”I? 9.49;"? ”NJ' . ::""""':, u ‘15 I"! .I , ‘ _‘ -: 1 1' ' .. 2 g A‘ ’ '- I ' 3 ' ".. \ . .2. IVn'n'. I. It. "' '. ‘0’."1‘ ‘ A .2 ”'1 xx ' ‘1} 01" (IN. "I I '2)! 2'3? and“ R" )IUI'JFWVI” "A‘I' [25" "5%? 2'- -'r;« "2 15’2"“. "" 'IEIJI'," i‘. P"% .3?" A}. "' W} «'25- ?‘5‘; 3.45%: 3'75" "I?" (\WYVWMTWWIZLRX' gig 5‘2 { ,2 ' . I '0“ 1‘ 5.5" 01%;;th .'T‘£\*'i.:" . Jfiiw :"I. "Vififl'dl‘ m .14- :4 .'- "5 1535;" '5’)” 2' g. 5 .."'v 1- I. n'sw. l \ "L" '4" 'o -Yf\""""' 5'5""? Xvi}, 1")“ 3'" “Kit" "¢"'l-':?Sf'¢' ‘3'" 5331225.: 2 , '2I . 1‘4““, . 5 '5'&:J'£IW"'PQH;K 5""'\' v.15}, I-I‘: ¢n..é I 2! "‘g'g'wlI-2'5 ‘4 _ ("OI all-’2‘ ri.‘ ffif, 52‘ ~‘ 52""‘y'fi' -'-...- {-5- .2 ;II:.. I .\ I“ 0).} 2“} ‘r. "'5; "If} “NI" #1:. '1'“ I \2‘ 'b *WQ; \J.‘ . . 2 .. V. '2;."j"§"‘l' ¢‘ '.;-‘ .2 '1‘“:c::{.'I" :'L%h‘""""""§N""':hl"“‘f" :1": 23"“. (, If.‘ tgkfi'ufi . V "'i .V‘VJ; . “ . ' 552*" "-"5"'2 2.22 2:122:22; 2.2222222»: """'$&:.t‘«12$2 .2222 2.222.222»— 2:24» v n o I -~ 2‘ ""2 I' t I“. "'.- Kn. ' 'f. A ..' . "~'.o . ' I' “ W5" 'I’I‘J-I’» 32““ “1:5.It-tI"“‘-2'-wuumo a m>o~uo o ¢Hx¢>oqmomumuo massage m >unu ocmxiomcm z somuocmocwa H umumwumz o Smdu cmsaswch z Uncommon I xmum O ocmzu Hszm a >zprimary xylem are the large metaxylem tracheids which are mostly pentagonal in cross section measuring about 50-60.um in diameter, They exhibit bordered pitting ‘wall-thickening,.Alternate, crowded, horizontal, elongated pits are present on‘both radial and tangential walls (Pl. 2, Figs. 5, 6). A few metaxylem tracheids adjacent to protoxylem have relatively small diameter (30-40 am) and 29 are characterized by a scalariform thickening pattern (Pl. 2, Figs. 5, 6). In the metaxylem zone, there are a few scattered , parenchyma cell strands that are mostly one cell wide and 10-15 cells high in longitudinal section. Secondary xylem is conspicuous and up to 3.0 cm thick. It is composed of tracheids and xylem rays. The xylem rays are interspersed to divide the xylem into radially aligned segments, the width of which varies from one to 4 cells (Pl. 2, Fig. 3). Tracheids of the one- cell wide segments are pentangular in cross section, about 35-55.nm in diameter. 4-7 rows of alternate, crowded, angular, bordered pits with crossed apertures are present on the radial‘walls similar to those of metaxylem. The xylem rays are uniseriate or multiseriate. Vascular cambium and secondary phloem are rather poorly preserved. Between immature secondary tracheids and phloem there is a vascular cambium zone.'There is a thin zone of phloem composed of sieve cells and phloem rays. Sieve cells are 20-25.um in diameter. Sieve area and sieve pores have been observed on both cross section (Pl. 2, Fig. 3) and longitudinal section (Pl. 2, Fig. 7). Phloem rays are opposite to xylem rays forming continuous rays from secondary xylem through secondary phloem (Pl. 2, Fig. 3). Cells of the phloem rays are roughly isodiametric measuring 20-25.um in diameter. Conspicuous canals that contain amber-colored 30 material surround.the secondary phloem and are scattered in the outer cortex (P1. 2, Fig. l). The canals are oval in cross sectional outline and up to 0.2 x 1.0 mm in cross section. They are elongated vertically measured to 9.0 mm without ends visible (Pl. 2, Fig. 4). The canals have been interpreted as resin canals (Rothwell and Taylor, 1972). Primary cortex is about 2.0 mm thick. In the outer part of the cortex there are radially elongated sclerenchymatous strands. Neither fusing nor branching of the sclerenchyma strands has been observed in longitudinal section (Pl. 2, Fig. 2). No periderm is present. An isolated petiole with an abaxial, toothed vascular bundle and an attached rachis with an Y-shaped vascular bundle has been Observed (Pl. 2, Fig. 8). Heterangium crossii sp. nov. (PLATES 3 -6) Pteridospermales Lyginopteridaceae Heterangium Heterangium crossii Feng Diagnosis: Stems up to 1.1 cm in diameter with mixed protostele: 1-, 2-, or 3-celled groups of metaxylem tracheids are randomly arranged in ground parenchyma. Primary xylem is pentagonal in cross section and divided 31 into 5 sections by longitudinal parenchyma plates that radiate from the stem center. Protostele is exarch. The protoxylem strands are distinct at periphery of the primary xylemu Some protoxylem strands are in pairs, one strand on each side of a parenchyma plate: some are single» Total number of the protoxylem strands: as many as 18 counted. Secondary xylem conspicuous, 2 mm thick with uni-, bi-, or triseriate rays. Sieve cells of secondary phloem with numerous sieve pores on walls without differentiation of sieve plate. Prominent periderm layer on periphery of the secondary phloem consists of phelloderm and phellem. Peripheral cortex is composed of 6-7 cell layer of normal ground parenchyma and 2-3 cell layers of small, sclerotic cells at outermost cortex. Epidermis consists of one layer of small cells. The cuticle is about 10.nm in thickness. The petiole contains a bilobed vascular bundle. Holotype: Slides and peels of specimen of coal ball MSUCB SB are deposited in the Paleobotanical Collections, Department of Geological Sciences, Michigan State University, East Lansing, MI 48824. All figures present on Plates 3-6. 