CHAPTER 3

Palynology and paleoecology of lignites from the Manning Formation (Jackson Group) outcrop in the Lake Somerville Spillway

By A. Raymond, M. K. Phillips, and J. A. Gennett

Department of Geology and Geophysics, Texas A&M University, College Station, TX 77843

ABSTRACT

Analysis of the palynomorphs in three lignite beds and the associated sediments from the Manning Formation (Jackson Group) outcrop in the Lake Somerville spillway of Texas reveals the presence of four palynomorph associations whose parent plants contributed to the formation of the lignite beds. The Cicatricosisporites-Laevigatosporites fern association occurs in the middle and upper seam and probably derived from freshwater marsh vegetation. The Momipites coryloides-Arecipites columellus-Liliacidites association occurs in all three seams and probably derived from a freshwater swamp forest with an open canopy or a wetland shrub community. The Momipites coryloides-Nyssa-Rhoipites angustus association occurs in the lower and middle seams and probably derived from a different freshwater swamp forest. The Cupuliferoipollenites-Cupuliferoidaepollenites liblarensis-Siltaria cf. S. scabriextima association occurs at the top of the lower and middle seams, which are overlain by marine sediments, and probably derived from a salt-tolerant swamp forest. Ailanthipites berryi, possibly derived from a tree or shrub related the 'tree of heaven', predominates in the claystone underlying the lower lignite seam and may have colonized the coastal plain prior to the initiation of peat accumulation.

INTRODUCTION

Three lignite seams of the Manning Formation (Jackson Group, Eocene) crop out in the Lake Somerville spillway of Texas. Gennett (1993) investigated the palynology and paleoecology of the lower and middle lignites at the spillway; her preliminary results suggested a strong similarity between the palynoflora of the Lake Somerville spillway lignites and that of the San Miguel lignite deposit from Atascosa and McMullen Co., Texas. In both deposits, the juglandaceous grain, Momipites, and the fagaceous grain, Cupuliferoipollenites, occurred commonly and fern spores and palm grains also contributed to the lignite palynoflora.

The lower two lignite seams at Lake Somerville showed a pronounced pattern of floral turnover. In the lower seam, Momipites dominated the base and middle of the seam and Cupuliferoipollenites dominated the upper 10 cm of the lignite. The middle seam differed from the lower seam in that fern spores dominated the base of that seam. However, Momipites dominated the middle seam from 10 to 80 cm above the base, and Cupuliferoi-pollenites dominated the upper 10 cm.

Phillips and Raymond undertook additional palynological studies at the Lake Somerville spillway outcrop in order to describe the palynoflora of the upper seam, to confirm the pattern of floral succession reported in the lower and middle seams, and to investigate the horizontal variability of the palynoflora at the top of the middle seam. In this contribution, we present palynomorph profiles of the three Lake Somerville lignite seams and the results of replicate sampling in the upper portion of the middle seam. We discuss the seam profiles in the context of work by Yancey (this volume) on the depositional environments and stratigraphic setting of the Lake Somerville spillway outcrop (1) to determine the cause of the shift from fern spore- to Momipites-dominated lignites at the base of the middle seam, (2) to determine the cause of the shift from Momipites-dominated to Cupuliferoipollenites-dominated lignites at the top of the lower and middle seams, (3) to determine if the upper seam shows the same pattern of floral dominance, and (4) to investigate how the claystone layers, interpreted as altered volcanic ash, were deposited within the lignite and how these depositional events affected the lignite-forming community.

ACKNOWLEDGMENTS

Conversations with T.E. Yancey, J. Over, P.D. Warwick, S.S. Crowley, L.F. Ruppert, F. Oboh, W. Elsik and F. Fleming and insightful reviews by N. Fredericksen and E. Robbins contributed to the ideas presented in this paper. Partial support of this research was provided by a fellowship granted to M.K. Phillips by the U.S. Department of the Interior Mineral Institutes Program administered by the Bureau of Mines under allotment grant G1134248 to the Texas Mining and Mineral Resources Research Institute and by grants from the Center for Energy and Mineral Resources of Texas to A. Raymond, which supported the dissertation and thesis work of Gennett and Phillips. The U. S. Geological Survey provided funds to analyze replicate samples from the top of the upper lignite. R. Day drafted the pollen diagrams; E. Heise drafted the lignite sections.

METHODS

Yancey (this volume) describes the Lake Somerville spillway section. The detailed stratigraphy and palynomorph profiles of the three lignite seams at Lake Somerville and the levels sampled by Gennett (1993) and Phillips and Raymond appear in figures 1-3. Gennett (1993) sampled the lower and middle seams at 10 cm intervals. In order to establish the relationship between lignite-forming communi-ties and lithological variations within the seam, Phillips and Raymond focused their sampling on intervals of lithologic change.

