LIGNITE DEPOSIT OF LATE EOCENE AGE, SOUTH TEXAS
Volume II
A Dissertation
for the degree of
Approved as to style and content by:
_________________________ _________________________
Anne Raymond Vaughn Bryant, Jr.
(Co-Chair of Committee) (Co-Chair of Committee)
_________________________ _________________________
Christopher Mathewson Elenor Cox
(Member) (Member)
_________________________ _________________________
Robert Stanton John Spang
(Member) (Head of Department)
A Dissertation
for the degree of
Approved as to style and content by:
_________________________ _________________________
Anne Raymond Vaughn Bryant, Jr.
(Co-Chair of Committee) (Co-Chair of Committee)
_________________________ _________________________
Christopher Mathewson Elenor Cox
(Member) (Member)
_________________________ _________________________
Robert Stanton John Spang
(Member) (Head of Department)
iii
ABSTRACT
Palynology and Paleoecology of the San Miguel Lignite
Deposit of Late Eocene Age, South Texas. (December 1993)
Co-Chairs of Advisory Committee: Dr. Vaughn Bryant, Jr.
Dr. Anne Raymond
Eight pollen and spore sequences from the high ash, high sulfur San Miguel lignite deposit indicate a Late Eocene, Jackson age. In general, the sequences follow a pattern, with high Nyssa kruschii and Rhoipites angustus percentages at the base, and high Cupuliferoipollenites sp. percentages at the top; the presence of Nyssa (tupelo) strongly suggests a freshwater swamp. Cluster and reciprocal averaging analysis reinforce these patterns, and suggest two vegetational-environmental gradients. One of these involves N. kruschii/R. angustus and Cupuliferoipollenites. The second involves these taxa in opposition to fern spores such as Laevigatosporites. It is not certain whether the gradients represent progressively shallower, deeper, or more saline environment. Similar palynomorph sequences were also recovered from the Momipites coryloides-dominated Late Eocene Lake Somerville lignites. The Lake Somerville lignites may represent a cyclic increase in depth and salinity, and the San Miguel lignites may have had an analogous depositional history.
No strong correlation was found in the San Miguel lignites between ash, sulfur, and palynomorph content.
iv
Analysis of palynomorph samples from horizontal sequences indicates a great deal of lateral variability, but the use of reciprocal averaging and running averages show that vertical variability is greater and that, although the data is "noisy," vertical changes in palynomorph content are valid and meaningful.
v
DEDICATION
For
Emma, Ian, and Erin
vi
ACKNOWLEDGMENTS
This study was funded in part by the Center For Energy and Mineral Resources. The San Miguel cores were provided by Sam Gowan; thanks also go to the drill crew including at various times Sam Gowan, Dave White, Stephanie Shelvey, Harlan Jennings, Glen Lowenstein, and Scott Armstrong. My field assistants at the Lake Somerville site were Emma Day-Gennett, Joshua Schoen, and John Jones. The Yegua sample from 29th Street in Bryan was collected by Chuck Thornton. Richard Day drafted Text-Figures 1-3, the "pollen diagrams," and the composite multivariate plots, using Generic CADD. Multivariate programs were provided by Warren Kovach and Bill Parker.