32 Etymology: The specific epithet "crossii“ is in honour of Dr. Aureal T. Cross for his contributions to paleobotany and for his collection and study of the Derringer Corners coal balls studied in detail in this dissertation. Locality and Age: A local coal below the Washingtonville marine shales and above the Middle Kittanning Coal at Derringer Corners, Lawrence County, western Pennsylvania: the age is the upper Middle Pennsylvanian, Allegheny Formation. Description: The new species is based on 3 stem segments. The stems are up to 11 mm in diameter. The primary xylem is monostelic and the mixed protostele measured 3-4 mm in diameter, Metaxylem tracheids of the primary xylem are 1-, 2- or 3- celled in groups and surrounded by parenchyma cells.'The metaxylem tracheids are roughly circular in cross section and range up to 260.um in diameter (Pl. 3, Figs. 1, 5). Longitudinal sections show metaxylem tracheids with tapered ends that exhibit scalariform thickening or multiseriate, crowded, bordered pits on the walls (P1. 4, Figs. 3, 4). Parenchyma cells which form a one- or two- cell thick layer surrounding the metaxylem tracheid strands are small and elongated peripherally. 33 Parenchyma cells between the metaxylem tracheids are relatively larger and up to 80m in diameter. The protostele of the primary xylem is divided into 5 sections by longitudinal parenchyma plates that radiate from the stele center (P1. 3, Fig. l). The parenchyma plates are 3-4 cells wide and connected with rays of the secondary xylem (Pl. 4, Fig. 1). The cells of the parenchyma plates are square or slightly elongated parallel with the length of the stem and exhibit scalariform thickening on the walls, while the cells of the rays of the secondary xylem.are horizontally elongated and without sclerotic thickening (Pl. 3, Fig. 4). Among the parenchyma cells of the protostele, the same type of scalariform cells also can be observed on both cross- and longitudinal sections (P1. 3, Fig. 2). The primary xylem is exarch: protoxylem strands occur on the periphery. Some protoxylem strands are single, others are in pairs, one strand on each side of the parenchyma plate. The single protoxylem strands are larger and more distinct than the paired ones (Pl. 4, Figs. 1, 2). As many as 18 protoxylem strands have been identified in a single stem. The protoxylem strands protrude into the secondary xylem, resulting in an undulating inner margin of the secondary xylem in cross section (Pl. 4, Figs. 1, 2). The elements of protoxylem are 10-20.um in diameter and show spiral and/or scalariform wall thickenings (Pl. 4, Figs. 3, 4). 34 Secondary growth is conspicuous in two specimens. Secondary xylem is about 2 mm thick (Pl. 4, Fig. 1: P1. 5, Fig. 7). The tracheids of the secondary xylem vary 'in size and shape. Generally, they are smaller than those of the metaxylem and are as much as 160 sum in diameter. They are rectangular or polygonal in cross section with multiseriate. crowded, bordered pits present on the walls (Pl. 4, Fig. 4). Uni-, bi-, or triseriate rays are interspersed in the secondary xylem (Pl. 4, Figs. 1, 3). The rays that abut the protoxylem strands or the parenchyma plates of the primary xylem are usually, if not always, wider than others. No parenchyma cell with secondary thickening has been observed in the rays. Vascular cambium and secondary phloem are poorly preserved. However, in some places, sieve cells of the secondary phloem can be recognized in both cross section (Pl. 4, Fig. 5) and longitudinal section (P1. 4, Figs. 6, 7: Pl. 5, Figs. 1-4). Outside the secondary phloem zone there is a layer of cells most of which have very thick walls and small lumens. This is phellem. The cells are almost isodiametric (P1. 5, Figs. 5, 6). Between the secondary phloem zone and the phel 1em area there is a layer of tangentially-elongated, thin-walled parenchyma cells. This layer may represent phelloderm. Most outer cortex was destroyed due to the formation of periderm. In one cross section, a small 35 amount of outer cortex and epidermis are intact (P1. 5, Fig. 6). Outside of the phellem the cortex is composed of a 4-5 cell-wide zone of parenchyma cells and a 2 cell-wide sclerotic layer: no sclerotic strands have been identified. Epidermis is a layer of small compact cells with rather thick cuticle (about 10 um). In serial cross sections, 3 leaf traces have been Observed. The specimen is not long enough to determine the phyllotaxy. All three leaf traces occur adjacent to the parenchyma plates of the primary xylem and are bilobed: i. e., each of these leaf traces arises from a pair of the protoxylem strands located on each side of the parenchyma plate. However, it is observed that, in longitudinal section, a leaf trace arises from a single protoxylem strand not adjacent to parenchyma plate. The leaf traces depart from the stem very slowly. In a distance of about 10 mm, the leaf trace is almost parallel to the stem axis (Pl. 6, Fig. 4). After passing through the secondary xylem, leaf traces are surrounded by phellem (P1. 