To process samples for palynomorphs, Phillips and Raymond pulverized and treated samples with 35% HCl. They removed silicates with 70% HF and oxidized samples with concentrated HNO₃. They removed humics with KOH and used ZnBr2 (specific gravity 2.0) for heavy density separation. They washed, mixed and centrifuged all samples between steps, stained the final residues with safranin and stored them in glycerine. In addition, they acetylated samples from the middle lignite seam. Gennett used 10% HCl and 50%HF in the acid treatment, removed humics with 5% NH₃OH and oxidized samples with 20% HNO₃. She used heavy liquid separation only for the clastic sample.

For the lower and middle lignite seams, we include both the profiles of Phillips and Raymond (figs. 1a; 2a, c, d) and those of Gennett (figs. 1b, 2b) because the location of profiles along the outcrop, the sampling interval, and processing techniques differed.

RESULTS

LOWER SEAM PROFILES

The lower lignite seam is 70 cm thick and overlies a volcanic ash deposit (Yancey, this volume). The upper portion of this deposit contains stump casts and a compression-impression leaf assemblage (fig. 1a). This volcanic ash in turn overlies a paleosol horizon containing lignitized logs and root systems. Lenses of claystone, interpreted as reworked volcanic ash, occur within the lower 10 cm of this seam. Argillaceous sandstone overlies the lower seam. This sandstone contains the marine diatom Cocinodiscus, sponge spicules, and abundant trace fossils including Gyrolithes, which also penetrate the underlying lignite and occur commonly in the upper 10 cm of the seam.

The species, Ailanthipites berryi (31%) and Caprifoliipites tantulus (9%), predominate in the carbonaceous claystone, interpreted as a volcanic ash (Yancey, this volume) that underlies the lower seam (1a). Ailanthipites berryi (37%) and Momipites coryloides (12%) predominate in a replicate sample of this claystone; both samples were gathered from within 6 cm of the base of the lignite. M. coryloides (35%), Cupuliferoipollenites spp. (14%), and C. tantulus (8%) predominate in a claystone lens derived from reworked volcanic ash (fig 1a). M. coryloides dominated the bulk of the lower lignite in both profiles (figs. 1a,b). In one profile, Cupuliferoipollenites spp. dominated the top of the lower seam (fig. 1b); in the other, M. coryloides dominated the top of the lower seam and Cupuliferoipollenites spp. contributed only 8% of the palynomorph sum at this level (fig. 1a). Brackish or marine dinoflagellates (Gonyaulacysta; Tuburculodinium) and grains attributed to modern mangrove taxa (Proxapertites, Heritiera littoralis, Scyphophoria hydrophyllaceae, Excoecaria agallocha, Brownlowia tersa; Conocarpus erectus; Avicennia eucalyptifolia) also occur at the top of the lower seam.

Momipites coryloides, Cupuliferoi-pollenites spp., and Cupuliferoidaepollenites liblarensis predominate in the marine argillaceous sandstones overlying the lower seam (fig. 1a). Mangrove grains encountered in these samples include: Spinizonocolpites echinatus (Nypa), Avicennia sp., A. schaueriana, and Rhizophora sp.

MIDDLE SEAM PROFILES

The middle seam overlies a sandstone with root traces interpreted by Yancey (this volume) as a shore-zone sand. Two persistent claystone layers, interpreted as altered volcanic ash, occur in the middle seam (fig. 2a). One thin layer of uniform thickness (2 cm) lies 55 cm above the base of the seam. A second claystone layer of variable thickness (<0.5 - 10.0 cm) lies 80 - 90 cm above the base of the seam. As in the lower seam, argillaceous sandstone overlies the middle seam. This sandstone contains the marine diatom Cosinodiscus and sponge spicules, as well as abundant trace fossils including Gyrolithes, which penetrate into the underlying lignite and occur commonly in the upper 10 cm of the seam.

Fern spores predominate in lignite samples taken from the base of the seam in both profiles (figs. 2a, b). Momipites coryloides along with Caprifoliipites tantulus, Rhoipites angustus, Cyrillaceaepollenites megaexactus, Nyssa and the monocot taxa, Arecipites columellus and Liliacidites predominate throughout the bulk of the middle lignite, from 10 cm above the base up to the lignite underlying the variable ash. Cupuliferoi-pollenites spp. dominated the top of the seam. We sampled the top of the middle lignite, the variable ash, and the lignite immediately beneath the variable ash in two additional places and found a range in Cupuliferoipollenites spp. and M. coryloides percentages at the top of the middle seam, from 17% Cupuliferoipollenites spp. and 17% M. coryloides to 50% Cupuliferoipollenites spp. and 5% M. coryloides (fig. 2a-d). Although Cupuliferoipollenites spp. shows a wide range of values in the upper layers of the middle seam, percentages of Cupuliferoipollenites spp. consistently peak near the top of the seam. In general, the palynomorph flora of ash samples matched that of the surrounding lignites (figs. 2a, c, d).