vii
Page
INTRODUCTION 1
Historical Perspective 1
Introduction 2
CHARACTERISTICS OF THE DEPOSIT 4
The San Miguel Deposit 4
The Lake Somerville Deposit 7
PREVIOUS RESEARCH 10
Geology and Paleontology of the San Miguel Lignites 10
Palynology 11
Paleoecology 14
METHODS 16
SYSTEMATIC PALEONTOLOGY 35
BIOSTRATIGRAPHY 138
Lake Somerville Spillway 139
The San Miguel Deposit 140
Stratigraphy and Biostratigraphy 140
Dominant Taxa 141
Indicator Taxa 145
Conclusion 148
DESCRIPTION OF PALYNOMORPH SEQUENCES 150
C Seam 150
Sequence A 150
Sequence B 152
Sequence C 152
Sequence D 153
Sequence E 154
Sequence F 154
Sequence G 155
Sequence H 155
D Seam 156
Sequence B 156
Sequence C 156
Sequence E 159
Sequence F 160
Sequence G 160
Other Seams 161
Sequence B, A Seam 161
Sequence E, A Seam 161
iv
Page
Sequence E, B Seam. 161
Sequence E, E Seam 161
Sequence F, E Seam 162
The Overburden 162
Sequence B, Overburden. 162
Sequence E, Overburden 162
Partings 162
Sequence B, A/B Parting 162
Sequence B, B/C Parting 163
Sequence E, B/C Parting 163
Sequence B, C/D Parting 163
Sequence E, D/E Parting 163
Lake Somerville 163
Lower Seam 163
Clastic Sample 165
Upper Seam 165
STATISTICAL ANALYSIS 166
Multivariate Analysis: Cluster Analysis 166
Introduction 166
Analysis and Results 166
Multivariate Analysis: Reciprocal Averaging (Correspondence Analysis) 181
Introduction 181
Analysis and Results 183
Comparison of Reciprocal Averaging with Cluster Analysis 189
Multivariate Analysis of Levels (Seam C) 191
San Miguel Sequences 191
Composite Diagram 194
Lake Somerville 196
Discussion 198
Efficacy of the Multivariate Technique 199
Horizontal Sequences 199
Sampling Methods. 200
Methods and Results: Introduction. 201
Methods and Results: Confidence Intervals 201
Methods and Results: Multivariate Analysis 208
Methods and Results: Running Averages 210
Discussion 228
Diversity 230
Introduction 230
Methods and Results 230
Palynomorph Concentration 250
Introduction 250
Methods and Results 250
FORMS OF SULFUR ANALYSIS 258
Introduction 258
ix
Page
Results 264
PALEOECOLOGY 277
Anchors 278
Suggestors 281
"Minor Taxa" 285
Taxon Summary 286
The History of the San Miguel Deposit and the Nature of Wetlands 288
Scenario One: Shallowing 288
San Miguel 288
Other Seams. 291
Partings 291
Lake Somerville 292
Scenario Two: Deepening 293
Scenario Three: Mangroves 295
Paleoecological Conclusions 297
CONCLUSIONS 301
REFERENCES CITED 303
SUPPLEMENTAL REFERENCES 328
APPENDIX 1: DESCRIPTION OF UNKNOWNS 329
APPENDIX 2: MULTIVARIATE DIAGRAMS 338
APPENDIX 3: DATA 384
VITA 490
xix
Plate
1. Major Taxon Palynomorph Diagrams. . . . .(in pocket)
2. Complete Palynomorph Diagrams of A, B,
and C Sequences . . . . . . . . . . . . .(in pocket)
3. Complete Palynomorph Diagrams of D, E,
and F Sequences . . . . . . . . . . . . .(in pocket)
4. Complete Palynomorph Diagrams of G, H,
and Somerville Sequences. . . . . . . . .(in pocket)
INTRODUCTION
Historical Perspective
Fossil pollen and spores (palynomorphs) have often been used as tools to obtain paleoenvironmental information contained in sediments. Although palynomorphs in clastic sediments are useful in determining paleoclimates, clastic palynomorph spectra reflect a large regional, wind and water transported component. Palynomorph suites from coals probably represent a more local flora, and variation in these assemblages can be interpreted to show changes in the vegetation near the site of deposition.
The first palynomorph studies of lignites were done on the middle Eocene German Braunkohles by Potonié (1934), and later, Pflug (1950). Traverse's paper on the Oligocene Brandon lignites of Vermont, published in 1955, was the first American paper in this field. Traverse presented a pollen diagram from both the lignite and surrounding clastics, and made both water depth and climatic interpretations from these data. He favorably compared the Brandon paleoswamp to wetlands in present-day southeastern North America. More recent papers describing North American Tertiary (excluding Texas) lignite palynofloras include Miocene studies by Rachele (1976), Taggert and Cross (1980); Cross and Taggart (1983), Piel (1977), Martin and Rouse (1966); Yeend (1977); Wolfe et al. (1966); an Oligocene study by __________
The format employed by Palynology is used for this dissertation.
Piel (1971); Eocene papers by Newman (1981), Hopkins (1967), and Hopkins and Sweet (1977); and Paleocene studies by Martinez and Urban (1971), Anderson (1960),
Moore and Urban (1975), Leffingwell (1971), Spindel (1975), Kremp et al. (1961), Trotter (1963) and Griggs (1970).