6, Fig. l). The leaf trace is composed of spiral and scalariform protoxylem elements, metaxylem tracheids, possessing bordered pits with roughly horizontal elongated apertures, and primary phloem (P1. 6, Figs. 2, 3). At the base of petiole, leaf trace becomes alate in cross section near a dichotomy (P1. 3, Fig. 5). Another young stem specimen with little secondary 36 growth is 4 mm in diameter. In primary xylem, only 3. instead of 5, parenchyma plates are Obviously present (Pl. 3, Fig. 5). It seems likely that the number of the parenchyma plates varies and increases with stem growth. Most extrastelar tissue have been destroyed so that the outer cortex and epidermis are not preserved. Discussion: Heterangium is a‘well-known pteridosperm.genus which mainly includes stem and petiole remains. The genus was founded by Corda in 1845, based on fragmentary stem specimens from the coal balls of Bohemia, of middle Coal Measures age..As the generic name indicates, originally' Corda interpreted the wood to be composed of two kinds of vessels, large ones and small ones. The larger elements are the tracheids of primary wood: but the smaller ones were confused with the intermingled parenchyma cells. According to the comprehensive work of Scott (1917), the genus is characterized mainly by a mixed mesarch, protostelic stemwwith sclerenchyma bands in the outer cortex and horizontal sclerenchyma plates in the inner cortex. Scott recognized two subgenera based mainly on leaf trace emission: l) subgenus Heterangium (= Euheterangium: the prefix Eu- is not allowed according to Article 21 of the Botanical Code (ICBN)): leaf traces are simple when initiated, and the protoxylem area is poorly 37 defined : 2) subgenus Polyangium: leaf traces are initially double: protoxylem strands are well defined and approach the exarch condition. Most species of the genus have been reported from England. Heterangium americanum, the first species described from North America (Andrews, 1942), was distinguished by the small size of metaxylem tracheids groups (2- or 3-celled).'The species is assigned to the subgenus Polyangium. Shadle and Stidd (1975) reported four orders of branching of fronds of Heterangium americanum which is vascularized‘by 2-10 traces depending on the levels where the sections were cut. According to the authors, the fronds of H. americanum may be correlated with the compression fossil Sphenopteris Obtusiloba.'The assignment of the fronds to H. americanum is not based on organic connection but upon anatomical features common on both stem and frond, such as cortical sclerotic plates, outer fibrous cortex, epidermal features (including cuticle), and xylem pitting. Another monostelic, monotypic genus Microspermopteris was initially described by Baxter (1949, 1952), and emended by Taylor and Stockey (1976). The mixed protostele of M. aphyllum is very close to that of Heterangium. The differences between the two genera are as follows: 1) the primary xylem of Microspermgpteris is divided into 5 sections by longitudinal parenchyma plates 38 that extend radially'from the stem center, whereas the parenchyma plates are absent in the primary xylem of Heterangium: 2) in Heterangium, the protoxylem strands are mesarch and occur singly on the periphery, whereas exarch protoxylem strands occur in pairs, one strand on each side of a parenchyma plate in Microspermopteris: 3) Heterangium fronds possess 2-10 vascular bundles, but only a single C- shaped trace is present in Microspermopteris. Stidd (1979) has, however, described a new species, Heterangium lintonii, the second species of Heterangium reported from North America. The species appears to share more features with Microspermopteris ME in exarch protoxylem and the radially aligned parenchyma in primary xylem, though, according to Stidd, the longitudinal continuity of the radially aligned parenchyma was not observed in H. lintonii. The principal difference between the two species is the position of protoxylem strands and leaf traces. In H. lintonii, a single trace becomes double in the petiole base, and protoxylem strands occur singly on the periphery: whereas the petiole has a C-shaped single trace and protoxylem strands occur in pairs in Mircospermopteris. Therefore, H. lintonii shows very close affinity to M. aphyllum. Pigg at e1. (1986) described a stem of Microspermopteris aphyllum with an axillary branch and ~petiole attached. According to these authors, the frond of 39 the stem consists of three orders of branching, and the pinnules are 2-, 3- or 4- lobed. This study adds