Momipites coryloides, Cupuliferoi-pollenites spp., Cupuliferoidaepollenites liblarensis, Caprifoliipites tantulus, and Siltaria cf. S. scabriextima predominate in the marine argillaceous sandstones overlying the middle seam. Mangrove pollen encountered in these samples include: Avicennia sp., Heritiera littoralis, and Brownlowia sp.

UPPER SEAM PROFILES

This upper seam overlies a sandstone with root traces interpreted by Yancey (this volume) as a shore-zone sand. Unlike the lower and middle seams, the upper seam has an interval of carbonaceous claystone bearing permineralized and lignitized logs (fig. 3). This carbonaceous claystone interbed is approximately 1 m thick and separates the upper lignite into two zones. Freshwater siliciclastic sediments overlie the upper seam (Yancey, this volume). Very few burrows from the overlying siliciclastic sediments penetrate the lignite at the top of the upper seam. The lower zone of this seam is fissile at the base and contains lignitized logs at the top.

Fern spores along with Momipites coryloides, Chrysophyllum brevisulcatum, and Tetracolporopollenites brevis predominate in the sample taken from the fissile lignite that composes the basal 12 cm of the lower lignite zone (fig. 3). Momipites coryloides, fern spores, Fraxinoipollenites, and palms (Arecipites columellus) predominate in the sample from the top of the carbonaceous claystone interbed. M. coryloides along with Cyrillaceae-pollenites megaexactus and two monocot pollen species, Liliacidites vittatus and L. tritus, predominate in the sample from the base of the upper lignite zone. M. coryloides, Ailanthi-pites berryi, and Caprifoliipites tantulus predominate in the lignite samples from the middle and top of the upper lignite zone of this seam. Counts of M. coryloides, Caprifoliipites tantulus, Cupuliferoipollenites spp., Cupuliferoidaepollenites liblarensis, and Siltaria cf. S. scabriextima present in two additional samples taken from the top centimeter of this seam confirm the predominance of M. coryloides (48% - 49%) and the virtual absence of Cupuliferoipollenites spp., C. liblarensis, and Siltaria cf. S. scabriextima (0 - 2%).

DISCUSSION

PATTERNS OF FLORAL DOMINANCE IN THE LAKE SOMERVILLE LIGNITES

The lignites of the Lake Somerville spillway share a similar floristic composition with the San Miquel lignite deposit from Atascosa and McMullen Co., Texas and may be nearly the same age (Gennett, 1993). However, these two lignite deposits differ in the strong pattern of floral zonation apparent in the lower and middle lignite seams of the Lake Somerville spillway outcrop. The middle seam of the Lake Somerville spillway outcrop shows the zonation pattern most clearly: ferns dominate the base of the seam; Momipites coryloides dominates the middle of the seam; and Cupuliferoipollenites spp. dominates the top of the seam. Phillips and Raymond confirmed the pattern of floral zonation originally reported by Gennett (1993), but found that the percentage of Cupuliferoipollenites spp. in the upper 10 cm of the middle seam varied considerably, from 10% to 50%.

The pollen percentages of Middle and Late Eocene lignites from the Texas Gulf Coast appear quite variable. Gennett (1993) also found wide variation in percentages of Cupuliferoipollenites spp., from 26% to 83%, in 10 replicate samples of lignite from the base of the C seam of the Eocene San Miguel lignite deposit, Atascosa and McMullen Co., Texas, sampled along the highwall of the mine at 30.5 m intervals. Thus, the consistent rise in the percentage of Cupuliferoipollenites spp. at the top of the middle seam may carry more significance than the actual percentage values.

The lower lignite seam of the Lake Somerville spillway outcrop appears to lack a fern-dominated layer. In one profile, Cupuliferoipollenites spp. replaces Momipites coryloides as the dominant palynomorph at the top of this seam (fig. 1b). In the other profile, M. coryloides predominates throughout the seam and only a modest percentage of Cupuliferoipollenites spp. (8%) occurs at the top of the lignite (fig. 1a). Within that profile, the greatest percentage of Cupuliferoipollenites spp. occurs in a claystone lens, which appears to represent volcanic ash transported into the swamp. Although the pollen percentages of these lignites exhibit a great deal of horizontal variability, with the exception of the claystone lens, the highest percentages of Cupuliferoipollenites spp. occur at the top of both the lower and middle seams.

In contrast to the Cupuliferoipollenites- Momipites transition observed at the top of the lower and

middle seams, Momipites coryloides predominates throughout the upper zone of the upper seam and Ailanthipites berryi contributes over 10% of the pollen sum to lignite samples from the middle and top of the upper zone of this seam. Replicate sampling confirmed the low percentage of Cupuliferoipollenites spp. (0 - 2%) at the top of the upper seam.

