCOPODS ESI PTERIDOSPERMS ^ SPHENOPSIDS S/R= SHOOT/ ROOT PERCENT BIOMASS RATIOS OF PEAT COMPOSITION S/R ^ I I F-USAIN = 26% S/R = 096 FUSAIN = 4 5% -rJ b^ 70 COAL BALLS 2,117 cm^ 110 COAL BALLS 3,744 cm^ 74 COAL BALLS 4,105 cm^ 97 COAL BALLS 5,380 cm2 Q-<> OKLAHOMA KANSAS 99 COAL BALLS 5,446 cm^ -op SECOR COAL WESTHOFF BROS MINES WEST LIBERTY AND ENTERPRISE FLEMING COAL PITTSBURG-MIDWAY MINE NO 19 20 40 60 80 SCALE IN MILES IOWA ROCK ISLAND I NO I) COAL OF ILLINOIS URBANDALE AND SHULER MINES Fig. 10. Comparative transect of coal-swamp vegetation in the lower Desmoinesian of the Western Interior Coal Region showing percent volume (biomass) of plants in peat. The Secor Coal sites are in the Arkoma Basin, the Fleming Coal in the Cherokee Basin and the sites of the Iowa coal, approximately equivalent to the Rock Island (No. 1) Coal Member of Illinois, in the northern part of the Forest City Basin (see Fig. 4). This transect does not represent equivalent coals (see Fig. 1 for stratigraphic positions of each). Summaries of peat constituents are given in Table 4. 70 z fi is ?3 0 o 0 ^ ,^ OT P3 E > ^ n a ^ >! .g c ?1 M 0 w 0 atf i-J S- Tl u ^^ e I-) es 0) 3 ?a 0 H B. S u CO 01 Tf M rH rH r- N 01_ OT 0_ C0_ lO TjT TH w Tjl" M' CO N co' to' iH in OlOtOtD'^OO'N OStDNC^t~t~i-l' UJ CO iD in TtC0rl 1%) to paralic coal swamps in any younger Pennsylvanian coals in Illinois. The change in coal-swamp vegetation at the time of the Springfield Coal may represent a zenith in available freshwater in the Interior Coal Province. It also represents the further development of some of the largest deltaic coal swamps of all geologic time from the Colchester to the Herrin Coal. How- ever, the prevalent trend to larger coal swamps should not overshadow the actual time of the most significant change in the vegetation. This was near the beginning of the Desmoinesian and relates, in part, to marine transgres- sions and regressions and brackish influences in some areas of the swamps. While brackish conditions had an influence on expanding cordaites, we sus- pect that there were other significant environmental changes that allowed tree ferns to actually radiate into coal swamps. These are discussed later. Coal-swamp vegetation and peat constituents in the Illinois Basin Coal Field The Illinois Basin Coal Field has served as a comparative section for asses- sing patterns of change in the vegetations of Pennsylvanian-age coal swamps (Kosanke, 1947, 1950; Peppers and Phillips, 1973; Phillips et al., 1974). The sporomorph plots of average relative abundance are shown for the major plant groups in Fig. 12. The general pattern of lycopod domination through two-thirds of the Pennsylvanian, followed by that of tree ferns is the most basic pattern. There are interruptions of these generalized patterns on a coal-by-coal basis. Diversity in miospore floras and relations to environmental changes Studies of average diversity and turnover of spore species in Illinois Basin coals provided a basis for inference of the "first drier interval" by Peppers (1979). This work has been expanded into an interpretative plot of average species diversity, intended to serve as an indirect measure of the relative magnitude of environmental change. Coal-swamp diversity, as measured by miospores, is significantly greater than measures based on extensive floristic lists from coal-balls because of the nature of sampling (dispersed spores versus in situ peat and greater number of sampling sites for palynology) and because of biases in coals which may receive large spore rains from plants 74 FERNS CORDAITES SPHENOPSIDS PTERIOOSPERMS HERBACEOUS ARBORESCENT MARATTIACEOLS FILICINEAN < SPRINGFIELD IN. HOUCHIN CREEK SURVflNT COLCHESTER(Nt DAWSON SPRINGS ROCK ISLAND (No I WILLiS TARTER REYNOLDSeuRG GENTRY A I Fig. 12. Average relative abundance of major Pennsylvanian-age coal-swamp plants based on spore floras from channel samples in the Illinois Basin Coal Field. Modified and ex- panded from Phillips et al. (1974) with additions from Winslow (1959). outside of the coal swamps ? particularly in small coal swamps. Regardless of potential sampling of non-coal-swamp plants, coal palynology samples indicate regional changes in diversity and can suggest with precision when major environmental changes occurred. Beginning low in the Pennsylvanian, the species diversity is extremely low near the position of the Gentry Coal Member (Fig. 13) and, with the exception of the large diversity in the upper Caseyville, the first rise in diver- sity occurs just below the Smith Coal of western Kentucky and does not drop significantly until the decline between the Murphy sboro and Colchester (No. 2). The large diversity of spores in several upper Caseyville coals may be due to contributions from plants surrounding the local coal swamps (Ravn, 1979; Ravn and Fitzgerald, 1982). The net rise below the Smith Coal appar- 75 ently represents the collective result of the "first drier interval" with some introductions of plant species from outside the coal swamps, some evolution from within, minor extinctions, and continued rain of exotic spores from non-swamp plants. Maximum turnover rate for the Pennsylvanian occurs at about the Rock Island (No. 1) Coal with little net loss of species diversity (Peppers, 1979). Peaks of high species diversity occur just below the Dawson Springs (No. 4) coal bed of western Kentucky and between the Colchester (No. 2) and Springfield (No. 5). Between these high levels in a significant drop in diver- sity between the Murphysboro and Colchester. In terms of loss (extinction and/or paleogeographic migration out of the area) this marks the last evi- dence from coal balls of Pennsylvanioxylon birame and Lepidodendron vas- culare in the Illinois Basin ? the producers of 86% of the biomass. The maximum species diversity was reached in the interval between some of the largest coal swamps (Colchester and Springfield) with declines prior to the Herrin and Danville. The Herrin, which constitutes the largest coal resource in the Illinois Basin Coal Field, had one of the most homogeneous floras that we have observed. The maximum loss of species occurred prior to deposition of peat for the Lake Creek Coal (Fig. 13). This low species diversity is comparable to that occurring below the Smith Coal. The low density in the Lake Creek Coal represents a time of considerable variation in coal-swamp and marsh vegeta- CO AL S R EY N O LD SB r g i 1 1 O z 2 > g ?a NE W HA VE N FR IE N D SV IL z> o ?a TR O W BR ID G " W O O DB UR Y ICJSEYVK^LE^ _ aeeoTT ATOKAN - SPOON 1 CARSONDALE MODESTO j BONO MATTOON FORMATION MORROWAN DESWOINESIAN 1 MiSSOURiAN VIRGILIAN SERIES LGWE^ 1 M DOLE PEN MSYLVANIAh UPPER PENNSYLVANIAN SERIES vVFS"'PHALIflN AWESTPHaLIAN B WESTP HALIA^ C WESTPHALIAN D STEP HANlAN STAGE UPPER CARBONIFEROUS SYSTEM Fig. 13. Graph of species diversity of miospore floras in channel samples of Pennsylvanian- age coals in the Illinois Basin Coal Field with selected coals indicated. Adapted from Peppers (1979) and expanded with additional data. There is apparently a marked masking effect due to the incursion of spores from extra coal-swamp floras in the Morrowan, as in Iowa (Ravn, 1979). The first significant rise in species diversity occurred midway in the lower Middle Pennsylvanian; the maximum turnover rate (Peppers, 1979) occurred at Rock Island Coal time; there was a loss of important (dominant and subdominant) species between the Dawson Springs and Colchester Coals; the maximum level of endemic species is in the lower Carbondale and the maximum loss of species between the Danville and Lake Creek Coals. 76 tion following the extinction of all the previously important lycopod genera and several of the tree-fern species. This interval represents a transition dur- ing which coal-swamp communities were restructured from lycopod-domi- nated to tree-fern dominated ones. While the subsequent Upper Pennsylvanian coal-swamp floras increased again in average species diversity, they never again attained the levels of diversity that existed in the Desmoinesian. Furthermore, the Upper Pennsyl- vanian diversity may be largely a function of the presence of many amphibi- ous plants that had population centers in both coal-swamp and clastic-sub- strate areas. The coal-swamp flora was not as distinct from adjacent floras as it had been in the Desmoinesian. Peat constituents of the coals The shoot/root ratios of coals in the Spoon and Carbondale Formations overlap with but are generally higher than those from older coals (Table 5). While ratios of less than one are found at some sample sites in most of the coals, the shoot/root ratios generally increased stratigraphically with 1.6 as the average for the Murphysboro equivalent, 2.13 for the Summum, 2.29 for the Springfield and then 1.8 for the Herrin. The lowest and highest ratios (0.89 and 3.67) are from the Springfield Coal with the highest from lycopod- tree fern swamps within 0.5 km of the Galatia Channel. Most of the samples of the Springfield were taken from near paleochannels. Variations in ratios in three channel samples along a 10 m face in the Herrin Coal at AM AX Delta Mine of southern Illinois are 1.95, 2.72 and 2.46. A random sample of coal balls from the Herrin Coal at Sahara No. 6 Mine of southern Illinois was intermediate (1.48) between the two profiles (1.17 and 1.93) in the same area. Most of the botanical constituents of the Springfield and Herrin peat deposits are aerial plant material with periderm generally accounting for 33.1?38.7% of the biomass and wood 5.8 to 6.4%; periderm/wood ratios are in the general range of 5/1 to 6.7/1 (see also Phillips et al., 1977). Despite the high shoot/root ratios the fusain content of the coals at most sites is relatively low, being usually in the range of 2.4?6.9%. The exceptions of 8.6?8.9% in the Springfield and 8.1?10.4% in the Herrin (see Harvey and Dillon, 1985) are in close proximity to major paleochannels except for the site in Burning Star Mine No. 4 in Illinois. Regional variation in the vegetation of the Springfield and Herrin Coals Some studies of the Springfield (Eggert et al., 1983) and Herrin Coals (Phillips et al., 1977; Phillips and DiMichele, 1981; Mahaffy, in press; Di- Michele and Phillips, 1985) have emphasized both successional and lateral changes in coal-swamp vegetation, particularly in relation to the major paleo- channel system contemporary with each of these major coal swamps. In the Springfield coal swamp, Lycospora (mostly from Lepidophloios) was greater in abundance near the Galatia paleochannel than elsewhere (Fig. 14). The area north of the plotted paleochannel system that has 40 to 60% Lycospora 77 < CO a E 6 0 s > 0 u 0. c o XI a ?.2 ? 2 W O O O (D ?>^ tO^ irj 01 N l> H O^ 00 TH t^ CO rH O iH ?* M l> CD CO CO M TtT ? '^t -?f ?j' TjT t^ ffi 01 CQ' N ? lO lO' C4 w' C-' ? rH TH H IH H COCOtOOlr-teO^iHr-iOrH 00C000lOO?O01rHtD?O i-H r-l W M TP CO TH COCOOlOtOTjlinOOO'trH ifj TJI ?J to uj to to '^' iO O 00 :Doooocot-ooiow?t-oi CD01010)rHTttO>t-'?t5000 N d O T-i H TH H w' C4 iH rH CO 01 lO i-i ?* O) y? rH CO t- H iH T)I ^' Tf Tf CO* N CO W d ''t t^oOfHTtcoNxr-Tfco d oi ^ CO* ?o CO '^ 00 oi t-* to* OCOC^OS^COiHCDCiXOl iri T)I M '^J* CO rH oi CO CO t^ 00 l^COr-COtDCDt^t~CDI>CO O W 00 01 CD 05 rH q r- O r-t CO* ci CO ?^* CO* ?^* iH H CO M d iHWi-liHrHiHrHi-(iHrHT-< q q lO iH 01 CO rH ^ 00 r-( in ddddddddrndd V V c4 c4,0^ <3i ca ei + NiHOiTltiONHClCOCOT}* cocncowwcQcowwajcQ tH 3 iJ ^ w ?j ct d ""^ iJ CD CO CO CD '^^ Ww S S S S**'^*!' 'P^^ c3 3 3 oi' ?c c S z S a ? -:??" m c e 2 5 ? ? J m g >. c 2 -o o ? S " a| 3 iJ a! S ifi -^ O Ol CO CO IN* oi Tf ^ Tjl (O c^ CO M ? N a> t-; iO Oi Oi X CD H C4 rH d rH r-I f; ?O'00 ?