differentiating features of branch and frond between the two genera, Heterangium and Microspermopteris. Taylor and Stockey (1976) pointed out that until the reproductive organs are known from‘both Microspermopteris and Heterangium, petiole anatomy would continue to be the distinguishing character between these two genera. It is of interest to note that the Derringer Corners specimen shows the features of both E. lintonii and M. aphyllnnm It.bears extremely close resemblance to H. lintonii in primary xylem structure, leaf trace emission and prominent periderm layers: but its radially aligned parenchyma forms longitudinally continuous plates dividing the protostele into 3 to 5 sections similar to the condition of M. aphyllum. The exarch protoxylem strands of the Derringer Corners specimen occurs on the periphery, some of which singly randomly distributed, as in H. lintonii, and some in pairs, one strand on each side of parenchyma plate as in M. aphyllum. The sieve cells of secondary phloem of the Derringer Corners specimen are without definite sieve plates which differs from the presence of horizontal sieve plates in sieve cells of H. americanum (Hall, 1952). Since the petiole anatomy is thought to be the distinguishing feature for the two genera, Heterangium and 40 Microspermopteris, the Derringer Corners specimens, with double traces at the petiole base, is assigned to Heterangium. Because some leaf traces are initiated singly, while some are initially double, it seems to be proper not to place the Derringer Corners specimens into any subgenus of Heterangium. With the link of the Derringer Corners specimens, the new species M. crossii, the structures of stems, especially vasculature, M. lintonii, M. crossii, and M. gpgyilggiare so close to each other that it is possible that the three species may represent different portions or different developmental stages of the plants within the same species, or the same genus. The difference of frond and branch between these two genera may indicate only an interspecific difference. It is better, for the time being, to leave the relationship of the three species Open to question until more information about fronds and reproductive organs are available. MEDULLOSACEAE Medullosan Stems: Sutcliffia insignis (PLATES 7-10) Scott, 1906, p. 69-72, pl. 7-10, figs. 1-22. de Fraine, 1912, p. 1032—1035, figs. 1-5: p. 1036, figs. 6-7: p.1038, fig, 8: p. 1040, fig. 9: p. 1043, fig. 10: p. 1044, fig. 11-13: seg em Pro Pan mi). 00M tree Stra 41 Phillips and Andrews, 1963, p.33, fig.l: p. 39-51. p1.I-VII. Stidd, Oestry and Phillips, 1975, p. 51, fig.l: Pe58' PleI: Pe61, p1e2e - 11-22. Comments: Sutcliffia insignis, a medullosan stemigenus, was originally described by Scott (1906).'The specimen from the Lower Coal Measures of Lancashire, England, measuring 12 X 65 cm in diameter, and the detached petioles were of about the same size. The vascular system of the stem is composed of massive stelar segments of which one or two large ones occupied the central part of the stem, and were more or less surrounded by smaller stelar segments that departed from the central ones and in turn continued to divide to produce the leaf traces. The large stelar segments consist of a massive amount of primary xylem enveloped by weakly developed secondary xylem. The exarch primary xylem is composed of protoxylem and large metaxylem tracheids with interspersed parenchyma so that the large stelar segments are exarch “mixed" protostele. The leaf traces were thought to be concentric (traces with phloem surrounding xylem) and each trace was surrounded by a ring of sclerenchymatous strands. de Fraine (1912) described the same species also 42 from the Lower Coal Measures of Lancashire, England and prepared a restoration drawing of the cauline vascular system in which one or two large central segments divide and anastomose repeatedly, with some smaller vascular divisions departing as leaf traces. It was not until 1963 that Phillips and Andrews reported the occurrence of Sutcliffia with the petiole attached from a middle Pennsylvanian coal ball from Illinois. The Illinois' specimens were not very well preserved anatomically and were regarded as being closely related to the English species with the exception of the presence of rather massive cortical emergences. Therefore, Phillips and Andrews named the Illinois specimens as a variety of Sutcliffia insignis, Sutcliffia insignis Scott var. tuberculata. Stidd et al. (1975) described Sutcliffia foliage members from.seven localities of the Illinois Basin ranging in age from the Middle to Upper Pennsylvanian. According to the authors, the fronds were large, probably 3-4 m long and 4-5 times pinnate. A diagrammatic drawing depicting a suggested relationships among isolated frond members of Sutcliffia was made. With the exception of the petiole attached to the stem, the attachment of these frond members to the stem was not known. The