AFFINITIES AND PALEOECOLOGICAL INTERPRETATION OF DOMINANT TAXA

Laevigatosporites and Cicatricosisporites. Fern spores similar to Laevigatosporites occur in the modern families Aspidiaceae, Aspleniaceae, Blechnaceae, Gleicheniaceae, Lomariopsidaceae, Polypodiaceae, and Pteridaceae (Frederiksen, 1980). Frederiksen (1980) attributed Cicatricosisporites to the schizaeacean genera Anemia or Mohria. Graham (1985) felt this grain resembled Ceratopteris in the Parkiaceae, although Fredericksen (pers. commun., 1995) felt that the ribs of Ceratopteris were much further apart than those of Cicatricosisporites. Although their exact affinities remain uncertain, the parent plants of Cicatricosisporites and Laevigatosporities were almost certainly herbaceous.

High percentages of Cicatricosisporites and Laevigatosporites in Lake Somerville lignite and detrital sediment samples may relate to the presence of clastic sediments in the wetland and rising water levels. Relatively high percentages of mineral matter occurred in fern-rich samples from the base of the middle seam (Gennett and others, 1990); and high percentages of Cicatricosisporites and Laevigatosporites from the lower zone of the upper seam occurred in fissile clastic-rich lignites. The clay parting in the upper seam also contained relatively high percentages of Laevigatosporites. The distribution of these spores in the Lake Somerville outcrop suggests that their parent plants grew in a freshwater marsh.

Cluster analysis of lignite and detrital sediment samples from the lower and middle seams at Lake Somerville based on their palynofloras suggest that the fern layer of the middle seam possesses a unique palynomorph association. In these analyses, the fern layer at the base of the middle seam formed one of four clusters at the 0.49 level (Gennett and others, 1990; Gennett, 1993).

Other workers have noted the relationship between Cicatricosisporites and Laevigatosporites and flooded, siliciclastic environments. In the Middle Cretaceous Dakota Formation, Cicatricosisporites most commonly occurred in marshy, lake shore environments (Farley and Dilcher, 1986). In the Maestrichtian Hell Creek Formation, Laevigatosporites sp. commonly occurred in abandoned channels and flood-basin marsh deposits (Kroeger, 1985). In the Early Eocene of Wyoming, Farley (1990) found the greatest concentration of most spores, including Laevigatosporites, within the levee-crevasse splay environment, which he attributed to the ability of ferns to grow rapidly in damp, disturbed environments. Similar horizons dominated by Cicatricosisporites and Laevigatosporites occurred at the base and in the middle of the D seam of the Eocene San Miquel lignite deposit, Atascosa and McMullen Co., Texas (Gennett, 1993).

Momipites coryloides, Arecipites colum-ellus and Liliacidites. The ornamentation, shape, and size of Momipites coryloides (finely scabrate porate, diameter 21 - 33 mm) as well as its abundance in Tertiary palynofloras, suggest a wind-pollinated parent plant (Frederiksen, 1980; Crepet, Daghlian, and Zavada, 1980). Crepet, Dilcher, and Potter (1975) and Crepet, Daghlian, and Zavada (1980) associated grains of similar shape to M. coryloides, but smaller diameter (mean diameter, 19.6 mm), with an Eocene catkin related to the extant Engelhardia-Oromunnea-Alfaroa complex, all of which are trees. Thus M. coryloides probably derived from a wind-pollinated tree in the Engelhardia-Oromunnea-Alfaroa complex. Although the affinities of the monocot taxon, Liliacidites, remain uncertain, Frederiksen (1980) and Gennett (1993) identified Arecipites columellus with the extant species Serenoa serrulata (saw palmetto), a low-growing palm.

The distribution of Momipites coryloides at Lake Somerville suggests a freshwater wetland plant. This species reaches peak abundance in the upper zone of the upper seam, and the middle of the lower and middle seams (22% - 69%). Lower percentages of M. coryloides occur at seam bases (2% - 34%) and at the top of the lower and middle seam (4% - 27%). Intermediate percentages occur in the marine argillaceous sandstones, which reflect the regional pollen rain (25% - 35%). Frederiksen (1981) also noted that this species reaches peak abundance in lignites.

An association between Momipites coryloides and either Liliacidites, or Arecipites columellus, or both occurs in lignite and detrital sediment samples from Lake Somerville (fig.1; figs. 2a,b; fig. 3). In cluster analyses of samples from the lower and middle seam on the basis of palynofloras, M. coryloides in association with A. columellus or Liliacidites or both formed one of four significant clusters (Gennett and others, 1990; Gennett, 1993).

Elsik (1978) and Mukhopadhyay (1989) identified lignites with low diversities of arboreal pollen and abundant Momipites in association with monocot grains (Calamuspollenites, Liliacidites, and Arecipites columellus) as marshes. This interpretation carries the assumption that Eocene marsh vegetation produced relatively little pollen and that trees growing outside the wetland, such as the parent plant of Momipites, contributed most of the pollen to 'marsh' lignites. Frederiksen (1985) argued against this interpretation because samples containing abundant wind-transported grains generally have high diversities. He felt that low diversities of arboreal pollen in Upper Eocene lignites indicated that those lignites accumulated in swamps. We found peltate leaf hairs attributed to a plant in the Engelhardia-Oromunnea-Alfaroa complex throughout the lignites of the Lake Somerville spillway, which further suggests that the parent plant of M. coryloides grew in peat-accumulating wetlands.