-; Oi q CO lO CO Ti* in* CO 00 q t- CD en o d lC3 CO d d rH t~ I^ Oi CO Oi o oi H ^' in d in l> t^ t^ M t- t- 00 00 rH Oi rH in CO d CO r-i iri CD q rH lO rH ? in oi t-' rH X t^ ^' to oi + t~ OJ'O + o*a rH rH in rH rH rH rH O in t- Oi to CO CO to 0) ?) CO CO to CO to > > > >. aJ >B! > >> > s ? 1 rH ?! g ? Coo ' r h )H !.ao o > 0 ,. .c < g Oi CO 1 C *^ 3 ? -C rH ) 0 ? 3 to cu a< P ..a ? o o a ^ a a" < [fl CO 5 O ci H ss 05 ^__; CJx CO . . Oi M ^ ?a rH Oi 0< ?JJ a, ^ Oi r^ <^ -^ a &4 Q . U J N S 0 3 a to (A ?a 1) a 3 ? o '^ CO CO ?^ S.S s ? ci S *- a ? ??? ?^ -| 0) O (o a'" c a *> Oi 3 a E >. CJ w ?a X 0) O CO iS -H a c S " s5 n (U O: Jin So" S " II ?o .a- o o S 78 is also near portions of the Galatia system not shown in Figure 14 (see Eg- gert et al., 1983). The greater abundance of Lycospora in proximity to the paleochannels is consistent with these as the most continuously wet areas (perhaps long periods of standing water), strong dominance of Lepidophloios, high shoot/root ratios and the thickest coal deposits (Smith and Bengal, 1975). A similar pattern for the Herrin Coal was shown by Phillips and Pep- pers (1984) where the maximum abundance of Lycospora paralleled the Walshville paleocbannel except in the vicinity of a lake or flood basin (Fig. 15) where the maximum abundance shifted to the lacustrine area. There is a significant difference in the levels of abundance of Lycospora in miospore floras in proximity to the paleocbannel systems of the Spring- field (35?60%) and Herrin (65?85%) Coals. Spores of tree ferns are very abundant in proximity to the Galatia paleocbannel system and despite the if Boundary of ?'/ Pennsylvanian System ? Golotia Chonne - Present Li Springfield (No 5) CooTv^ - Percent Abundonce of *^-- Lycospora KENTUCKY Fig. 14. Isocontours of relative percent abundance of Lycospora in channel samples of the Springfield Coal in the Illinois Basin Coal Field showing maximum abundance levels in proximity to the Galatia channel, especially the Indiana portions. High percentages shown north of the plotted paleocbannel system also agree with data compiled by Eggert et al. (1983) that indicate a paleocbannel is in that area. For further explanation see text and for comparison of similar patterns in the Herrin Coal Member see Phillips and Peppers (1984, fig. 11). 79 1 OLD BEN C04L CO. MINE NO 24 2 AMAX DELTA MINE 3 SAHARA COAL CO. MINE N06 4 PEABODY EAGLE SURFACE MINE 5 PEABODY KEN MINE UM FERNS LYCOPODS PTERIDOS PERMS SPHENOPSIDS PERCENT VOLUME T 100 ^ Boundary of Pennsylva ?^? Limit of Hen ^^ Wolsfiuille poieocfionnel ( contemporaneous wilhi Herrin Cool Modified from ^ Smittiand Stoll (1975) ^.. ond Nelson (1983) \, '. I Thick Energy Stiale %^.ff-'\i Locustrine or flood basin Modified from Treworqy and Jocob50n(l979) KENTUCKY Fig. 15. West?east traverse across the Herrin Coal in selected mines in the southern part of the Illinois Basin Coal Field showing similar structural peat composition based on per- cent volume (biomass) of major taxa. The Herrin Coal is exceptional in the relative uni- formity of abundance of major taxa across the heavily mined southern portion, but significant differences in the peat constituents are observed in the sites nearest the Walsh- vill paleochannel and in the extreme eastern portion (see Table 5 for summary data). Plotted data are derived from Phillips et al. (1977), PhiUips and DiMichele (1981), and DiMichele and Phillips (1985). significant fern biomass contribution of 8.9?18.9% to the peat, they are markedly over-represented in the miospore flora. This suggests that the vege- tation growing along the levees could have been an important additional spore source (Eggert et al., 1983). Also, Medullosa occurs in greatest abun- dance at some of the sites (Table 5) and hence biases the estimates based on miospores. In the Springfield Coal there is a broad north?south area of low abundance of Lycospora in eastern Illinois along the north, northwest- ward trending La Salle Anticlinal Belt. The low of about 30% Lycospora along the Illinois?Indiana border is a sample from the AMAX Wabash Mine that is within 0.5 km of the Galatia channel, also the closest site to the chan- nel sampled by a peat profile from coal balls. In this case the dominant lyco- pods were Sigillaria in the lower part of the profile and Lepidophloios in 80 the upper part. There were abundances of other plants, such as Sutcliffia (medullosan seed fern), suggesting representation of what may have been some levee-type vegetation within the profile. In general, estimates of vegetation based on peat and spores are in much closer agreement in the Herrin than in the Springfield. One can encounter al- most as much variation in the peat composition along the coal face in a given mine (compare Sahara No. 6 with AMAX Delta, Table 5) as across most of the Herrin coal swamp (Fig. 15). The Springfield shows much more local variability with tree-fern spores most abundant to the east of the La Salle Anticlinal Belt. Lepidophloios is most abundant in the Herrin Coal at the Old Ben Mine No. 24 (VS3+5 in Table 5) near the Walshville channel (Fig. 15) and the Peabody Ken Mine at the eastern margin of our sampling in western Ken- tucky. These two sites are also among the highest in lycopod biomass, 77.8% and 75%, respectively. The Ken Mine peat deposits were quite high in peri- derm, 38.9%, and lower than usual in wood, 2.0%, with a periderm/wood ratio of almost 20/1. These may have formed in a lacustrine or oxbow depos- it; recent studies (Williamson and Williams, 1984) indicate a paleochannel system existed in that area, but it is not clear whether it was contemporary with the Herrin (Kentucky No. 11) Coal. Coal-swamp vegetation and peat constituents of selected coals in the upper Pottsville and Allegheny Formations in the Appalachians Miospore floras Palynology of coals in the Appalachian Coal Region in the upper part of the Pottsville Formation and in the Allegheny Formation up to the Lower Kittanning Coal (Fig. 16) indicates that lycopods, ferns, and sphenopsids were rather evenly represented in the floras. The two oldest coals shown in Fig. 16, the High Splint (Pennsylvania and eastern Kentucky) and Lower Mercer (West Virginia), contain a relatively small percentage of Lycospora. The large percentage of sphenopsids {Laevigatosporites desmoinesensis and L. vulgaris), ferns {Punctatosporites minutus, and Speciososporites minutus), and herbaceous lycopods (Densosporites) in the High Splint coal suggests more open swamp vegetation possibly with periodic dryness or exposure of the substrate and consequently with higher nutrient levels. Kosanke (1973) correlated the interval from the Princess No. 5B coal to the Princess (No. 9) coal with the interval from the Davis Coal Member to the Chapel (No. 8) Coal Member of Illinois. Palynology of the coals in this interval in the Appalachians is similar to that in the upper part of the Spoon Formation and the Carbondale Formation in Illinois (Peppers, 1970). The sphenopsid Laevigatosporites is greatly reduced in abundance above the Prin- cess 5B coal, and fern spores increase significantly in the Middle and Upper Kittanning coals, the Rider coal, and Lower Freeport coal. 81 SOURCES OF DATA a SCHEMEL(I95T) West Virginia b GRAY (1967) Ohio CK0SANKE(I973) E Kentucky tfPEPPERS Pennsylvonia E Kentucky COALS 257. MAHONING tf UPPER FREEPORT d LOWER FREEPORT o RIDER a UPPER KITTANNING a,b r^lDDLE KITTANNING a-c LOWER KITTANNING o-a P=1INCESS 5B c P^ilNCESS 5A c CLARION a UPPER MERCER a LOWER MERCER a HIGH SPLINT d Fig. 16. Percent abundance of major miospore taxa from channel samples of selected coals in the uppermost Pottsville, Allegheny and lowermost Conemaugh of the Central and Northern Appalachians, based on sources indicated in the figure and the following: High Splint Coal (MAC. 2696 A-C), U.S. Steel Mine 32, 2.4 km W. of Kentucky-Virginia state line on side of Black Mountain, Harlan Co., Kentucky; Upper Freeport Coal (MAC. 2816), Pittsburg Plate Glass Company, Creighton Mine, Allegheny Co., Pennsylvania; and Mahoning Coal (MAC. 2789), W. side of Route 164, NE 1/2, SW 1/4 Sec. 29, Wash- ington Township, Columbiana Co., Ohio. Princess 5A and 5B are names applied in the Princess Reserve District of eastern Kentucky; High Splint is a coal name applied in the Upper Cumberland Reserve District of eastern Kentucky. The two youngest coals, the Upper Freeport and Mahoning coals, are dominated by Lycospora and probably formed in a very wet swamp popu- lated by essentially the lycopod trees Lepidophloios and Lepidodendron. Florinites (cordaites pollen) is quite rare in all the coals represented in Figure 16. The paucity of cordaites in the lower part of the section contrasts sharp- ly with the lower Desmoinesian patterns of the Interior Coal Province. The abundance of tree-fern spores is generally greater in the Allegheny Forma- tion than that in the Spoon and Carbondale of Illinois and exhibits a sus- tained increase up to the two youngest Lycospora-dominated coals, Upper Freeport and Mahoning, the last of which is in the Conemaugh Formation. Composition of peat With the exception of the Tennessee coal referred to as the "High Splint" there are only two coal-ball sources known from the uppermost Pottsville to the top of the Allegheny Formation. Coal balls were reported from a coal referred to as above the Middle Kittanning Coal (Cross, 1967) or the Upper Kittanning (Ferm and Cavaroc, 1969) from western Pennsylvania. A collec- tion of 35 coal balls (1,937 cm^) from the site indicates that the biomass is composed of 45.5% lycopods, 34% seed ferns (mostly Medullosa), 12.5% ferns (mostly Psaronius), and 8% sphenopsids. As in coal of comparable age in the Illinois Basin Coal Field cordaites are rare or absent. The abundance of ferns is comparable to that of Illinois Basin coals, but such an abundance 82 of seed ferns is found only in Illinois Basin coals in patchy sites close to the paleochannel system in the Springfield Coal (see Eggert et al., 1983). The shoot/root ratio is 1.38, and the fusain content of 2.6% is low. The amount of wood present is higher than usual, 10% (mostly calamitean), and periderm composes only 7% of the biomass. The second occurrence of coal balls was reported from the Lower Free- port Coal of eastern Ohio by McCullough (1977, p. 128) who reported: "Lycopod periderm was the only identifiable floral constituent in 36% of the concretions. Twenty percent of the coal balls contained some amount of fusain-like material..." He also provides a table showing percent occurrences in 100 peels with lycopod periderm in 84, stigmarian roots in 4:8,Myeloxy- lon (Medullosa assemblage) in 44 and Psaronius in 24. COMPOSITION OF COAL SWAMPS AND VEGETATIONAL CHANGES IN THE TRANSITION TO LATE PENNSYLVANIAN Coal-swamp vegetation and peat constituents in the Illinois Basin Coal Field Peat deposits occur in six Upper Pennsylvanian coals in the Illinois Basin Coal Field, extending across the Missourian into the base of the Virgilian (Fig. 1, also see Scheihing, 1978). In the Missourian the vegetation was strongly dominated by Psaronius in all the known peat deposits (Table 6, and the Parker Coal Member which is not included in the table), constituting 64.3 to 83.5% biomass (Willard, 1984; Galtier and Phillips, in press). Seed ferns, principally Medullosa, were subdominant accounting for 11.7?24.2% of the biomass. Lycopods, sphenopsids, and cordaites were generally minor contributors to the peat (up to 3.9%) except in particular coals and areas of the swamps. The distribution of other plant groups may well have been patchy in most Missourian coal swamps in south central Illinois; sampling is usually inade- quate to document this. Sigillaria (lycopod trees) accounts for about 10% of the peat in the Bristol Hill Coal Member and in one random sample from the Calhoun Coal Member. In another sample of the Calhoun Coal calamites contributed almost 10% and seed ferns were twice as abundant as in the other Calhoun estimate. In the Friendsville Coal cordaites account for al- most 10% of the peat, a record high for a paralic coal basin in the Upper Pennsylvanian. The Friendsville Coal was sampled in a grid system (Willard, 1984), and almost all the cordaitean peat was localized in one small part of the sample area. All shoot/root ratios of Upper Pennsylvanian peats are near 0.5 (range 0.42 to 0.6) and must be interpreted differently than those of the Lower and Middle Pennsylvanian. The bulk of peat is composed of Psaronius roots, which largely formed the buttressing mantle of the trunks as well as the root systems in the substrate. Conservative estimates that treat 50% of the root material as aerial (Phillips and DiMichele, 1981) would reverse the "shoot"/ MIDCONTINENT NORTHERN APPALACHIAN (too ^ *o *o -to 3Ci iis?