criteria used by Stidd et a1. (1975) to identify isolated Sutcliffia frond fragments are as follows: 1) fewer and larger bundles: 2) concentric bundles enclosed by fiber rings or 43 embedded in outer sclerenchyma zone: 3) arrangement of vascular bundles: 4) the quite large emergences on the surface. In their report, no pinnule had been found attached to any branching order of the pinna. Based only on the association of LinOpteris pinnules with Sutcliffia in every coal ball used in their study, Stidd et a1. (1975) suggested that LinOpteris pinnules were borne on Sutcliffia fronds. An anatomically preserved pteridosperm pollen organ was identified by Stidd (1978) with the compression fossil, Potoniea. Because the Potoniea specimens were associated with "net-veined (Reticulopteris-Linopteris), Stidd (1978) further suggested that Potoniea might have been borne on fronds of LinOpteris which , in turn, was believed by Stidd etia1.(1975) to‘have been borne on Sutcliffia stems (Stidd, 1978). Description of the Derringer Corners Specimens: Several specimens‘T‘ ~ ‘\§\\ \\\\\\§\\§\“ {7' V $.39 ‘9 .~ -. swamps ” 3 d k“ “1‘: ‘\\\\\\\\\\\1\\ _ A <1t vQ .4 g o 3 \ \ we \\\\\\\\\§\§§\ 5 S? gig , mocommcm conomcs \ \ \\ \‘ u . “‘Q‘ "T 1“ / ..\ 2 g < 4 \\\\\\\\:\\‘\\\\§§ ; f} 3. 5? «ONERGENCE or MAJOR Lvoowoos : E u 'm‘m‘m“ ”“5““ \\\\‘\\\t chopoos :1 Eunmemcau commas a _ ‘ " PARALYCOPOOIYES ) sum as Lvoossoms spears u m rmsr DRIER INrERVAL . LEPIOOPHLOIOS , «masoucnou or rat: runs 3 . \ nancounm )sout unasnuuo vaoeunou _ \\\\\ «summon or LYGINOPTIRIS Tr" \ ' ; Ltmoootnonou , g 4 \\\\\\\ t ; VASCULARE ' wenasr a\ omesr l 1 . . . . . . Fig. 5. Generalized stratigraphic patterns of abundance of Pennsylvanian coal-swamp vegetation in the United States with a relative wetness curve indicating the alternation of the wet and drier intervals (from Phillips at e1.. 1985) synchronously (or nearly so) near the Lower-Middle Pennsylvanian (Westphalian A-B) and the Middle-Upper Pennsylvanian (Westphalian D- Stephanian) boundaries, respectively. These vegetative changes were consistent with two drier intervals of the Euramerican tropical belt during the Pennsylvanian Period, which is supported by evidence from paleobotany, palynology, coal petrography, sedimentology, distributions of coal beds and evaporites, and paleotectonics. The massive reduction of arborescent 1yc0pods except for Sigillaria, and the rising dominance 127 of tree ferns in the coal swamps coincided with the secondary interval (Fig. 5). However, it is evident, on the other hand, that the massive reduction of lycopod trees at the Middle-Late Pennsylvanian boundary is a complex process, and cannot be attributed only to wet-dry alternation of the paleoclimate. This will be discussed later. SOME LIMITATIONS OF QUANTITATIVE ANALYSIS OF VEGETATION PATTERNS OF COAL SWAMPS Identifying Some Dismembered Plant Remains: There is great difficulty in identifying and classifying some dismembered plant parts preserved in coal ‘balls for quantitative analyses (Phillips et al., 1985). Identification of coal ball plants is based on their anatomical features. Distinguishing a fossil plant to generic or specific level generally requires at least part of the vegetative (stem, leaf, and root) or the reproductive organs, and, in many instance, a combination of both vegetative and reproductive features. Generally, ‘well-preserved coal ball plants can be identified by using serial sections, complete cross sections, medium- 1ongitudina1 sections, or a combination of the cross sections with the longitudinal sections of plant parts or organs. But, during a quantitative analysis procedure, it 128 is sometimes very difficult to identify some plant remains in coal balls because the identification is only based on one random section which may be a cross-, longitudinal—, or Oblique-section. An example of the difficulty in making such identification is illustrated by the following examples. The main difference between two cordaitean stem genera Mesoxylon and Pennsylvanioxylon is, for instance, that the leaf traces of Mesoxylon appear to be double and the sympodia (separate protoxylem strands), from.which the leaf traces arise, are mesarch (Traverse, 1950: Baxter, 1959: Cohen and Delevoryas, 1959), while the leaf traces of Pennsylvanioxylon rise singly from an endarch sympodial complex (Vogellehner, 1965: Taylor, 1981, p.425-426: Stewart, 1983, p. 328-329). Obviously, it is impossible to distinguish the two genera from each other without making serial longitudinal sections to show the structure of the sympodia and to follow the course of a leaf trace. The same difficult situation exists in distinguishing genera of lycopod trees. For example, Paralycopodites, first described by Morey and Morey (1977), differs from Lepidophloios and most species of Lepidodendron in having persistent, short, scale-like leaves (DiMichele, 1980L.This means that if random sections of