The Momipites coryloides-Arecipites columellus-Liliacidites association may indicate a swamp forest with an open canopy, especially if the parent plant of Arecipites columellus was a low-growing palm. Evidence that the parent plant of M. coryloides grew in the swamp coupled with the low overall diversity of the palynoflora at Lake Somerville (Gennett, 1993) argues against interpreting this association as a marsh.

Nyssa and Rhoipites angustus. Gennett (1993) identified the pollen taxon Nyssa with the extant arboreal genus of the same name, which is an insect-pollinated tree (Lewis and others, 1983) that grows in freshwater wetlands and forests of the Eastern United States. Although the taxonomic affinities of Rhoipites angustus remain uncertain, this grain probably derived from an insect-pollinated plant (Gennett, 1993; N. O. Frederiksen, written commun., 1988; Wodehouse, 1933).

Percentages of Nyssa > 5% indicate freshwater swamps. This percentage seems low; however small percentages of insect-pollinated taxa within a palynomorph assemblage indicate the presence of the parent plants in the wetland. The palynomorph assemblage of Nyssa-dominated communities from the Okefenokee contained only 7% to 15% Nyssa grains (Cohen, 1975).

In Lake Somerville lignites and detrital sediments, Nyssa and Rhoipites angustus have the distribution of freshwater wetland plants. Abundances of these taxa co-vary, particularly in lignite samples from the lower part of the lower seam (figs. 1a, b). Momipities coryloides dominates all samples that contained >5% Nyssa, which suggests that the three taxa grew together within the wetland. In cluster analyses of lignite and detrital sediment samples from the lower and middle seam based on their palynofloras, samples containing Momipites coryloides in association with Nyssa and R. angustus formed one of four clusters at the 0.49 level (Gennett and others, 1990; Gennett, 1993).

Cupuliferoipollenites, Cupuliferoidae-pollenites liblarensis and Siltaria cf. S. scabriextima. Frederiksen (1981) suggested that Cupuliferoipollenites derived from the extinct fagaceous genus, Dryophyllum and perhaps the extant fagaceous genera Castanea and Castanopsis. Nichols (1970) referred grains with similar morphology, which he called Cupuliferoipollenites cingulum, to Castanea. Crepet and Daghlian (1980) found pollen similar to Cupuliferoidaepollenites liblarensis in a castaneoid (Fagaceae) inflorescence from the Claiborne Group of Tennessee. Nearly all of the Fagaceae are trees or shrubs. Frederiksen (1981) noted a morphological gradient linking Cupuliferoipollenites spp. and Siltaria, but did not identify a probable parent plant for Siltaria; although Traverse (1955, p. 51) implied that Siltaria could belong to the Fagaceae. Gennett (1993) noted the resemblance of Siltaria to Agiceras, an extant mangrove taxon in the Myrsinaceae.

In the Lake Somerville spillway lignites, high percentages of Cupuliferoipollenites spp. appear related to the presence of marine sediments overlying the seams. Marine argillaceous sandstones overlie both the lower and middle seam, in which the high percentages of Cupuliferoipollenites occur. Additional evidence for the presence of brackish or marine water at the top of the lower seam includes the presence of brackish or marine diatoms and pollen grains identified as belonging to extant or extinct mangrove taxa. The upper seam, overlain by freshwater sediments, contains negligible percentages of Cupuliferoipollenites spp. and high percentages of Momipites (36% - 62%) and relatively high percentages of Ailanthipites berryi (11% - 12%).

In the lignites and detrital sediments of the Lake Somerville spillway, Cupuliferoidae-pollenites liblarensis and Siltaria cf. S. scabriextima have a similar distribution to Cupuliferoipollenites spp. (figs. 1, and 2). The distribution of all three taxa, particularly Cupuliferoipollenites spp., suggests that their parent plants grew in the wetland. The highest percentages of Cupuliferoipollenites spp., up to 63%, occur in lignite samples from the tops of the lower and middle seams; percentages in the overlying marine argillaceous sandstones, which reflect the regional pollen rain, ranged from 8% to 21%. Percentages of Cupuliferoipollenites spp. in lignite samples from below the top of the lower and middle seams ranged from 0 to 17%. In cluster analyses of lignite and detrital sediment samples from the lower and middle seam based on their palynofloras, samples from the top of the lower and middle lignite and the sample of argillaceous sandstone overlying the lower seam, which contain Cupuliferoipollenites ssp. in association with Cupuliferoidaepollenites liblarensis and Siltaria cf. S scabriextima, formed one of four clusters at the 0.49 level (Gennett and others, 1990; Gennett, 1993).