_ UPPER LOCUST CREEK MISSOURI -?^Mmii n GRAIN VALLEY MISSOURI "HEPLER"AV?/V5/)5 k. (C *?. (T *|?^ * t ^ ^ 5 uj (0 t~- c 1^ So o^o -J k i IL CHAPEL (No.8) iii n ATHENSVILLE iiia_Ek: LAKE CREEK S5; BRUSH CREEK' Ik nnni ncp 5 S 5 c\5 S5 ?0 lo o> o; g % n^ fi S "0 J ? - p "rip ?1 3NN?3ria38oa r^ K li . d rH d O CO 6 d Bi Hi 3 3 O O O d o 0; v u g >, >, I 3 S i! S X 0 " ?; >,-s j- (0 CQ .g" Q o o ::3 flj . <^ a O o u o o i! ? S c Si It ?a "o c ^a .5 C c O' OJ C ? > o a 0) 0) "T ^ U '-I ?J rt !? U OJ ^ o. a n: g eg S S a 2 2 ? Q Q o 88 root ratios to about 2.0, but we have no anatomical means of distinguishing consistently between subterranean and aerial roots. The principal botanical components change in the transition to typical Upper Pennsylvanian peat deposits, from abundant bark and associated cortical tissues to mostly air- chambered (aerenchyma) roots, which were not only less resistant to decay, but also much more compactible. Peat to coal ratios would be much higher than in the Lower and Middle Pennsylvanian. Resin-rodlet abundance is noted in the coals (Lyons et al., 1982) because of the importance of medul- losan seed ferns in the vegetation. The fusain in Upper Pennsylvanian peat is generally low (1.4?4.1%). The highest fusain level recorded, 14.9%, is in the Bristol Hill Coal (see comparable inertinite percentage in Harvey and Dillon, 1985). Spore assemblages near Desmoinesian?Missourian and WestphalianStephanian boundaries The major floral changes that took place at the Desmoinesian?Missourian and equivalent boundaries are illustrated by histograms of spore assemblages from coals adjacent to the boundaries (Fig. 17). Most of the coal samples from the Midcontinent Region were collected by Heckel and his colleagues (see Schutter and Heckel, 1985) as part of studies on the stratigraphic rela- tionships and paleontology of different lithofacies in the cyclothems ad- jacent to the boundary. The histograms for the Sarre-Lorraine Basin were constructed from range charts showing the relative abundance of major spores in the Westphalian and Stephanian coals (Alpern et al., 1967). Histo- grams for the Donets Basin represent the spore populations of several coals that were diagrammed together within spore zones by Inosova et al. (1975). The general patterns noted in spore assemblages from below to above the Desmoinesian?Missourian and the equivalent Westphalian?Stephanian boun- dary are the following: (1) Lycopod trees, represented by Lycospora, dominate or nearly domi- nate the spore floras of coals immediately below the boundary. Above the boundary Lycospora is generally insignificant or absent. (2) Coal-swamp Lepidodendron, represented by Cappasporites (except as reported in Europe), also disappears at the boundary. (3) Marattiaceous tree ferns, represented by T/zymospora, decrease progres- sively in abundance toward the Westphalian?Stephanian boundary, while those represented by Punctatosporites minutus increase in abundance. (4) However, across the boundary into the Stephanian both Thymospora pseudothiessenii and Punctatosporites minutus nearly disappear. (5) Marattiaceous spores dominated most of the paralic coal-swamp floras in North America and Europe above the boundary, but there is a distinct change in the taxa that predominate on each side of the boundary;Punctati- sporites minutus, Punctatosporites granifer, and Cyclogranisporites rise to predominance in the Stephanian. 89 (6) Chaloneria, an herbaceous lycopod, represented by Endosporites, became abundant in coal swamps and marshlands of the Stephanian. PERSPECTIVES AND INTERPRETATIONS OF INTERREGIONAL CHANGES IN VEGETATION Perspectives Impact of fluctuating environments Pennsylvanian (Late Carboniferous) tropical coal-swamp environments reflect regional paleogeography. The fluctuating distribution of epicontinen- tal seas, freshwater wetlands, lowlands of slight relief and the Allegheny (Hercynian) orogeny created the milieu for coal-swamp development. The extensive water cover along the southern half of the paleocontinent of Laurussia markedly influenced atmospheric moisture levels at a time when the paleoclimate also was being altered by the pole to pole formation of Pangaea, by a long-term net emergence of land, and by the orographic belts of Laurussia and Gondwana (Rowley et al., 1985). Oscillations in Pennsyl- vanian climate and tectonics created a dynamic mosaic of emergent land and wetlands that expanded and contracted. This caused the extinctions of many plant species and allowed establishment in swamps of many others. ''Holding capacity " between coal-swamp expansions Swamplands, lakes, rivers and other wet habitats had "holding capacities" that affected the nature of sequential contraction and expansion phases of coal and clastic swamps. The nature of "holding environments" in which shrinking coal-swamp populations may have survived between times of major coal-swamp occurrences reflects both paleoclimate and basinal geology. It is from such areas, should coal swamps have become temporally or regionally discontinuous, that the competitive colonization of new coal-swamp areas would have occurred. Thus, vegetational analysis of a sequence of coals in a region indicate some of the selective environmental factors of the "holding areas" ? namely "holding capacity". One extreme example is the "holding capacity" between coal swamps at the Middle?Late Pennsylvanian transition in the United States, which was negligible and marked by mass extinction. At the other extreme are the coal swamps represented by seams in the Car- bondale Formation in the Illinois Basin Coal Field and equivalent coals in the Western Interior Coal Region. In this case, a potentially large "holding capacity" (including perhaps the overlapping of coal swamps through time) was part of the ecological fabric that allowed swamp floras to reach their maximum diversity. The quantitative composition of coal-swamp vegetation changed signif- icantly in some "holding areas" that existed between regional expansions of coal swamps. The fluctuations in the ecology of a single coal swamp during peat accumulation also have a bearing on the colonization and composition of subsequent coal-swamps. 90 Edaphic control and high threshold responses Coal swamps were buffered against all but the most severe climatic (most- ly moisture) changes because of the water-holding capacity of a peat sub- strate. The low pH, low availability of mineral nutrients, flooded conditions, and low oxygen levels of peat substrates, at least during part of a season, created edaphic conditions which excluded all but highly specialized plant species. As major environmental changes occurred, differential effects on coal-swamp species followed, suggesting distinct and different thresholds to ecological disruption. Coal swamps were not only environments with unusual edaphic condi- tions, but they were affected by repeated fluctuations in water availability and significant disturbances, such as incursions of brackish water, fires and clastic influx. The dominant trees, such as lycopods or ferns, were highly adapted to disturbances. They grew quickly, required minimal nutrients, were composed mostly of non-woody tissues, had determinant growth or relatively small stature, generally large reproductive allocations, and repro- ductive bodies that were dispersed rapidly and extensively by wind and water. On a peat substrate the dominant trees exhibit relatively high threshold responses to regional environmental changes and consequently provide an average guide to the most severe changes in climatic factors. When major floristic changes occurred they are found between successive expansions of coal-swamp sequences from one seam to the next, and these are the ex- ception rather than the rule. Interregional differences in coal-swamp vegetation There are differences as well as similarities in coal-swamp vegetation among regions within the United States and between the United States and Europe. These differences are even greater between the Euramerican coal belt and its Cathaysian extension in the Stephanian and Permian. However, within the conterminous Euramerican area the timing of vegetational change is approximately synchronous in the major coal regions. Regional differences in tectonic setting, historical aspects of the vegetation, and local environmen- tal differences may make the patterns of vegetational change somewhat different in each coal region, but these do not mask the synchroneity of the timing of change, which appears to be truly interregional. Midcontinent and Appalachian regions Rather subtle differences are detectable in spore floras of the Illinois Basin Coal Field and the Central Appalachians. New spore species appear at the beginning of the "first drier interval" in several coals lower in the Illinois Basin Coal Field than in the Appalachians (Phillips and Peppers, 1984, p. 213). At higher stratigraphic levels (near the Paralycopodites zone), new species appear in the Appalachian region several coals lower than in the 91 Illinois Basin. This shift in timing is consistent with an initial west to east drying trend (or increased seasonability) followed by a return to more moist conditions (or reduced seasonability) from the opposite direction. Euramerican coal regions The drastic changes in coal-swamp vegetation are transitional across the Middle?Upper Pennsylvanian boundary and are very similar in each of the major coal regions of the United States and Europe. However, an important difference between the vegetation on either side of the Appalachians is the continued presence of some Lycospora-bearing lycopods in Europe. This is possibly the result of a more asymmetrical change in the drying of the cli- mate there (Hedeman and Teichmiiller, 1971). There were apparently gradual transitions between Westphalian and Stephanian floras in the southern parts of Europe (the so-called Cantabrian Stage as in Spain; Wagner, 1966). In both the Sarre-Lorraine (Alpern et al., 1967) and the Donets Basins (Inosova et al., 1975) the spore genus Lycospora continues to be found in coals of the Stephanian and Kasimovian but at markedly reduced abundance. In the Massif Central of France Lycospora-producing lycopods survived throughout the Stephanian in the limnic basins, detected as small percent- ages of the spore floras (Liabeuf et al., 1967). Occasional specimens of Lyco- spora have been observed in Virgilian coals in the Illinois Basin Coal Field (Peppers, 1985) and in the Appalachian Basin (Clendening, 1974). Lepidodendron is reported from compression floras of the Stephanian of Spain (Lorenzo, 1979) and Ulodendron occurs in compression floras from the Kasimovian