Introduction
This study details the pollen and spore composition of the Texas Eocene San Miguel lignites and the spatial changes that occur in this composition. There are several reasons for this undertaking. In the broadest sense, a general overview of the flora can provide a comparison with other palynofloras from other environments and ages, particularly from the Gulf Coast. Within lignites the type of vegetation indicated by the palynomorphs gives clues to the climate, the communities which formed the deposits, and by inference, the depositional environment of the deposit. Both vertical and horizontal variations in palynomorph composition indicate changes in environment that are often not elucidated by other methods. Changes in palynomorph content may be correlated with other variables of interest to coal geologists, such as coal quality. These changes may be a result of biological succession, which would extend our understanding of Eocene paleocommunities. The flora itself is a clue to a more precise date for the San Miguel lignites and associated units. In turn a compilation of the taxa present at San Miguel will enlarge the biostratigraphic data available to other researchers attempting to date Gulf Coast sediments. A study of the present-day ranges of palynomorph taxa assignable to modern genera and species ranges will aid in the study of the regional climate during the Eocene.
The San Miguel Deposit
The San Miguel Lignite is a low quality but extensive deposit located in northeastern McMullen and southern Atascosa Counties, Texas (Text-Figure 1). Lignite from the San Miguel deposit averages 32% moisture, 25% ash, 2.0 % sulfur, and 5,318 Btu/lb. (L. Dick, written commun., cited in Ayers, 1986). It is Middle to Upper Eocene in age, and is referred to as "Yegua-Jackson" lignite because of the difficulty in establishing the Middle-Upper Eocene boundary in South Texas (Ayers and Kaiser, 1986).
In the area of this study, five major seams and several riders comprise the San Miguel. Gowan (1985) designated these seams, from the top down, as A, B, C, D, and E. These seams are separated by silty partings. At Gowan's Highwall section B, the A-B parting averages 0.25 m (.81 ft.) thick, the B-C averages 0.24 m (.78 ft.), and the C-D averages 0.30 m (.97 ft.). The total thickness of these three partings ranged to over 1.2 m (4 ft.) in Gowan's study area. Only the C and D seam are substantial in thickness. Gowan gave coal core thicknesses for the C seam of 0.84 to 0.95 m (2.75 to 3.1 feet). The thicknesses increases to the southeast. His core thickness for the D seam ranged from .85 m (2.8 ft.) to .98 m (3.2 feet).
The San Miguel lignites formed in what was, during the Late Eocene, part of the South Texas Strandplain-
Barrier Bar System of Fisher et al. (1970). Text-Figure 2 shows both the barrier bar system and the Fayette
Text-Figure 1. Location of the San Miguel and Lake Somerville lignites.
Text-Figure 2. Depositional systems in the Jackson Group of Texas during the Late Eocene.
Fluvial-Delta System, which contributed clastic sedimentto the former barrier bars.
The exact mode of formation of these lignites has been controversial. According to Snedden (1979) and Snedden and Kersey (1981), these lignites formed in coastal wetlands at the landward side of a lagoon; partings in the lignite represent "fire-splays," formed by the rupture of river levees by the burning-off of surrounding peats during dry periods. A modern analogy is described by Staub and Cohen (1978) for the Snuggedy Swamp of South Carolina. Snedden's theory implies a fluvial source for the clastic partings. Gowan (1985), on the other hand, believed that a freshwater swamp formed behind a closed barrier bar. Temporary breaches in the bar brought in marine sediments, and saltmarshes replaced freshwater swamp vegetation. Coal petrographic work by Mukhopadhyay (1989), based largely on channel samples, indicated that the B and C seams were deposited in freshwater swamp-marsh environments because of the "...high ulminite content and low humodetrinite and liptodetrinite content..." (p. 78). He believed the A and D seams to have formed in saline marshes because of the "...unbanded and nonxylitic lithotype, rich humodetrinite and liptodetrinite," and "sparse marine phytoclasts (dinoflagellates)..." (p. 78).
The Lake Somerville Deposit
The lignite beds at Lake Somerville (Text-Figure 1) may correlate with seams of the Manning Formation mined at the Gibbons Creek Mine in Grimes County, Texas (Yancey and Davidoff, 1991; T. Yancey, oral commun., 1993). The three lignite seams are separated widely by siliciclastic deposits. The lower lignite occurs upon an erosion surface (paleosol) on top of a friable sandstone which overlies volcanic sediments (K. Phillips, oral commun., 1993). The lowest seam is 0.2 to 1.0 m thick, locally grades to a carbonaceous shale, and has a burrowed top (Yancey and Davidoff, 1991). Yancey (1992) also found a thin ashy zone near the base. This lower seam is separated from the middle lignite by silty shale overlain by sandstone. Yancey believes these burrowed shales, which contain marine diatoms (K. Phillips, oral commun., 1993), to represent flooding of the lignite with standing water. The sandstone is less burrowed and contains muddy interzones. Planar cross-beds appear near the top of the sand. Yancey and Davidoff (1991) described the middle lignite as "...1.0 m thick in the middle of the outcrop, and thickening toward the dam, with a variable mud content. The lignite contains a thin interbed of tuff near the top [10 cm. from the top] and the top is penetrated by small burrows filled with muddy sand from the overlying bed. Layers within the lignite and sands underlying the lignite are cut with root traces." (p. 101) Yancey (1992) described a similar sequence of shales and sandstones separating the middle lignite (referred to as the "upper lignite" in this paper) from an uppermost lignite layer.