Because nearly all modern Fagaceae are shrubs or trees, the Cupuliferoipollenites-Cupuliferoidaepollenites liblarensis-Siltaria cf. S. scabriextima association probably indicates a forested or shrubby wetland. The distribution of these taxa in the lignites and detrital sediment samples of Lake Somerville spillway suggests that their parent plants could tolerate the presence of brackish or salt water. However, few, if any, modern fagaceous species grow in brackish or marine wetlands and although Siltaria may not belong in the Fagaceae, Cupuliferoidaepollenites liblarensis almost certainly does. Perhaps ancient fagaceous plants inhabited a wider range of habitats than their modern descendants. A better understanding of the paleoecology of this association requires detailed studies of the community paleoecology of other Late Eocene lignites from the Gulf Coast.

Caprifoliipites tantulus. Frederiksen (1980) noted a similarity between Caprifoliipites tantulus and pollen of the extant genus Viburnum (Caprifoliaceae), which, in North America, consists of small trees and shrubs (Harrar and Harrar, 1962). According to Frederiksen (1981), Caprifoliipites tantulus appeared insect-pollinated but could belong to a transitional insect- to wind-pollinated plant. Caprifoliipites tantulus occurs commonly within the Lake Somerville lignites and shows a distribution similar to Momipites coryloides, which suggests a freshwater swamp plant. Peak abundances of C. tantulus (up to 33%) occur in the middle of the lower and middle seams and in the upper zone of the upper seam. Lower abundances occur in the marine argillaceous sandstones (1% - 7%), which reflect the regional pollen rain; similar lower abundances (1% - 8%) occur in samples from the top of the lower and middle lignites, which may contain a salt-tolerant plant community. Frederiksen (1981) also found C. tantulus more common in lignites than in clastic samples of Jackson Age.

Although Frederiksen (1981) identified C. tantulus as 'possibly shrubby or herbaceous', its distribution in Lake Somerville lignites appears consistent with that of a swamp tree (either canopy or understory) or shrub. In the Lake Somerville samples, this grain does not appear to co-vary with the Momipites coryloides-Arecipites columellus-Liliacidites association, which may derive from swamp forest with an open canopy, or the Cicatricosisporites-Laevigatosporites association, which may derive from marsh vegetation (figs. 1-3).

Ailanthipites berryi. The shape (tricolporate with long furrows, conspicuous furrow and pore rims) and ornamentation (reticulate exine with rows of pits) of Ailanthipites berryi suggest an insect-pollinated plant (Wodehouse, 1933; Frederiksen, 1980; Gennett, 1993). Frederiksen (1980) suggested possible affinities of this grain with Anacardiaceae, Burseraceae, Leguminoseae, Sapindaceae and Simaroubaceae. Wodehouse (1933) assigned it to the extant genus Ailanthus (Simaroubaceae), a group of insect-pollinated trees (the 'tree of heaven' belongs to this genus) and noted its prevalence in the oil shales of the Eocene Green River Formation. Although the Simaroubaceae are primarily insect-pollinated, they may be facultatively wind-pollinated (Lewis and others, 1983).

The high percentage of Ailanthipites berryi (30%) in the upper, reworked portion of the volcanic ash that underlies the lower lignite suggests that the parent plant of this grain grew in the coastal plain prior to the formation of the wetland. This species also contributes over 10% of the pollen sum to lignite samples from the upper zone of the upper seam. Normally, this abundance of grains derived from an insect-pollinated plant would suggest that the parent plant grew in the swamp; however this reasoning may not hold for facultatively wind-pollinated plants. On the other hand, Ailanthipites berryi contributes only one or two percent of the pollen sum of the marine argillaceous sandstone samples overlying the lignites, which reflect the regional pollen rain. Thus, its abundance in the lignite samples from the upper seam may indicate that it grew in the swamp.

Palynoflora of the claystone layers and lenses. Two claystone layers, interpreted as altered volcanic ash (Yancey, this volume) occur within the middle lignite at Lake Somerville. Ruppert and others (1994) suggested a detrital origin for these layers because they consist of mixed layer clays rather than kaolinite and contain sub-angular and sub-rounded grains. However, the claystone layers do not contain an exotic palynoflora, which might be expected in a transported sediment: the palynomorph assemblages of the thin lower claystone match those of the surrounding lignite; assemblages of the variable upper claystone either match those of the underlying lignite, or appear transitional between the underlying and overlying lignite. One rare grain (<1%), Lycopodium heskemensis, occurs in six of eight middle seam claystone samples and does not occur in lignite samples from any seam. However, other rare grains that occur in more than one claystone sample from the middle seam occur in both lithologies.

The similarity of claystone and lignite profiles suggests that the palynoflora of the claystones in the middle seam derived from vegetation nearly identical to that growing in the swamp. Carbonaceous zones occur sporadically at the base of both claystone layers, which indicates some interaction with water during deposition. However, the similarity of claystone and lignite palynomorph profiles argues for a local origin for the claystone palynoflora. These sediments may not have been transported very far prior to deposition.