of the Donets Basin (Stschegolev, 1975). These provide megafossil evidence of the survival of lycopods in Europe that are missing in the Upper Pennsylvanian rocks of the United States. The particular kind of lepidodendroid lycopods that survived are not known with certainty, al- though that described by Lorenzo (1979) appears to be most like a coal- swamp Lepidodendron, the kind lacking infra-foliar parichnos. So far as we can ascertain the lepidodendrons did not extend into the Permian of Europe. Cathay sian paleofloristic province Based on compression floras vegetational changes that occurred in the Euramerican floristic province at the Westphalian?Stephanian boundary coincided with differentiation of the paleotropical belt into western Eurame- rican and eastern Cathaysian floristic provinces (Li and Yao, 1982). How- ever, in contrast to the patterns in the United States and Europe, lycopod trees thrived in coal swamps and clastic environments of deposition (partic- ularly seat earths and some roof shales but not in mineral-rich partings of coals) in the Stephanian and Early Permian of North China and in those of the Late Permian of South China. The lycopods include species of Lepidoden- dron as well as Sigillaria. Instead of a diminution in Lycospora at the West- phalian?Stephanian (Middle?Upper Carboniferous) boundary in North China (Shanxi Province), Lycospora attained its zenith in the coal swamps of 92 the Stephanian, along with abundant Marattialean spores (Ouyang and Li, 1980). Other Euramerican trees found in Chinese coal-swamps include ca- lamites and cordaites. Although quantitative peat data are not yet available from the Upper Carboniferous coal balls in the Xi Shan coal field of Shanxi Province, those from the uppermost Permian of southwestern China (Guiz- hou Province) are composed of Lepidodendron-like trees, Psaronius and cordaites in that order of importance. Much of the same kind of tropical coal-swamp vegetation that characterized the Euramerican coal belt of West- phalian time continued to flourish to the end of the Permian in Cathaysia. Stratigraphic patterns and environmental inferences ? major plant groups Most of the paralic coal swamps of the Early and Middle Pennsylvanian of the eastern half of the United States were dominated on a "whole seam basis" by lycopod trees (Fig. 18). The dominant genera were Lepidophloios, coal-swamp Lepidodendron, true Lepidodendron, and Paralycopodites, all of which became extinct during the Middle?Late Pennsylvanian transition in the United States. Other important tree genera were Mesoxylon, Pennsyl- vanioxylon, Psaronius and Medullosa. These genera formed most of the peat biomass in the only coal swamps of the Midcontinent not dominated by lycopods during the Desmoinesian (in Iowa). All were subdominant to lyCO- PERCENT VOLUME OF PEAT 90 70 50 30 10 0 INTERCALATED POLYSPORIA MARSHLANDS EXTINCTION OF LEPIDOPHLOIOS LEPIDODENDRON PAPAL YCOPODITES -?-ONSET OF MOST EXTENSIVE SWAMPS MIDCONTINENT CORDAITES I?DIVERGENCE OF MAJOR LYCOPODS EURAMERICAN CORDAITES SHIFT IN LYCOSPORA SPECIES Q|-<-INTR0DUCTI0N OF TREE FERNS ' SOME MARSHLAND VEGETATION ?i-EXTINCTION OF LYBINOPTERIS Fig. 18. Generalized stratigraphic patterns of abundance of Pennsylvanian coal-swamp vegetation in the United States with a relative wetness curve indicating the three most important times and kinds of changes. The vegetational diagram is based on distribution of spores and plants as well as other data available before 1980; the relative wetness curve is based on abundance of identified bituminous coal resources compiled by Phillips et al. (1980) plotted on a semi-log scale. Adopted and modified from Phillips and Peppers (1984, fig. 8). 93 pods in varied coal swamps during the Middle Pennsylvanian, and all the genera continued into the Late Pennsylvanian. Lycopods All of the lycopod tree genera w^ere present in the earliest Westphalian coal swamps of Europe and the United States, but only Sigillaria and herba- ceous lycopods such as Selaginella, Sporangiostrobus and Chaloneria survived into the Late Pennsylvanian in the United States. Despite the grossly similar architecture of lycopod trees, they exhibited significant differences in growth and ecological adaptations, especially as related to reproductive biology (Phillips, 1979; DiMichele and Phillips, 1985), resulting in segregation within coal swamps. Hence, they indicate different intraswamp conditions. Factors controlling lycopod distribution appear to be complex, but one of the major elements was the water regime of the swamp. On a relatively wet to drier scale within the swamps, Lepidophloios probably dominated in the wettest areas with standing water. The next wettest habitats were occupied by lepidodendrons but appear to have been modified by disturbances to a considerable extent. Some species of coal-swamp Lepidodendron were toler- ant of slightly brackish influences. True Lepidodendron trees also inhabited moderately wet but variable habitats and were dominant in some wet en- vironments where the nutrient level was perhaps higher than usual for coal- swamps. True Lepidodendron and Sigillaria were common in wet-clastic en- vironments of deposition during the Early and Middle Pennsylvanian; this is in marked contrast to the coal-swamp centered Lepidophloios and coal- swamp Lepidodendron. Species of Paralycopodites occurred in transitional environments from clastic to coal-swamp conditions or within coal-swamp sequences following disruptions and degradation of peat (lower water tables). Consequently, Paralycopodites was seldom dominant on a "whole-seam basis". Sigillaria was usually the least important lycopod tree in coal swamps except in the Upper Pennsylvanian. Early and Middle Pennsylvanian coal swamps were developed mostly in the Lepidodendron-Lepidophloios habitat ranges. In some swamps, especial- ly those of massive peat accumulation, the greatest overlaps occurred among lycopods. However, Lepidophloios seems mostly to co-occur, on a whole- seam basis, with either coal-swamp Lepidodendron or with true Lepidoden- dron, but not with both. The relative importance of lycopod-tree genera changed at the Lower?Middle Pennsylvanian and during several intervals within the Middle Pennsylvanian. With each of these changes there were accompanying increases in the importance of one or more of the other major groups of trees. Lycopod communities of the Lower Pennsylvanian were mixed; Lepido- dendron vasculare was generally most abundant, followed by Lepidophloios harcourtii and Lepidodendron hickii, with Paralycopodites and Sigillaria still less abundant. In the vegetational transition spanning the Lower?Middle Pennsylvanian boundary (near the Westphalian A?B boundary) Lepidophloi- 94 OS harcourtii, cordaites, calamites, and herbaceous lycopods increased in abundance. A transitional stage, consisting of several coals in the lower Middle Pennsylvanian in which Paralycopodites was dominant, might be regarded as indicative of the end of the "first drier interval". Other plant groups increased in abundance more steadily, and dominance shifted from Lepidophloios harcourtii to trees that produced Lycospora granulata micro- spores. During these changes in lycopod dominance the first major marine trans- gression occurred in the Illinois Basin Coal Field; this is represented by the Lead Creek Limestone Member in the upper Atokan of Indiana and western Kentucky. Subsequently, other marine transgressions occurred from the west into the Midcontinent during the early Desmoinesian. The abundance of lycopods dropped to the lowest level in the Middle Pennsylvanian in the Western Interior Coal Region. There was an influence of brackish conditions mostly in the Western Interior Coal Region but extending across the Illinois Basin Coal Field. As trees generally intolerant of brackish influence, most of the lycopods were displaced to freshwater habitats, with major exceptions among the coal-swamp lepidodendrons, especially Lepidodendron vasculare, which is inferred to be somewhat tolerant of such conditions. It was during and after this disruptive change in Midcontinent coal swamps that other species of coal-swamp Lepidodendron appear in abundance; L. vasculare subsequently disappears from the coal swamps. In the upper Pottsville-Allegheny transition in the Appalachian Region, at about the same interval as the onset of brackish influence in the Midcon- tinent, Lepidodendron hickii appeared as a major element in those coal- swamp floras we have studied. The data are based on few peat deposits. Coal-ball deposits from the "High Splint" coal in Tennessee are dominated by Lepidodendron hickii. The only Allegheny coal balls that are quantified are from the coal above the Middle Kittanning coal of western Pennsylvania. They also exhibit a dominance of L. hickii. Lepidophloios also occurs in these peat deposits from Tennessee and Pennsylvania. We have insufficient coal-ball data to make comparisons with changes in vegetation indicated by spore floras. The spore floras indicate an increased abundance of Lycospora pellucida (L. harcourtii) from the Lower Mercer to the Clarion coal interval followed by a rise in abundance of Lycospora granulata. Cordaites The cordaites were the only gymnosperms known to achieve dominance or sole subdominance in any coal swamps (whole seam basis) in the Middle Pennsylvanian of the eastern half of the United States. In paralic coal basins the rise and decline of cordaites was entirely within the Middle Pennsylvani- an. The two main kinds of assemblage are Mesoxylon with Mitrospermum ovules and Pennsylvanioxylon with Cardiocarpus ovules (Costanza, 1983). In the lower Middle Pennsylvanian Mesoxylon rose in importance from the beginning of the "first drier" interval until the middle Middle Pennsylvanian. 95 The abundance of roots of Mesoxylon, penetrating peats dominated by Lepidophloios harcourtii, a wet indicator, suggest that the two kinds of trees did not grow together and were adapted to different intra-swamp environ- ments. The Mesoxylon organ assemblage (one natural species) existed from the earliest Westphalian. There is no evidence of aerating tissues in the root systems (Cridland, 1964) as in all the other coal-swamp trees. Instead, the larger roots (Amyelon) usually exhibit eccentric growth rings, apparently disruptions of root growth not evident in the trunk and branches. Such "growth rings" may be responses to seasonal or sporadic flooding for pro- longed periods, a phenomenon somewhat similar to the formation of eccen- tric growth in modern cypress knees. The abundance of Mesoxylon without aerating systems apparently indicates tolerance of periodic flooding in coal- swamp environments at a time when the other main group of gymnosperms (seed ferns) and amphibious tree ferns were not yet important in coal swamps. It is interesting to note, in general, that cordaites and, to a lesser extent, calamites filled significant habitat gaps in the early Middle Pennsyl- vanian until Psaronius and Medullosa expanded as subdominant plants. Only in what we infer to be brackish-influenced coal swamps of the early Desmoinesian did cordaites co-occur in peat deposits with abundant seed ferns and tree ferns. Pennsylvanioxylon was the most important genus, par- ticularly P. birame with Cardiocarpus spinatus ovules (Costanza, 1983). The extreme cordaitean dominance and diversity occurred in the northern part of the Forest City Basin at the onset of the Desmoinesian. Cordaites were less abundant in younger coals of the Illinois Basin Coal Field and the Arkoma Basin than in the Forest City Basin. The stratigraphic importance of cor- daites, as a whole, ended almost synchronously across the Midcontinent