The Lake Sommerville spillway lignites are encompassed in the Fayette Paleodelta system of Fisher et al. (1970) (Text-Figure 2). Gilmore (1981) believed that the lower lignite was "probably laid down in a low area between distributary channels [i.e., blanket peat] (p. 45)", whereas the upper lignite of this study may have been a "channel fill" lignite. Yancey and Davidoff (1991) hypothesized deposition by periodic flooding into coastal plain swamps or lacustrine basins. They believed the regional setting to be "..a coastal plain which encompassed large fluvial channels separated by low energy plains with swamps and lake environments..." and "...part of a highstand systems tract." T. Yancey (oral commun., 1992) further explained that the lignite deposition was contained in a paleotopographic low. Water depth increased during lignite deposition, with the peak transgression occurring in the overlying shales. The paleosols at the top of the sands represent periods of maximum regression.
Mukhopadhyay (1989) investigated the coal petrology of the Lake Sommerville lignites. He considered the two lower seams to have been derived from "...blanket peats, primarily containing marsh vegetation, which was deposited at the transition of the upper and lower delta plain (p. 78)." Only one channel sample was examined from each seam in Mukhopadhyay's study.
Megafossils are found in abundance in clastic layers that occur higher in the section at Welch's park. Specimens collected by A. Raymond, R. Murry, and C. Thornton included seeds similar to those of modern Nyssa (A. Raymond, oral commun., 1987) and leaves similar to those of Oreoroa claibornensis Dilcher and Manchester 1986.
Geology and Paleontology of the San Miguel Lignites
Because of San Miguel's economic importance, numerous studies have been initiated on the deposit. Maxwell (1962) considered the deposit to be part of the Upper Yegua Formation. Kaiser (1974) referred to lignite of commercial thickness in McMullen and Atascosa Counties. He believed these lignites to be of Yegua-Jackson age. High ash and sulfur values, and low Btu/lb led Kaiser to conclude that the South Texas lagoonal lignites were the poorest quality lignites in Texas.
Fisher et al. (1970) located the San Miguel lignite in the South Texas lagoonal-coastal plain system. McNulty (1978) summarized available knowledge of the San Miguel, as well as presenting simplified cross sections of the seams.
Little previous research has been done on the palynology of the San Miguel. Mukhopadhyay (1989) selectively and qualitatively reported palynomorphs present in five channel and five lithotype samples from four San Miguel seams. In accordance with the views of Elsik (1978), he believed that Momipites was "...often associated with the marsh community (p. 58)" and used Nyssapollenites, Alangiopollis, Symplocoipollenites, and Sphagnum as swamp indicators. In seams B and D, "marsh" pollen was more abundant than "swamp" pollen. Nyssa and minor Engelhardia were observed in seams C and D. Mukhopadhyay noted that the San Miguel contained less Momipites than East Texas lignites of the same age.
Gowan (1985) reported mostly dicotyledonous woody fragments from the San Miguel coal seams. In addition, Gowan found leaf impressions, which he believed to represent both monocotyledonous and dicotyledonous plants. Lath shaped plant fragments present in the partings were believed to belong to monocotyledonous taxa, notably, "...grasses, sedges, and reeds..." with a few Typha (cattail) leaves. Gowan also described pyritized roots.
The coal petrology of the ten San Miguel lignite samples led Mukhopadhyay (1986, 1989) to believe that the "...vegetation was dominantly of a reed marsh type (1986, p.139)." Dinoflagellates were observed in the D seam samples. This evidence led Ayers (1986) to consider the San Miguel to have formed in a "...fresh-water reed/marsh complex (p. 57)". Mukhopadhyay (1989) believed seams A and D to have been deposited in a marsh environment behind a barrier bar and that seam B and C peats formed in a fresh-water swamp-marsh barrier-bar environment.