The claystone lens near the base of the lower lignite, interpreted as winnowed volcanic ash, has the most distinctive palynomorph profile of any of the interbedded claystones. Both the claystone lens and the surrounding lignites have the same dominant species (Momipites coryloides). However, Cupuliferoi-pollenites spp. reaches15% in the lens and Caprifoliipites tantulus reaches 8% in contrast to 1% - 2% in the surrounding lignites. The co-occurrence of abundance peaks in Cupuliferoipollenites spp. and C. tantulus, taxa which have dissimilar distributions in the lignites, suggests that the claystone lens in the lower seam contains a mixed palynoflora, and supports the interpretation of these lenses as winnowed, transported deposits.

Given the similarity of palynomorph profile above and below the thin lower claystone of the middle seam, the deposition of this layer did not greatly affect the wetland community. The position of the upper variable claystone layer in the middle seam coincides with the position of the Momipites-Cupuliferoipollenites transition, which raises the question of whether the deposition of this layer caused the change in dominant taxa. However, we view the deposition of the variable claystone and the Momipites-Cupuliferoipollenites transition as independent events because no claystone occurs in the lower seam at the level of the Momipites-Cupuliferoipollenites transition (fig. 1a).

CONCLUSIONS

FLORAL PALEOECOLOGY OF LAKE SOMERVILLE LIGNITES AND DETRITAL SEDIMENTS

The lower lignite at Lake Somerville replaced a coastal plain community dominated by shrubs or trees that produced Ailanthipites berryi, which probably grew on drained soils. Forest or shrub communities with the parent plant of Momipites coryloides, a juglandaceous tree or shrub belonging to the Engelhardia-Oromunnea-Alfaroa complex produced most of the lignite in the three seams. The Momipites coryloides-Arecipites columellus-Liliacidites association may have derived from an open canopy swamp or wetland shrub community; the Momipites coryloides-Nyssa-Rhoipites angustus association probably derived from a closed-canopy swamp.

Marine sediments overlie the lower and middle lignite beds. At the top of both these seams, a forest or shrub community with the parent plant of Cupuliferoipollenites spp., possibly related to the Fagaceae, replaced the Momipites coryloides plant and its associates within the mire. In addition to the Cupulifero-ipollenites spp plant, the parent plants of Cupuliferoidaepollenites liblarensis and Siltaria cf. S. scbriextima probably grew in this community. In the lignites and detrital sediments of the Lake Somerville spillway, the Cupuliferoipollenites-Cupulifero-idaepollenites liblarensis-Siltaria cf. S. scabriextima association appears to correlate with the presence of marine or brackish water: 1. This association predominates at the top of seams overlain by marine sediments; yet very low percentages of these taxa occur at the top of the upper lignite seam overlain by freshwater sediments; 2. This association co-occurs with brackish or marine dinoflagellates and palynomorphs attributed to extant mangrove genera in lignites from the top of seams overlain by marine sediments; lignite samples from the top of the upper seam overlain by freshwater sediments did not contain fossils indicative of brackish or marine water.

The correlation between brackish or marine indicators and the Cupulifero- ipollenites-Cupuliferoidaepollenites-Siltaria association suggests that salt-water influx into the wetland could account for the Momipites-Cupuliferoipollenites transition near the top of the lower and middle seams. If so, the ecological tolerances of the Fagaceae, which do not grow in marine or brackish water at the present time, must have been broader in the Eocene. Finally, increased water depth and clastic influx may have contributed to the change in dominant grains, from Momipites to Cupuliferoipollenites.

Both the middle and upper lignite seams formed on top of shore-zone sands. At the base of the middle seam, a unique fern-dominated assemblage occurs which is replaced by Momipites-dominated assemblages within 10 cm of the base of the seam. Given the propensity of ferns to colonize disturbed environments, the fern-Momipites transition at the base of the middle seam may reflect an environmental change in the wetland such that the disturbances required to maintain the fern community ceased.

Two intervals of volcanic ash deposition occurred during the accumulation of the middle seam. The palynoflora of these layers matches that of the surrounding lignites and appears locally derived. Although these layers show evidence of detrital transport (Ruppert and others, 1994), the transport distance may have been small.

An open canopy swamp or shrub wetland dominated by fern spores and Momipites coryloides formed the lignite at the base of the lower zone of the upper seam. These same palynomorphs joined by Arecipites columellus (saw palmetto) occur in the claystone interbed in the middle of the upper seam. This palynomorph profile could derive from an open canopy swamp, a shrubland, or a swamp-marsh complex; the claystone parting probably accumulated in fresh water. An open canopy swamp or wetland shrub community with the parent plant of M. coryloides and the monocot grain, Liliacidites, formed the lignite at the base of the upper zone. The parent plants of M. coryloides, Caprifoliipites tantulus (perhaps a small tree or shrub related to the extant genus Viburnum), and possibly Ailanthipites berryi (a shrub or tree related to the 'tree of heaven') contributed to the lignite in the middle and top of the upper zone of the upper seam. Conversely, the parent plant of Ailanthipites berryi may have lived on drained clastic substrates outside the wetland and contributed wind-borne pollen to the lignite. Freshwater siliciclastics overlie the upper zone of the upper seam.