at about Iron Post and Springfield Coal time. In contrast to the lack of aerating tissues in the older Mesoxylon assemblages, the roots of cordaites in brackish- influenced coal swamps exhibited aerenchyma (Costanza, 1983). The brackish tolerance of early Desmoinesian cordaites is founded on several lines of deduction. However, the mangrove-like habitat inferred for one particular species, Pennsylvanioxylon birame, should not be applied to all early Desmoinesian cordaites. The distribution of forest types in such swamps is thought to be fundamentally different from other coal swamps, but is not, as yet, resolved (Raymond, 1983; Raymond and Phillips, 1983; Raymond et al., 1984). Psaronius tree ferns The patterns of change suggested by the various estimates of relative abun- dance of Psaronius are the most dramatic and puzzling of any of the tropical trees. The sensitivity of miospore floras provides an interesting record of Psaronius in the Euramerican coal belt long before these plants can be docu- mented by other means as an actually significant component of clastic-sub- strate or coal-swamp communities during the middle Middle Pennsylvanian. There appears to be some disparity between the earlier tree-fern spore record 96 and the peat deposits up to near the beginning of Desmoinesian time; how- ever, peat and spore data are in general agreement in the HamUn Coal, which exhibited relatively few tree ferns. While such a spore record could be dis- missed as extra-swamp in origin, there is also an apparent disparity with the generalized stratigraphic pattern of compression floras ? the logical extra- swamp source for over-representation of tree-fern spores. Pfefferkorn and Thomson (1982) report an increase in tree-fern foliage in compression floras that could parallel that of tree-fern expansion in coal swamps; however, they lump their assemblages into such major stratigraphic intervals that compari- sons are uncertain. Such extra-swamp abundance could add to the over- representation of tree-fern spores in some Desmoinesian and equivalent coal swamps. The only modern analogs for the environmental tolerances of Psaronius are the living Marattiales, which hardly compare with the tree sizes of Psaro- nius. The Marattiales occur in the moist tropics as understory plants along the lower slopes of mountains up to about 2000 m in tropical and montane rainforests, especially on islands (Troll, 1970). Only one species (Marattia fraxinea) is not entirely tropical, occurring in the temperate but highly mari- time parts of eastern Australia and on the North Island of New Zealand. Other modern tree ferns of the Filicales overlap the Marattiales environ- mentally and are much more widespread. They occur in mild temperate low- lands and wet tropical mountains; these tree ferns seem to be more tolerant of cool temperatures than the Marattiales. Moisture and altitude seem to be the limiting factors in their tropical or near tropical distribution (Troll, 1970). The possibility exists that Psaronius might have been tolerant of slightly cooler temperatures, as exhibited by modem Filicalean tree ferns, and could have inhabited lower altitudes of mountains in the tropical belt; if not, then an alternative may be more temperate latitudes of Euramerica or other land masses. Such speculative considerations are perhaps appropriate in the case of Psaronius considering the conflict between miospore data and those from the peat and the compression-fossil records of the lowlands. Spore abun- dance that started to rise during the "first drier interval" may reflect "spore rain" from remote regions prior to actual expansion across the paleotropical lowland belt. Tree ferns did not dominate any of the compression floras prior to the upper Westphalian D (upper Desmoinesian). However, such ferns dominated 11 of 14 floras, being first or second in abundance in upper West- phalian D and younger floras (Pfefferkorn and Thomson, 1982). We agree with the general interpretation of Pfefferkorn and Thomson (1982) that the response of the floras preserved in clastic rocks may be more rapid than those of the coal swamps. However, the change in tree-fern abun- dance in both clastic and peat-accumulating environments is one of rising im- portance during the wettest interval, the time of the largest coal swamps, and evidently, the most extensive shifts in habitable space. Once Psaronius be- came established in paleotropical habitats, its enormous capability for swift 97 dispersal and fast growth led to rapid integration into existing vegetation especially in disturbed habitats. The shift to tree-fern dominance just above the Middle?Upper Pennsyl- vanian boundary is the most drastic permanent change in coal-swamp vegeta- tion following the lycopod extinctions. This change is much more complex than the generalized data plots indicate. Several major species of Psaronius also suffered extinction or near extinction. Lycopods and several other kinds of plants dominated the intervening coal-forming environments between the abrupt decline in Lepidophloios, Lepidodendron and Paralycopodites and the clear establishment of Psaronius dominance. Psaronius generally domi- nated coal swamps thereafter and extended into the Permian as a major tree in peat-accumulating environments of limnic (Galtier and Phillips, in press) and paralic coal basins of the paleotropical belt. Lyginopteris and Medullosa seed ferns The stratigraphic patterns of change in distribution of seed ferns consists of two parts, one for the liana-like to shrubby Lyginopteris and one of. Me- dullosa trees. In the Early Pennsylvanian the larger seed ferns were mainly represented by a robust vine, Lyginopteris. It was cosmopolitan like many other seed ferns in clastic and peat-accumulating environments. Lyginopte- ris became extinct just above the Lower?Middle Pennsylvanian boundary in the United States and just above the Westphalian A?B boundary in Europe (Patteisky, 1957; Gillespie and Pfefferkom, 1979). This extinction at the on- set of the "first drier interval" is most conspicuous and coincides with the extinction of the spore Schulzospora. Medullosa was larger than Lyginopteris but still small in relation to lyco- pod trees. Medullosa was uncommon in early Westphalian peat deposits, and there is a decided lag between its rise in importance and the earlier ex- tinction of Lyginopteris (Phillips, 1980). Medullosa did not substantially increase in abundance until near the middle of the Middle Pennsylvanian at abovat the same time that Psaronius became abundant. While medullosans were generally diverse and dominant kinds of trees in moist lowlands of the Pennsylvanian (Pfefferkorn and Thomson, 1982; Peppers and Pfefferkorn, 1970), species of Medullosa apparently did not expand into coal swamps during the "first drier" interval, as did cordaitean gymnosperms. Medullosa's most characteristic distributional pattern is one of fluctuating and increasing abundance upward to the upper Middle Pennsylvanian, followed by a more sustained level of abundance thereafter. Medullosa exhibited more of a patchy distribution than the other main types of trees. In general, Medullosa parallels the abundance patterns of Psaronius usually at lower and more inconsistent levels. Abundance of Medullosa suggests higher nutrient levels in the peat; it occurs near mineral-rich bands within the coal, and is associ- ated often with high fusain levels to which Medullosa was a major contribu- tor. 98 Interregional vegetationalpatterns in relation to environment While this study deals principally with one suite of environmental condi- tions, those of coal swamps, the Pennsylvanian Period was an age of wetlands in which the patterns of vegetational changes in coal-swamps were hardly separable from those of other vegetation of the tropical belt. With the onset of the Pennsylvanian, the changing paleogeography of drylands and fresh- water wetlands seems to have reinforced the evolutionary differences in reproductive biology and growth of trees that originated in the Mississippian or earlier. Gymnosperms were the best adapted to dispersal and slow growth in the dryland habitats, while lower vascular plants, especially the hetero- sporous lycopods and the calamitean trees, dominated wetlands. The greatest overlap of habitats was in the clastic-rich wetlands. The combination of a dynamic tropical climate in concert with markedly changing habitats greatly influenced the adaptive patterns that emerged among the five basic kinds of trees. Trees were the most important elements of most plant communities and every major group exhibited arborescent forms. The limits imposed by growth-rate and habit, physiology, and repro- ductive biology were such that at the beginning of Pennsylvanian time: (1) lycopods totally dominated the coal swamps and represented most of the arborescent diversity; (2) other lycopods and calamites dominated the clastic swamps; (3) pteridosperms characterized the moist lowlands; and, (4) cor- daites were apparently most abundant in the drier lowlands where conifers evolved during the Early Pennsylvanian (Peppers and Pfefferkorn, 1970; Oshurkova, 1977; Scott and Chaloner, 1983). The transitional environments that changed between drylands, clastic-rich wetlands and coal swamps ap- parently constituted recurring habitat space that any of the main kinds of trees could occupy, depending on migration rates and duration of the habi- tat. Preemption of habitat space was important, but we have no compelling evidence that communities were largely structured by competition between major kinds of trees (lycopods, sphenopsids, cordaites), which had markedly different adaptations. "First drier interval" The beginning of vegetational change near the Early?Middle Pennsylvanian transition (Peppers, 1979) was detected by changes in relative abundances of many kinds of spores and plants. The principal coal-swamp changes in- cluded: (1) the rapid expansion of Lepidophloios harcourtii (a wet indicator) and a significant decrease in the importance of Lepidodendron vasculare; (2) concurrent fluctuating abundance of herbaceous lycopods (Denso- spore marshlands) within coal swamps, and the appearance of Chaloneria (marshlands); (3) increase in abundance of small cordaitean trees to subdominant levels, and increase in calamites; (4) extinction of Lyginopteris; and 99 (5) the "rain" of tree-fern spores, increasing stratigraphically in diversity and quantity but lacking confirmation in peat deposits. The overall pattern of vegetational change suggests a fluctuation in fresh- water regime that resulted in oscillating wet and drier conditions within in- dividual coal swamps and between sequential coal swamps (seams). This pattern is consistent with erratic monsoonal circulation and varied seasonal availability of freshwater. If the "first drier interval" can be explained as marked fluctuation of wetness due to monsoonal circulation, then the most likely paleogeographic change is the rise of a transequatorial orographic barrier, the ancestral Appalachian highlands (Rowley et al., 1985). Floristic patterns of change that can be interpreted as migration outside the coal swamps are provided by the compression-floral zonation of Read and Mamay (1964), which recognizes Zone 7 (Zone of common occurrence of Megalopteris spp.) as occurring concurrently with the beginning of the "first drier interval". The Megalopteris flora, which includes cordaites and apparently some of the earliest Pecopteris foliage from tree ferns, is con- sidered a drier-type flora. The observations of Read and Mamay (1964), Cross (1977), Leary (1981), Leary and Pfefferkorn (1977) suggest the pos- sibility of a "Megalopteris flora" migration eastward