lycopod stems, present on peels used for quantitative analysis, do not pass through attached leaves, there is no way to separate Paralycopodites from other lycopod genera. Also, the differentiation of 129 Diaphorodendron, a lycopod stem genus established by DiMichele (1985), from other lycopod stem genera requires the identification of a combination of several anatomical features of both vegetative and reproductive organs (DiMichele, 1985). With these few examples, one can understand the difficulty in identifying some dissociated plant parts in coal balls, based on only a random sectionlavailable on a peel, in making a quantitative analysis of the constituent plants, comprising the peat, which are preserved in a coal ball. When such difficulties do appear occasionally, it is better not make any further guess beyond that which the random sections can illustrate because correct identification is required for quantitative analysis and incorrect identification would significantly reduce the reliability of the quantitative data. Shoot/Root Ratio: Another limiting factor is the shoot (stem and its appendanges)/root ratios of coal ball plants. The amount of aerial plant parts is crucial for community analysis. From living taxa, it is estimated that 20-30% of original biomass of 1ycopods was root system and 70-80% was aerial part, i.e., a shoot system (Phillips and Peppers, 1984).'Therefore if a‘whole tree was preserved as peat, it would result in a shoot/root ratio of roughly 4/1. But the actual shoot/root ratios of peats are 130 commonly close to 1/1 indicting extensive disproportionate loss of aerial plant parts. This greatly reduces the accuracy of reconstructing the relative amounts of plants which make up the community structure of the vegetation of peat deposits represented by coal‘balls (Phillips and Peppers, 1984). THE PURPOSE OF THE PRESENT ANALYSIS: Previous investigations of coal swamp vegetation have been concentrated on the Illinois basin and the central Appalachian Basin in the United States because of the well-known coal ball plants from the Illinois Basin and of the many studies of the coal geology of the Paleozoic in the central Appalachian Basin. The Appalachian Basin presents an extensive geological record of the upper Paleozoic (Gillespie and Pfefferkorn, 1979). In southeastern West Virginia and southwestern Virginia, sedimentation was continuous across the Mississippian-Pennsylvanian boundary (Englund, 1969). In many other parts of the Appalachian Basin and in other basins of the United States there is an unconformity at the Mississippian-Pennsylvanian‘boundary: the central Appalachian basin contains the most complete biostratigraphic and lithostratigraphic record of the coal-bearing Pennsylvanian and the Lower Permian strata (Cecil et al., 1985). A comparison of the changes in coal swamp 131 vegetation of the Appalachian and the Midcontinent regions between the first and the second drier intervals of the middle part of the Middle Pennsylvanian has been difficult because of the dearth of permineralized peat deposits in coal beds in the Appalachian Basin (Phillips et al., 1985). The Derringer Corners coal balls studied here were collected from near the Pennsylvania-Ohio state line in the Northern Appalachian Basin, geographically the easternmost coal ball locality known to date in the United States.'The stratigraphic position of the locality is in the Kittanning coals of the Allegheny Formation, near the Middle-Late Pennsylvanian boundary at the time of the second drier interval. It is very close stratigraphically' to the well-known, coal ball-rich seams, Herrin (No. 6) or Harrisburg (No. 5) coals in Illinois. Phillips et a1. (1985, p. 81) made a preliminary quantitative analysis of 35 pieces (1,937 cm2) of the Derringer Corners coal balls from.our collections. The present quantitative analysis of the Derringer Corners coal‘balls, the only.A11egheny coal ball collection that has been quantified, provides some significant data for study of coal swamp vegetation of the Appalachian Basin near the second drier interval, and for comparison of the vegetation pattern in coal swamps between the Appalachian Basin and Illinois Basin. 132 TECHNIQUES OF ANALYSIS: The techniques used in the present study are basically those given by Schopf (1938b), Phillips (1976), Phillips et a1. (1977), Phillips and DiMichele (1981), and Raymond et a1. (1984). Coal balls were cut with a rock saw, serially, into pieces about 2.0 cm apart. Cellulose acetic peels were made from each surface. Plant parts included on the peels were measured using a grid system under a dissecting microscope. The measured plant parts of different groups were classified and the percentages of each were then computed. Some aspects of the techniques in this study are described below in more detail. 