REFERENCES CITED

Cohen, A. D., 1975, Peats from the Okefenokee Swamp-Marsh complex: Geosciences and Man, v. 11, p. 123-131.

Crepet, W. L., and Daghlian, C. F., 1980, Castaneoid inflorescences from the Middle Eocene of Tennessee and the diagnostic value of pollen (at the subfamily level) in the Fagaceae: American Journal of Botany, v. 67, p. 729-757.

Crepet, W. L., Daghlian, C. F., and Zavada, M., 1980, Investigations of Angiosperms from the Eocene of North America: a new Juglandaceous catkin: Review of Palaeobotany and Palynology, v. 30, p. 361-370.

Crepet, W. L., Dilcher, D. L., and Potter, F. W., 1975, Investigations of Angiosperms from the Eocene of North America: a catkin with Juglandaceous affinities: American Journal of Botany, v. 62, p. 813-823.

Elsik, W. C., 1978, Palynology of Gulf Coast lignites, the stratigraphic framework and depositional environments, in Kaiser, W. R., ed., Proceedings of the Gulf Coast Lignite Conference: Geology, Utilization and Environmental Aspects: Austin, TX, University of Texas Bureau of Economic Geology. Report of Investigations, No. 90, p. 21-32.

Farley, M. B., 1990, Vegetation distribution across the Early Eocene depositional landscape from palynological analysis: Palaeogeography, Palaeoclimatology, Palaeoecology, v. 79, p. 11-27.

Farley, M. B., and Dilcher, D. L., 1986, Correlation between miospores and depositional environments of the Dakota Formation (Mid-Cretaceous) of north-central Kansas and adjacent Nebraska, U.S.A.: Palynology, v. 10, p. 117-134.

Frederiksen, N. O., 1980, Sporomorphs from the Jackson Group (Upper Eocene) and adjacent strata of Mississippi and western Alabama: U.S. Geological Survey Professional Paper 1084, 75 p.

Frederiksen, N. O., 1981, Middle Eocene to Early Oligocene plant communities of the Gulf Coast, in Gray, J., Boucot, A. J., and Berry, W. B. N., eds., Communities of the Past: Stroudsburg, PA, Hutchinson Ross, p. 493-550.

Frederiksen, N. O., 1985, Review of Early Tertiary Sporomorph Paleoecology: American Association of Stratigraphic Palynologists Contribution Series, No. 12, 109 p.

Gennett, J. A., 1993, Palynology and paleoecology of the San Miguel lignite deposit of Late Eocene age, South Texas: College Station, Texas, Texas A&M University, Ph.D. dissertation, 490 p., 121 figs.

Gennett, J. A., Raymond, A., and Wong, C., l990, Distinguishing regional floras from local peat-forming communities: An example from two Upper Eocene lignite beds in Washington Co., Texas. Geological Society of America Abstracts with Programs, v. 22, no. 7, p. A201.

Graham, A., 1985, Eocene communities of Panama: Annales of the Missouri Botanical Garden, v 56, 504-534.

Harrar, E. S, and Harrar, J. G., 1962, Guide to Southern Trees: New York, Dover Publications, Inc., 709 p.

Kroeger, T.J., 1985, Paleoenvironmental significance of Upper Cretaceous palynomorphs in the Hell Creek Formation, Butte County, South Dakota: Rapid City, South Dakota, South Dakota School of Mines and Technology, 143 p.

Lewis, W. H., Vinay, P. and Zenger, V. E., 1983, Airborne and allergenic pollen of North America: Baltimore, MD, Johns Hopkins University Press, 245 p.

Mukhopadhyay, P. K., 1989, Organic petrography and organic geochemistry of Texas Tertiary coals in relation to depositional environment and hydrocarbon deposition: Austin, TX, The University of Texas Bureau of Economic Geology, Report of Investigations, No. 188, 118 p.

Nichols, D. J., 1970, Palynology in relation to depositional environments of the Wilcox Group (Early Tertiary) in Texas: University Park, Pennsylvania, The Pennsylvania State University, Ph.D. dissertation, 467 p.

Ruppert, L. F., Warwick, P. D., Crowley, S. S., and Pontolillo, J., 1994, Tonsteins and clay-rich layers in coal-bearing intervals of the Eocene Manning Formation, East-Central Texas: Transactions of the Gulf Coast Association of Geological Societies, v. 44, 649-656.

Traverse, A., 1955, Pollen analysis of the Brandon Lignite of Vermont of Lower Tertiary age: U. S. Bur. Mines Rep. Invest. 5151.

Wodehouse, R. P., 1933, Tertiary pollen II: The Oil Shales of the Eocene Green River Formation: Torrey Botanical Club Bulletin, v. 60, p. 470-524.

Chapter 2.

Chapter 4.

Created by the EERT WWW Staff.