to, but not beyond the transequatorial Appalachians, while the Lonchopteris flora existed on the European side (Read and Mamay, 1964; Dix, 1934). This would be con- sistent with a reduction in moisture from the east. Cross (1977) provides another explanation of Megalopteris distribution. End of the "first drier intervaV and onset of wetter environments The paleobotany of coal swamps between the waning of the "first drier interval" and the middle Middle Pennsylvanian is beset with the incongruities of marine transgressions from the east prior to those from the west. Changes in the paleogeography that took place during this time are the initial shift of major coal swamps from the southern half of the central Appalachians to the northern half and expansion of major coal-forming areas in the Midcon- tinent. If the interregional stratigraphic correlations (Fig. 1) are approxi- mately correct in the middle part of the Middle Pennsylvanian, the differ- ences in patterns of coal-swamp vegetation in the Appalachians and Midcon- tinent are at least slightly asynchronous. The lack of close synchrony is per- haps compatible with the key environmental controls, which differed from east to west. If a modest mountain range, the transequatorial Appalachians, were a paleoclimatic factor in the tendency for monsoonal circulation in the "first drier interval", the gradual return of wetter conditions toward the middle of the Middle Pennsylvanian and progressively into the Desmoinesian of the Midcontinent would require either a lowering of such a barrier or the further elevation of such mountains to form what Rowley et al. (1985) term a "high- altitude heat source". The latter model would be compatible with increased rainfall in the Appalachians as a developing interior mountain range within 100 Pangaea. As Rowley et al. (1985) pointed out, such a high-altitude heat source could theoretically offset some easterly flow (as an orographic barrier) and also the tendency for north?south monsoonal circulation by shifting some flow from the west (Rowley et al., 1985). If substantial moisture from transgressive epicontinental seas existed in the west, during Desmoinesian time, abundant rainfall could have been delivered by westerly flow into the Appalachians. Rainfall would have been augmented by the Michigan River System in the Midcontinent. The "high altitude heat-source" model of Rowley et al. (1985), although highly speculative, would be consistent with a major Appalachian orogeny and barrier, increased erosion and sedimentation rates, and sustained water supply to the largest coal swamps. In the Midcontinent the beginning of major marine transgressions no doubt played an important role in the ame- lioration of a more moist paleocHmate as did the freshwater damming effect on large deltaic plains resulting in extensive coal-swamp development. It also is plausible that, given an adequate freshwater regime, subsidence patterns would effect the stratigraphic pattern of coal abundance ? thus, making it an environmental variable independent of paleoclimate. Such subsidence- tectonic factors could create stratigraphic patterns difficult to separate from paleoclimatic wetness. Most of the Midcontinent was topographically unsuitable for significant coal deposits before the Desmoinesian; exceptions are noted in the Harts- horne coal beds of the Arkoma Basin and the Bell coal bed of western Ken- tucky and Lower and Upper Block Coal Members of western Indiana in the Illinois Basin Coal Field. As sedimentary infilling and deltaic platforms in- creased in size there was a concomitant change in coal-swamp vegetation suggesting an increase in freshwater availability in the Midcontinent. This offset the earlier brackish influences that began with the Mariah Hill Coal bed in Indiana and western Kentucky and, a little later, the Rock Island Coal and equivalent coals. "Wetter" swamps probably reflect a combination of more water influx and markedly changed paleogeography due to basin infilling. This progressively wetter regime is well documented regardless of its proximate cause, by the rising importance of Lepidophloios, which is indicative of the wettest conditions, in both the Western Interior Coal Region and the Illinois Basin Coal Field. The level of wetness, possibly the least seasonal climate, seems to have peaked in the Midcontinent at about the time of the Springfield and Herrin Coals and equivalents, which are charac- terized by Lepidophloios, coal-swamp lepidodendrons and Psaronius general- ly in that order of importance. Onset of the driest interval The threshold level of late Middle Pennsylvanian coal-swamp trees was not significantly affected until conditions reached a catastrophic level near the Middle?Late Pennsylvanian boundary. It is the mass extinction of coal- swamp vegetation, particularly some lycopod genera, that originally recalled 101 the attention of paleobotanists to the paleoclimatic importance of the Des- moinesian?Missourian (Westphalian?Stephanian) boundary (PhiUips et al., 1974). This does not constitute the lowest moisture level for the early Late Pennsylvanian, but it does suggest that at some point the "holding capacity" of wetlands reached the lowest for the Pennsylvanian up to that time. This low "holding capacity" was of sufficient duration as to cause the inter- regional extinction of many coal-swamp centered trees. The truncation of stratigraphic ranges is so apparently geologically abrupt as to elicit concern over unconformities or disconformities in which the actual stratigraphic succession of change has been lost (Tschudy, 1969). While this is a possibil- ity, the subsequent pattern of coal-swamp vegetational change, a series of coal by coal changes in dominance by trees or herbs other than the typical Middle Pennsylvanian lycopods or the Upper Pennsylvanian tree ferns, strongly suggests a mosaic transition rather than a major break in the strati- graphic succession. As with compression floras in the "first drier interval", there is a marked west to east migrational shift in drier adapted gymnosperm floras during the "second drier interval". In the far west conifers begin appearing in the late Desmoinesian (Read and Mamay, 1964) while cordaites and conifers appear in the middle to late Missourian deposits of Kansas (Elias, 1936; Winston, 1983: Mapes, 1984) and subsequently migrate into the major coal regions of the United States. The floral migrations of the late part of the "second drier interval" resemble some Early Permian distributional patterns when the paleotropical belt underwent further drying. While the return of conditions more favorable for peat preservation did occur at the beginning of Monongahela and equivalent time, there is little information on the actual coal-swamp vegetation except that Psaronius continued to be dominant, probably with subdominant Medullosa. There may have been increased provinciality (Phillips, 1980). In the United States, for example, Medullosa, Arthropitys (calamites) and Psaronius were locally abundant in that order in small random samples of peat from the Newcastle coal of Texas (Phillips, 1980) near the Pennsylvanian-Permian boundary. In the Autunian of France Psaronius and Arthropitys were the main biomass contributors to the peat (Galtier and Phillips, in press). The occurrence of a few significant coal deposits from the Midcontinent to the Dunkard Basin, beginning with the Pittsburgh coal is evident in the occurrence of the Nodaway and Elmo coals of Iowa and Kansas and coals in a borehole in western Kentucky, which are stratigraphically higher than any previously known from the Illinois Basin Coal Field (Hower et al., 1983; Phillips and Peppers, 1984). It should be noted that Hower et al. (1983 p. 1477) have reported abundant Thymospora thiessenii (a species abundant in the Pittsburgh coal) from the two upper coals (Virgilian) as well as a few percent of T. pseudothiesseni, which was not previously considered to ex- tend into the Virgilian in the Illinois Basin Coal Field. 102 Pennsylvanian bituminous coal resources and coal-swamp vegetation The stratigraphic distribution of identified bituminous coal resources of the Appalachian Coal Region and the Interior Coal Province were suggested (Phillips et al., 1980; Phillips and Peppers, 1984) as an approximate guide to relative wetness during the Pennsylvanian Period in the eastern United States; this implies that other environmental conditions were also conducive to peat accumulation and burial. The data were compiled in order to test relative wetness as suggested by the dominant trees (Fig. 19). The stratigraphic basis for comparison began at approximately the middle of the Lower Penn- sylvanian (lower Westphalian A equivalent) and extended into the basal Virgilian of the Interior Coal Province. Paleoclimatic factors, perhaps most importantly the availability of fresh- water or the balance from rainfall, runoff and evapotranspiration, probably had a major impact on the structure of coal-swamp communities. What we infer as relative wetness may, in fact, be increased or decreased seasonality LEPIDODENDRON PERCENT ABUNDANCE VEGETATION LEPIDODENDRON S L.EPIDOPHLO:CS CORDfliTES PSARONIUS MEDULLOSA ' ^ MIOCONTINENT SERIES ____^ 10 -? *>?r *-T?^- '^--- '"^^?-T-^ ?......,.-?^ ILLINOIS FORMATIONS CASEYVILLE| ABBOTT SPOON |CARBONDALE| MODESTO | BOND . _ ^ '. zy. Z ..~~.^ .. 1.-^.,.. ri^^ii,^inr-r-i n .. u i ccni in i A ?, MORROWAN lATOKflNJ DESMOINESj_AN_ tl-l^^ I NTERIOR COAL PROVINCE COAL RESOURCE: - APPALACHIAN COAL REGION PQCAHONTAS NEW RIVER L-GHLNY* CONEMAUGH|M ^L'iJDUNK NAMURIAN B-C WESTPHALIAN B IILT PROPOSED STRA"0')-^^J _PEN"NSYL~y_A_Nlf. N _ WESTERK FUROPEArj 3E Fig. 19. Comparative stratigraphic distribution of groups of Pennsylvanian plants and bituminous coal resources in the Interior Coal Province and the Appalachian Coal Region. Plots of identified bituminous coal resources (Phillips et al., 1980) are overlain with generalized patterns of the percent abundance of major groups of coal-swamp plants for the Interior Coal Province to show relative changes in relation to paleoclimate. The per- cent abundance of vegetation begins at the 10 percent level. 