1) In the analyses given by Phillips et a1. (1977), coal ball samples were collected stratigraphically through the entire thickness of the coal seams (vertical profile sample): the results of the analyses of the vertical profile samples provide greater accuracy than that of random samples. According to Raymond (in press), an analysis of a random sample based on 2,000 cm2 or more yields quantitative results closest to those of vertical profile samples collected from the same locality as the random samples. Although the Derringer Corners coal bal ls were randomly collected, i.e., without recording their orientation or position in the vertical profile (stratigraphically), all coal balls or broken pieces of 133 coal balls, from bottom to top, of the large concretionary mass of coal ball concretions which occupied the full thickness of the coal seam were collected for analysis by Dr. A. T. Cross and his students during 1965-1967. There may be little difference in results of the quantitative analysis of the present collection from those of coal balls from vertical profile samples. The sample size analyzed here (13,203 cm2) of the Derringer Corners 2 quantitative collection is much larger than 2,000 cm analysis of coal ball peels from a random collection of coal balls from Iowa reported by Raymond (in press). 2) Because of the rather large collection of the Derringer Corners coal balls, it was not reasonable to measure all peels made from the whole collection. Phillips et a1. (1977) used 189 coal balls or pieces of coal balls to test a "middle-peel" approach (measurement of the peels from the middle slice of each coal ball only) to determine quantitatively and qualitatively the relative proportions of the plant materials. The results obtained from the "middle-peel" method was very close to that obtained by the quantitative analysis of all peels from every slice, including the middle one, of each coal ball measured. Accordingly, the Derringer Corners collection (342 coal balls or pieces of coal balls) is large enough to use the ”middle-peel" method for analysis. In this study, 342 middle peels have been measured with two exceptions when two rare plant parts (a seed of Pachytesta and a piece of 134 Dolerotheca ) were identified on the other peels rather than the middle ones. In these cases, the two selected peels were chosen to replace the middle peels for analysis. 3) The method of measuring is modified from those given by Phillips et a1. (1977), Phillips and DiMichele (1981) and Raymond et a1. (1984). Those authors placed a 2 coordinate transparent plastic sheet marked in a one cm system over a peel from coal ball surface to be quantitatively and qualitatively analyzed. Using this grid system, they recorded only the taxonomic affiliations of the largest piece of plant remains seen in each grid square (1 cm2), and disregarded all other plant remains smaller than the largest one. This system was modified for reproductive organs. Phillips et a1. (1977) assigned one grid square (1 cm?) to each of the reproductive organs identified within a grid square, regardless of its size, whereas Raymond et a1. (1984) counted only the largest piece or whole reproductive organ within each grid square, and disregarded the presence of smaller reproductive organs. In the present study, a mapping method was used. A selected peel was placed over a paper grid system measuring of 0.259 cm2 for each square (the area of 100 squares is 5.09 x 5.09 cm), and a carbon paper was inserted between the peel and the grid paper. The outlines of all plant parts on the peel were sketched with a pencil under a dissecting microscope at a magnification of 10 x. 135 The resulting sketch of the plant remains on the grid paper shows the exact positions and areas occupied by all plant parts exhibited on the peel. Then the area of the squares occupied by different plants were counted and recorded. In some peels, certain areas were occupied by several different.kinds of plant debris (mostly leaves and rootlets): these plant units were sometimes so small and were mixed with each other so evenly that it was impossible to map and count each of these pieces or units separately. In such cases, besides mapping rare plant organs, the areas occupied by each plant group were estimated and mapped so that the mapping of these areas were not exactly consistent with the actual distribution of these plant parts on the peels. Because these areas were rather small relative totthe whole of the areas analysed, the bias caused by the imprecise mapping in these instances can be disregarded. It is believed that this mapping method should be more precise, then those applied by Phillips et a1. (1977, 1981) and by Raymond et a1. (1984), though it is not necessarily as efficient. 4) Because of the difficulty of identification of dissociated plant parts mentioned above, most plant remains were identified only to generic or group level in this study. 5) For calculating shoot/root ratios, one-half of 136 all Psaronius outer zone roots were considered to be aerial debris (Phillips and DiMichele, 1981L.This is in contrast to classifying all Psaronius outer zone roots, that appeared to intrude the peats, as root debris (Raymond, in pressL 6) Root penetration events were also computed. Penetrated units of plant debris were regarded as substrates of the penetrating roots, as proposed by Raymond (in press). The root penetration may indicate a successional sequence of coal ball plants, 14%, the growth of later plants on earlier litter or peat surfaces. RESULTS The quantitative analysis of 13,203 cm2 of the middle peels from each of 342 coal balls or pieces of coal balls from.the Derringer Corners, western Pennsylvanian of the later Middle Pennsylvanian age are presented below: 1) The coal swamp vegetation is composed of four major plant groups comprising about 30 genera, common in the Pennsylvanian Euramerican coal swamps: 1ycopods, pteridosperms (seed ferns), ferns, and sphenopsids (Table 1). 2) Lycopods were the dominant plants in the Derringer Corners coal swamps, contributing 66.6% of all identifiable plant parts preserved in the coal‘balls: pteridosperms were of the second importance (20.8%): ferns 137 TABLE 1 Peat Composition (%(near the mid-level) showing one primary rib: another primary rib is damaged. Because of the branching of two vascular bundles (at arrows), the total number of vascular bundles per valve of the integument is 10. x 3. Cross section of the seed at level E showing two primary ribs (at arrows) which are loss conspicuous than above, 8 vascular bundles per valve of the integument, cuticle of the endotesta, and remnant of the nucellus. x 7. Cross section of the seed at level F showing the integument, the cuticle of the endotesta, and the nucellus stalk. The ribs are difficult to identify so that the total number of vascular bundles of the integument at this level is uncertain. x 3. Cross section of the seed at level G showing double vascular system:'the outer ring of the vascular bundles enter the integument, and the inner ring of vascular bundles enter the nucellus. x 3. 5)! [V9 per e gen. do 905.. the 204 PLATE 21 Figs. 1-7. Pachytesta noei (continued). All Figures from 7. M.S.U.CB 116-6. (CUE=cuticle of the endotesta, ISO-inner part of the sclerotesta, OSC=outer part of the sol erotesta, MP=micropyl e, MX=metaxyl em, NU=nucellus, PX=protoxylem, SC=sieve cell of primary phloem, ST=sarcotesta) Longitudinal section of the seed integument showing a narrow layer of the sarcotesta (right), and two-zoned sclerotesta. x 70. Enlargement of Fig.1 showing the parenchymatous sarcotesta (right), and the horizontally elong- ated fibers (with single pits on cross section of the cell walls, at arrow) of the outer zone of the sclerotesta. x 304. Enlargement of Fig.1 showing the outer zone of the sclerotesta, the vertically elongated fibers with single pits of the inner zone of the sclerotesta, and the cuticle of the endotesta (at arrow). x 259. Enlargement of Pl.20, Fig.4 showing the fibers of the outer zone of the sclerotesta, and a strand of vertically elongated fibers (at arrow). x 242. Longitudinal section of a vascular bundle in the integument showing protoxylem, metaxylem, and primary phloem. x 1035. Longitudinal section of the seed showing the micropyle, and the tips of the integument where the fibers of the inner zone of the sclerotesta have shifted to be horizontally elongated similar to that of the outer zone of the sclero- testa. x 26. Longitudinal section of the seed near the base showing the nucellus remnant and nucellus stalk. x 48. PLATE 2| J City: . ‘1 O "t. L..: . w .2 a a . .. .hr.u""nw.u-.m-r. '7. . s 6...... . . {so .. l .1 l l . £ a I In. t h 9 .n1 tous ong 100 one of 1beru a me n m I: here esta :leIO' stalk use Figs. 1-8. 5. 6. 7. 8. 205 PLATE 22 Pachytesta noei (continued). All figures from M.S.U.CB 116-6. (CC=cannal cell, HR=hair, MX=metaxylem, SC=sieve cell, SH=vascu1ar bundle sheath, SM=stoma, PX=protoxylem, VB=vascular bundle) Longitudinal section of the seed showing irregular shaped cells of the nucellus remnant. x 242. Enlargement of Pl.20, Fig.8 showing the vascular bundle of the chalaza (the central hole). x 242. Longitudinal section of a vascular bundle in the nucellus. x 303. Cross section (above) and longitudinal section (below) of the seed base showing the remnant of the seed stalk (at arrow). x 70. Longitudinal section of the seed stalk remnant showing a vascular bundle and the bundle sheath. x 293. Enlargement of Fig.5 showing metaxylem, proto- xylem, sieve cells of the primary phloem of the vascular bundle in the seed stalk. x 938. Longitudinal section of the seed stalk showing a stoma on the stalk surface with papillate subsidiary cells . x 968. Longitudinal section of the seed stalk showing two multicellular hairs on the stalk surface. x 968. PLATE 22 e v e .1 5 ant. cular :24L n the ion tof ant 23th- the ing ing v. ‘v . 'h AN "'lllllmllllllillljfllillllllllll'“ 3 129