103 of moisture availability. Our data do not allow this level of resolution. Other factors also may have been as important as climate. Thus our perception of relative or seasonal wetness has to be considered cautiously as a rough guide. While it is consistent with numerous other lines of reasoning, such data are not exclusively a measure of paleoclimate. They reflect the interaction of paleoclimate and the causal and modifying factors of paleotectonics and its effect on eustatic changes of sea levels, paleogeography, basinal geologies and, hence, rates of subsidence and sediment accumulation. Thus, the abun- dance of coal resources provides a basis for determining approximately when major changes ensued in the paleoenvironment defined by these interacting factors. In general, the coal resource patterns suggest the following (Fig. 19); (1) A "first drier interval", perhaps due to increased fluctuation or sea- sonality of available moisture, and a stratigraphic decline in coal resources in the lower Middle Pennsylvanian in the Appalachian Region; (2) An increasingly wetter (less seasonal) interval of maximum coal re- sources in the upper Middle Pennsylvanian possibly modified by rates of subsidence that allowed thick coals to accumulate; (3) A second and more severe drier interval and a marked decline in coal resources to a minimum in the lower Upper Pennsylvanian. The most compelling evidence for a major paleoclimatic role in these patterns is the synchronous (or nearly so) times of change and the similar resulting vegetational patterns despite the interregional difference in floras and basinal tectonics. The general decline in lycopod dominance during the beginning of the "first drier interval" occurs at the same time as the decline in the coal re- sources of the Appalachians. Shoot/root ratios also decline to a minimum at this time in the Upper Path Fork coal, and fusain content rises, becoming relatively high for such a low shoot/root ratio. The decline of lycopods and concomitant rise of cordaites in the Interior Coal Province is significantly different from the pattern seen in the Appalachians. The ensuing vegeta- tional change back to lycopod dominance with rising abundance of tree ferns and seed ferns is accompanied by progressively higher shoot/root ratios and moderate fusain levels. These changes suggest wetter coal swamps more favorable for peat accumulation and preservation. The beginning of the "second drier interval" brings together pronounced changes in vegetation, particularly extinction, and a shift to a tree-fern dom- inance. These clearly are tied to loss of "holding capacity". The actual set of factors causing such a severe vegetational change is probably much more complex than simply a "drier climate" although that is the net result. A major regression (see Schutter and Heckel, 1985) near the Middle?Upper Pennsylvanian transition, a subsequent major marine transgression (Ames Limestone and equivalents) (Vail et al., 1977; Crowell, 1978; Busch and Rollins, 1984), decreased rainfall and runoff, and changes in erosion, sedi- mentation and subsidence rates may have combined to the point of causing a minimum of coal resources in the upper Conemaugh and equivalents. 104 ACKNOWLEDGMENTS For generous help in field work, stratigraphic correlations, access to infor- mation and sharing specimens and data we thank Matthew J. Avcin and Robert L. Ravn, formerly of the Iowa Geological Survey; Manfred Barthel, Humboldt-Universitat zu Berlin, D.D.R.; Lawrence L. Brady, Kansas Geolog- ical Survey; George Botoman and Horace R. Collins, Ohio Geological Survey; Russell A. Brant and Donald Chesnut, Kentucky Geological Survey; Aureal T. Cross, Michigan State University; Philip J. DeMaris, Richard D. Harvey, and W. John Nelson, Illinois State Geological Survey; Muriel Fairon-Demoret, University of Liege, Belgium; Robert Gastaldo, Auburn University; Donald L. Eggert, Indiana Geological Survey; John C. Ferm, Robert W. Hook, and James C. Hower, University of Kentucky; Samuel A. Friedman, Oklahoma Geological Survey; Jean Galtier and John Holmes, University of Montpellier, France; Philip H. Heckel, University of Iowa; Robert M. Kosanke, U.S. Geol- ogical Survey, Denver; Robert E. McLaughlin, University of Tennessee; Hermann W. Pfefferkorn, Universitat Heidelberg, West Germany; Anne Ray- mond, Texas A and M University; Jeffery Schabilion, University of Iowa; John Shepard, Shell Oil Company; Cedric Shute, British Museum of Natural History, London; Natasha S. Snigirevskaya, Komarov Botanical Institute, Leningrad; Tian Baolin and Wang Jie, China Institute of Mining, Xuzhou, P.R.C.; W.D. Ian Rolfe, Hunterian Museum, Glasgow; Zhang Shan-zhen, Nanjing Institute of Geology and Palaeontology, P.R.C. Field and labora- tory assistance was provided by Suzanne Costanza, Alicia Lesnikowska, Debra Willard and Richard Winston, University of Illinois and by James F. Mahaffy, Dordt College. We wish to acknowledge the assistance of Debbie Gains, Computer Sec- tion, Illinois State Geological Survey; Anne Patterson, Paleobotany Labora- tory, Department of Plant Biology, University of Illinois; and Carol Kubitz, Scientific Illustrator, School of Life Sciences, University of Illinois. This cooperative research project was supported in part by NSF Grant EAR 83-13094 entitled, "The Quantitative Analysis of Middle and Upper Pennsylvanian Coal-Swamp Vegetation in Relation to Coal in the United States and China (Stephanian)", to T.L. Phillips, University of Illinois, by the Illinois State Geological Survey, and by NSF Grant DEB 82-10475 en- titled "Evolutionary Patterns in Pennsylvanian Lepidodendrid Lycopods from Compression-Impression Paleoenvironments", to W.A. DiMichele. REFERENCES Alpern, B., Lachkar, G. and Liabeuf, J.J., 1967. Le Bassin Houiller Lorrain peut-il fournir un stratotype pour le Westphalien superieur? Rev. Palaeobot. Palynol., 5; 75?91. Bradley, W.H,, 1956. Use of series subdivisions of the Mississippian and Pennsylvanian Systems in reports by members of the U.S. Geological Survey. Bull. Am. Assoc. Pet. Geol., 40: 2284?2285. 105 Busch, R.M. and Rollins, H.B., 1984. Correlation of Carboniferous strata using a hierarchy of transgressive-regressive units. Geology, 12: 471?474. Clendening, J.A., 1974. Palynological evidence for a Pennsylvanian age assignment of the Dunkard group in the Appalachian Basin ? Part II. W. Va. Geol. Econ. Surv., Coal- Geol. Bull., 3, 107 pp. Costanza, S., 1983. Morphology and systematics of cordaites of Pennsylvanian Coal Sviramps in Euramerica. Ph.D. thesis. University of Illinois at Urbana, Illinois, 186 pp. (unpubl.). Cridland, A.A., 1964. Amyelon in American coal balls. Palaeontology, 7 : 186?209. Cross, A.T., 1967. A coal-ball flora from Pennsylvania. Am. J. Hot., 54: 652. Cross, A.T., 1977. Stratigraphic significance of Megalopteris (Pennsylvanian) in North America. Geol. Soc. Am., North-Central Section, Abstr. Programs, p. 587. Crowell, J.C., 1978. Gondwanan glaciation, cyclothems, continental positioning, and climate change. Am. J. Sci., 278: 1345?1372. Davies, D., 1929. Correlation and paleontology of the coal measures in East Glamorgan- shire. Philos. Trans. R. Soc, 2173: 91?154. DiMichele, W.A., 1979. Arborescent lycopods of Pennsylvanian age coal swamps: mor- phology, evolution, and paleoecology. Ph.D. thesis. University of Illinois at Urbana- Champaign, 252 pp. DiMichele, W.A., 1983. Lepidodendron hickii and generic delimitation in Carboniferous lepidodendrid lycopods. Syst. Bot., 8: 317?333. DiMichele, W.A. and Phillips, T.L., 1977. Monocyclic Psaromus from the lower Pennsyl- vanian of the Illinois Basin. Can. J. Bot., 55: 2514?2524. DiMichele, W.A. and Phillips, T.L., 1985. Arborescent lycopod reproduction and paleo- ecology in a coal-swamp environment of late Middle Pennsylvanian age (Herrin Coal, Illinois). Rev. Palaebot. Palynol., 44: 1?26. Dix, E., 1934. The sequence of floras in the Upper Carboniferous with special reference to South Wales. Trans. R. Soc. Edinb., 57: 789?838. Eggert, D.L. and Phillips, T.L., 1982. Environments of deposition ? Coal balls, cuticular shale and gray-shale floras in Fountain and Parke Counties, Indiana. Indiana Geol. Surv., Spec. Rep. 30, 43 pp. Eggert, D.L., Chou, C-L., Maples, C.G., Peppers, R.A., Phillips, T.L. and Rexroad, C.B., 1983. Origin and economic geology of the Springfield Coal Member in the Illinois Basin (Field Trip 9). In: R.H. Shaver and J.A. Sunderman (Editors), Field Trips in Midwestern Geology. Indiana University, Bloomington, 2: 121?146. Elias, M.K., 1936. Late Paleozoic plants of the Midcontinent region as indicators of time and of environmental. Int. Geol. Congress, 5th Washington, 1933, 1: 691?700. Felix, J., 1886. Untersuchungen Uber den inneren Bau westfalischer Carbon-Pflanzen. Abh. Geol. Specialkarte Preuss. Thuring. Staaten, 7 : 153?225. Ferm, J.C. and Cavaroc, Jr., V.V., 1969. A field guide to Allegheny deltaic deposits in the upper Ohio Valley with a commentary on deltaic aspects of Carboniferous rocks in the northern Appalachian plateau. Ohio Geol. Soc. Pittsburgh Geol. Soc, 21 pp. Feys, R., 1964. The paleopedology of coal basins. In: A.E.M. Nairn (Editor), Problems in Paleoclimatology. Interscience Publications, London, 66?75. Galtier, J. and Phillips, T.L., in press. Swamp vegetation from Grand Croix (Stephanian) and Autun (Autunian), France, and comparisons with coal-ball peats of the Illinois Basin. In: J.T. Dutro, Jr., and H.W. Pfefferkorn (Editors), Paleontology, Paleoecology, Paleogeography. Ninth Int. Congr. Carboniferous Stratigr. Geol. (1979, Urbana, Illi- nois), C.R. 5. Gillespie, W.H. and Pfefferkorn, H.W., 1979. Distribution of commonly occurring plant megafossils in the proposed Pennsylvanian System stratotype. In: K.J. Englund, H.H. Arndt and T.W. Henry (Editors), Proposed Pennsylvanian System Stratotype, Virginia and West Virginia. Am. Geol. Inst., Selected Guidebook Series No. 1, Falls Church, Va., pp. 87?94. 106 Gray, L.R., 1967. Palynology of four Allegheny Coals, northern Appalachian Coal Field. Palaeontographica, 121B: 55?86. Harvey, R.D. and Dillion, J.W., 1985. Maceral distributions in Illinois coals and their paleoenvironmental implications. In: T.L. Phillips and C.B. Cecil (Editors), Paleo- climatic Controls on Coal Resources of the Pennsylvanian System of North America. Int. J. Coal Geol., 5: 141?165. Hedeman, H.-A. and Teichmiiller, R., 1971. Le developpement paleogeographique du Carbonifere superieur. Fortschr. Geol. Rheinld. Westfalen, 19: 134?147. Hirmer, M., 1928. Uber Vorkommen und Verbreitung der Dolomitknollen und deren Flora, ler Congr. Stratigr. Carbonifere, Heerlen, 1927, C.R. 289?312. Hower, J.C, Fiene, F.L., Wild, G.D. and Helfrich, C.T., 1983. Coal metamorphism in the upper portion of the Pennsylvanian Sturgis Formation in western Kentucky. Geol. Soc. Am. Bull., 94: 1475?1481. Huddle, J.W. and Patterson, S.H., 1961. Origin of Pennsylvanian underclay and related seat rocks. Bull. Am. Assoc. Pet. Geol., 72: 1043?1660. Inosova, K.I., Kruzina, A.Kh. and Shvartsman, E.G., 1975. Fig. 6, The scheme of distribu- tion of characteristic genera and species of microspores and pollen in Carboniferous and Lower Permian deposits. Field excursion guidebook for the Donets Basin, 8th Int. Congr. Carboniferous Stratigraphy and Geology, Moscow, 1975, p. 287. Koopmans, R.G., 1928. Researches on the flora of the coal balls from The "Finefrau- Nebenbank" Horizon in the Province of Limburg (The Netherlands). N.V. Boek, En Handelsdrukkerij v/h Jos. Penners, Heerlen, 33 pp. Kosanke, R.M., 1947. Plant microfossils in correlation of coal beds. J. Geol., 55: 280? 284. Kosanke, R.M., 1950. Pennsylvanian spores of Illinois and their use in correlation. 111. State Geol. Surv., Bull. 74, 128 pp. Kosanke, R.M., 1973. Palynological studies of the coals of the Princess Reserve District in northeastern Kentucky. U.S. Geol. Surv., Prof. Pap. 839, 22 pp. Kukuk, P., 1909. Uber Torfdolomite in den Flozen der niederrheinische-westfalischen Steinkohlenablagerung. Gliickauf, 45: 1137?1150. Kukuk, P., 1938. Geologic des Niederrheinisch-Westfalischen Stein Kohlengebietes. Julius Springer, Berlin, 706 pp. Leary, R.L., 1981. Early Pennsylvanian geology and paleobotany of the Rock Island County, Illinois area, Part 1: Geology. 111. State Museum, Rep. Invest. 37, 88 pp. Leary, R.L. and Pfefferkorn, H.W., 1977. An Early Pennsylvanian flora with Megalopteris and Noeggerathiales from west-central Illinois. 111. State Geol. Surv., Circ. 500, 77 pp. Leclercq, S., 1925. Introduction a I'etude anatomique des vegetaux houillers de Belgique: Les "coal balls" de la couche Bouxharmont des charbonnages de Werister. Ann. Soc. Geol. Belg., Mem., Collect., 4; 1?79. Leclercq, S., 1930. Etude d'une coupe verticale dans une couche a "coal balls" du Houil- ler de Liege. Ann. Soc. Geol. Belg., Bull., 54: B63?B67. Lewis, R.O., 1978. Geologic map of the Hyden West Quadrangle, Leslie and Perry coun- ties, Kentucky. U.S. Geol. Surv., Map GQ-1468. Li, Xingxue, and Yao, Zhaoqi, 1982. A review of recent research on the Cathaysia flora in Asia. Am. J. Bot., 69: 479?486. Liabeuf, J.J., Doubinger, J. and Alpern, B., 1967. Caracteres palynologiques des charbons du Stephanian de quelques gisements fran^ais. Rev. Micropaleontol., 10: 3?14. Lomax, J., 1915. The formation of coal-seams in the light of recent microscopic investiga- tions, Part 1. Trans. Inst. Min. Egn., 50: 1?32. Lorenzo, P., 1979. Les sporophylles de Lepidodendron dissitum Sauveur, 1848. Geobios, 12: 137?143. Lyons, P.C, Finkelman, R.B., Thompson, C.L., Brown, F.W. and Hatcher, P.G., 1982. Properties, origin, and nomenclature of rodlets of the inertinite maceral group in coals of the central Appalachian Basin, U.S.A. Int. J. Coal Geol., 1: 313?346. 107 Mahaffy, J.F., in press. Profile patterns of coal and peat palynology in the Herrin (No. 6) Coal Member, Carbondale Formation, Middle Pennsylvanian of Southern Illinois. In: J.T. Dutro, Jr. and H.W. Pfefferkorn (Editors), Paleontology, Paleoecology. Paleogeog- raphy. 9th Int. Congr. Carboniferous Stratigraphy Geol., C.R. 5. Mapes, G.' 1984. Paleozoic conifers in North America. Contributed Papers, Second International Organization of Paleobotany Conference, Edmonton, p. 24. McCullough, L.A., 1977. Early diagenetic calcareous coal balls and roof shale concretions from the Pennsylvanian (Allegheny Series), Ohio. Am. J. Sci., 77: 125?134. McLaughlin, R.E. and Reaugh, A.B., 1976. Pennsylvanian coal-ball petrifactions in the Cumberland Plateau section. East Tennessee. Geol. Soc. Am., Abstr. Programs, 2: 228-229. Nelson, W.J., 1983. Geologic disturbances in Illinois coal seams. 111. State Geol. Surv., Circ. 530, 47 pp. Neves, R., 1958. Upper Carboniferous plant spore assemblages from the Gastrioceras sub- crenatum Horizon, North Staffordshire. Geological Magazine, 95: 1?19. Oshurkova, M.V., 1977. The principles and methods of facies-paleoecological studies of continental deposits containing plant remains. Paleontol. J., 2: 233?241 (Translation of Paleonol. Zh., 1977, 2: 105-114). Ouyang Shu and Li Zaiping, 1980. Upper Carboniferous spores from Shuo Xian, northern Shanxi. Proc. Int. Palynol, Conf., 5th, 16 pp. Patteisky, K., 1957. Die phylogenetische Entwicklung der Arten von Lyginopteris und ihre Bedeutung fiir die Stratigraphie. Mitt. Westfal. Berggewerkschaftskasse, Bochum, 12: 59?83. Peppers, R.A., 1970. Correlation and palynology of coals in the Carbondale and Spoon Formations (Pennsylvanian) of the northeastern part of the Illinois Basin. 111. State Geol. Surv., Bull. 93, 173 pp. Peppers, R.A., 1979. Development of coal-forming floras during the early part of the Pennsylvanian in the Illinois Basin. In: J.E. Palmer and R.R. Dutcher (Editors), Deposi- tional and Structural History of the Pennsylvanian System of the Illinois Basin. Guide- book to Field Trip 9, 9th Int. Congr. Carboniferous Stratigraphy and Geology, Part 2, Invited Papers. 111. State Geol. Surv., Urbana, 111., pp. 8?14. Peppers, R.A., 1982. Palynology of the unnamed Coal in the Staunton Formation, Maple Grove Mine. In: D.L. Eggert and T.L. Phillips (Editors), Environments of Deposition ? Coal Balls, Cuticular Shale and Gray-shale Floras in Fountain and Parks Counties, Indiana. Indiana Geol. Surv., Spec. Rep., 30: 27?33. Peppers, R.A., 1985. Comparison of miospore assemblages in the Pennsylvanian System of the Illinois Basin with those in the Upper Carboniferous of western Europe. 9th Int. Congr. Carboniferous Stratigr. Geol., C.R., 2: 483?502. Peppers, R.A. and Pfefferkorn, H.W., 1970. A comparison of the floras of the Colchester (No. 2) Coal and Francis Creek Shale. In: W.H. Smith, R.B. Nance, M.E. Hopkins, R.G. Johnson and C.W. Shabica (Editors), Depositional Environments in Parts of the Carbondale Formation ? Western and Northern Illinois. Field Trip, Coal Geology Division, Geol. Soc. Am., Milwaukee, 1970, 111. State Geol., Surv., Guideb. Ser. 8, pp. 61-74. Peppers, R.A. and Phillips, T.L., 1973. Pennsylvanian coal-swamp floras in the Illinois Basin. Abstr. Programs, 1972 Annu. Meet. Geol. Soc. Am., 4: 624?625. Pfefferkorn, H.W. and Thomson, M.C., 1982. Changes in dominance patterns in Upper Carboniferous plant-fossil assemblages. Geology, 10: 641?644. Phillips, T.L., 1979. Reproduction of heterosporous arborescent lycopods in the Missis- sipian?Pennsylvanian of Euramerica. Rev. Palaeobot. Palynol., 27: 239?289. Phillips, T.L., 1980. Stratigraphie and geographic occurrences of permineralized coal- swamp plants ? Upper Carboniferous of North America and Europe. In: D.L. Dilcher and T.N. Taylor (Editors), Biostratigraphy of Fossil Plants. Dowden, Hutchinson and Ross, Stroudsburg, Pa., pp. 25?92. 108 Phillips, T.L. and Chesnut, D.R., 1980. Coal balls in lower Middle Pennsylvanian strata of eastern Kentucky; occurrences and peat composition. Geol. Soc. Am., Abstr. Programs, 12, p. 498. Phillips, T.L. and DiMichele, W.A., 1981. Paleoecology of Middle Pennsylvanian age coal swamps in southern Illinois ? Herrin Coal Member at Sahara mine No. 6. In: K.J. Niklas (Editor), Paleobotany, Paleoecology and Evolution. Praeger, New York, N.Y., 1: 231?284. Phillips, T.L. and Peppers, R.A., 1984. Changing patterns of Pennsylvanian coal-swamp vegetation and implications of climatic control on coal occurrence. Int. J. Coal Geol., 3: 205?255. Phillips, T.L., Peppers, R.A., Avcin, M.J. and Laughnan, P.F., 1974. Fossil plants and coal: Patterns of change in Pennsylvanian coal swamps of the Illinois Basin. Science, 184: 1367?1369. Phillips, T.L., Avcin, M.J. and Berggren, D., 1976. Fossil peat from the Illinois Basin. A guide to the study of coal balls of Pennsylvanian age. 111. State Geol. Surv., Ed. Series 11,39 pp. Phillips, T.L., Kunz, A.B. and Mickish, D.J., 1977. Paleobotany of permineralized peat (coal balls) from the Herrin (No. 6) Coal Member of the Ilhnois Basin. In: P.N. Given and A.D. Cohen (Editors), Interdisciplinary Studies of Peat and Coal Origins. Geol. Soc. Am., Microform Publ., 7, pp. 18?49. Phillips, T.L., Shepard, J.L. and DeMaris, P.J., 1980. Stratigraphic patterns of Pennsyl- vanian coal swamps in Midcontinent and Appalachian Regions of the United States. Geol. Soc. Am., Abstr. Programs, 12: 498?499. Ping, R.G., 1977. Geologic map of the Cutshin Quadrangle, Leslie County, Kentucky. U.S. Geol. Surv., Map GQ-1424. Ravn, R.L., 1979. An introduction to the stratigraphic palynology of the Cherokee Group (Pennsylvanian) coals of Iowa. Iowa Geol. Surv., Tech. Pap. 6, 177 pp. Ravn, R.L. and Fitzgerald, D.J., 1982. A Morrowan (Upper Carboniferous) miospore flora from eastern Iowa, U.S.A. Palaeontographica, 183B: 108?172. Raymond, A., 1983, Peat taphonomy of recent mangrove peats and Upper Carboniferous coal-ball peats. Ph.D. thesis, University of Chicago, Chicago, 111., 275 pp. (unpubl.). Raymond, A., and Phillips, T.L., 1983. Evidence for an Upper Carboniferous mangrove community. In: H.J. Teas (Editor), Biology and Ecology of Mangroves. Tasks for Vegetation Science, 8: 19?30. Raymond, A., Phillips, T.L. and DiMichele, W.A., 1984. Middle Pennsylvanian swamp communities. Geol. Soc. Am., Southeastern and North-Central Section, Abstr. Pro- grams, 16: 188. Read, C.B. and Mamay, S.H., 1964. Upper Paleozoic floral zones and floral provinces of the United States. U.S. Geol. Surv., Prof. Pap. 454-k, 35 pp. Reaugh, A.B. and McLaughlin, R.E., 1977. Report on coal balls from Cross Mountain, Anderson County, Tennessee. Bot. Soc. Am. Misc. Series, Pub. 154, p. 41. Rowley, D.B., Raymond, A., Parrish, J.T., Lottes, A.L., Scotese, C.R. and Ziegler, A.M., 1985. Carboniferous paleogeographic, phytogeographic and paleoclimatic reconstruc- tions. In: T.L. Phillips and C.B. Cecil (Editors), Paleoclimatic Controls on Coal Re- sources of the Pennsylvanian System of North America. Int. J. Coal Geol., 5: 17?42. Scheihing, M.H., 1978. A paleoenvironmental analysis of the Shumway Cyclothem (Vir- gilian), Effingham County, Illinois. M.S. Thesis, University of Illinois, Urbana-Cham- paign, 184 pp. Schemel, M.P., 1957. Small spore assemblages of mid-Pennsylvanian coals of West Vir- ginia and adjacent areas. Ph.D. Dissertation in Geology, West Virginia University, Mor- gantown, W. Va., 188 pp. Schopf, J.M., 1941. Mazocarpon oedipternum sp. nov. and sigillarian relationships. 111. State Geol Surv., Rept. Invest., 75, 53 pp. Schopf, J.M., 1961. Coal-ball occurrences in eastern Kentucky. U.S. Geol. Surv., Prof. Pap., 424B: 228?230. 109 Schutter, S.R. and Heckel, P.H., 1985. Missourian (early Late Pennsylvanian) climate in Midcontinent North America. In: T.L. Phillips and C.B. Cecil (Editors), Paleoclimatic Controls on Coal Resources of the Pennsylvanian System of North America. Int. J. CoalGeol., 5: 111?140. Scott, A.C. and Chaloner, W.G., 1983. The earliest fossil conifer from the Westphalian B of Yorkshire. Proc. R. Soc. London, 220B: 163?182. Scott, D.H., 1920. Studies in Fossil Botany, 3rd edition. Vol. 1, Pteridophyta. A. & C. Black, Ltd., London, 434 pp. Smith, W.H. and Bengal, L.E., 1975. Coal reserves of Illinois, Harrisburg-Springfield (No. 5) Coal. 111. State Geol. Surv., Map. Smith, W.H. and Stall, J.B., 1975. Coal and water resources for coal conversion in Illinois. 111. State Water Surv. 111. State Geol. Surv., Coop. Resour. Rep. 4, 79 pp. Stopes, M.C. and Watson, D.M.S., 1909. On the present distribution and origin of the calcareous concretions in coal seams, known as "coal balls". Philos. Trans, R. Soc, 200B: 167?218. Stschegolev, A.K., 1975. Die Entwicklung der Pflanzenbedeckung im Suden des Euro- paischen Teils der UdSSR, von Ende del Mittelkarbons bis zum Perm. Umfang und Gliederung des oberen Karbons (Stefan). Septieme Congres Int. Stratigr. Geol. Carboni- fere, (Krefeld, 1971), C.R., 4: 275-280. Taylor, T.N. and A.C. Scott, 1983. Interactions of plants and animals during Carboni- ferous. Bioscience, 33: 488?493. Treworgy, C.G. and Jacobson, R.J., 1979. Paleoenvironments and distribution of Penn- sylvanian age low-sulfur coal in Illinois. Ninth Int. Congr. Carboniferous Stratigr. Geol. Abstr. Papers, p. 217. Troll, C, 1970. Das "Baumfarnklima" und die Verbreitung der Baumfarne auf der Erde. Tubinger Geogr. Studien, 34: 179?189. Tschudy, R.H., 1969. Applied palynology. In: R.H. Tschudy and R.A. Scott (Editors), Aspects of Palynology. John Wiley and Sons, New York, N.Y., pp. 103?126. Vail, P.R., Mitchum, Jr., R.M., and Thompson, III, S., 1977. Seismic stratigraphy and global changes of sea level, Part 4: Global cycles of relative changes of sea level. In: C.E. Payton (Editor), Seismic Stratigraphy ? Applications to Hydrocarbon Explora- tion. Mem. Am. Assoc. Pet. Geol., 26: 83?97. Wagner, R.H., 1966. Sur I'existence dans la cordillere cantabrique de series de passage entre Westphalien et Stephanien: la limite inferieure de ces formations cantabriennes. C.R. Acad. Sci., Paris, 262: 1337?1340. Wanless, H.R., 1969. Marine and non-marine facies of the Upper Carboniferous of North America. Sixieme Congr. Int. Stratigr. Geol. Carbonifere, (Sheffield, 1967), C.R., 1: 293-336. White, D. and Thiessen, R., 1913. The origin of coal. U.S. Bureau Mines Bull. 38, U.S. Government Printing Office, Washington, D.C., 390 pp. Willard, D., 1984. Upper Pennsylvanian coal-swamp vegetation of the Illinois Basin Coal Field. Master's Thesis, University of Illinois, Champaign-Urbana, 97 pp. Williamson, A.D. and Williams, D.A., 1984. Paleochannel system above the Springfield Coal in western Kentucky. Geol. Soc. Am., Southeastern and North-Central Sections, Abstr. Programs, 16: 205. Wilson, L.R., 1976. Desmoinesian coal seams of northeastern Oklahoma and their paly- nological content, coal and oil potential of the Tri-state area. Field trip guidebook, Tulsa Geol. Soc, pp. 19?32. Winslow, M.R., 1959. Upper Mississippian and Pennsylvanian megaspores and other plant microfossils from Illinois. 111. State Geol. Surv., Bull. 86, 135 pp. Winston, R.B., 1983. A Late Pennsylvanian upland flora in Kansas: systematics and en- vironmental implications. Rev. Palaeobot. Palynol., 40: 5?31.