Palynology
The Eocene palynomorph taxonomy and biostratigraphy of the Gulf Coast area have been described in papers by Engelhardt (1964), Fairchild and Elsik (1969), Tschudy (1970), Tschudy and Van Lonen (1970), Elsik (1974), Elsik and Dilcher (1974), Saunders et al. (1974), Martinez-Hernández et al. (1980), Jones and Gennett (1991) and Frederiksen (l980a, 1988). Details of these papers are given in Tables 1 and 2.
Table 1. Palynological characteristics of Claiborne lignites.
Site & Author Stratigraphy Samples Dominant Taxa Noteable Taxa
Palofax Mine Bigford Fm. 1 "Unidentified pollen Platycarya
(Webb Co., TX) Early Claiborne referred to as Nudopollis
Mukhopadyay (1989) Tricolpites,
Tricolporites,
Nudopollis terminalis
San Ignacio Bigford Fm. 2 tricolpates and Aesculidites
(Tamaulipas, Mexico) Early Claiborne tricolporates
Martinez-Hernandez and
others (1980)
Columbia Bigford Fm. 5 tricolpates and Nudopollis
(Nuevo Laredo, Mexico) Early Claiborne tricolporates Aesculidites
Martinez-Hernandez and
others (1980)
Lake Casa Blanca Laredo Fm. 1 Not Given Spinozonocolpites
(Webb Co., TX) Middle Claiborne (Nypa)
Westgate and Gee (Marginal Marine)
Panola Co., MS Middle Claiborne 1 Cupuliferoidaepollenites Nudopollis
Frederiksen (1981) Quercoidites absent
microhenricii
Tate Co., MS Middle Claiborne 1 Cupuliferoipollenites
Frederiksen (1981) Siltaria
Quercoidites
microhenricii
Miller Clay Pit Cockfield Fm. 27 Cupuliferoipollenites Nudopollis
(Henry Co., TN) Late Claiborne (with palms at top of Aesculidites
Potter (1976) (Oxbow Lake) seam) Anacolosidites
Hinds Co., MS Cockfield Fm. 1 Momipites Nudopollis
Frederiksen (1981) Late Claiborne absent
Madison Co., TX Yegua Fm. 3 Momipites Amanoa
Elsik (1978) Late Claiborne Nyssa
(Dominantly Fluvial)
Carter Creek Yegua Fm. 1 Rhoipites angustus Nudopollis
(Brazos Co., TX) Late Claiborne Momipites, Nyssa absent
This Study Liliacidites
Table 2. Palynologial characteristics of Jackson lignites.
Site & Author Stratigraphy Samples Dominant Taxa Noteable Taxa
Fayette Co., TX Manning Fm. 3 Momipites Nudopollis
Frederiksen (1981) Caprifoliipites absent
Walker Co., TX Manning Fm. 2 Momipites Nudopollis
Frederiksen (1981) Caprifoliipites absent
Gibbons Creek Manning Fm. 9 Momipites, sometimes Nudopollis
(Grimes Co., TX) with Caprifoliipites absent
Frederiksen (1981) Cyrillaceaepollenites
ventosus,
Cupuliferoipollenites
Brazos and Manning Fm. ? Momipites, Nyssa,
Grimes Co., TX sometimes C. ventosus
Elsik (1978)
Miguel Aleman Jackson Group ? Momipites Cupuliferoipollenites
Tamaulipas, Mexico Cicatricosisporites present
Martinez-Hernandez "Tricolporopollenites"
and others (1980)
Lake Somerville Manning Fm. 20 Momipites, sometimes
Washington Co., TX Cupuliferoipollenites
This Paper Cicatricosisporites
Paleoecology
Several attempts have been made to correlate Paleogene Gulf Coast palynomorphs with paleoenvironment. Gray (1960) made preliminary attempts to classify the ecological origin of Middle Eocene palynomorphs from Alabama. Nichols (1970) and Nichols and Traverse (1971) correlated broad depositional systems with palynomorph content in Paleocene Texas lignites; their central Texas tidal lagoon pollen spectra were characterized by chenopod (saltbush) pollen. Potter (1976) described a pollen sequence from an inland ox-bow lake from the Middle Eocene of Tennessee.
is lignite seam was characterized by low pollen diversity and was dominated by Cupuliferoipollenites, which he believed to have been derived from wetland plants. Using these data, he was able to determine floral changes in this unit. High diversity in clastic samples was accompanied by large numbers of Cupuliferoidaepollenites grains, which he interpreted as fluvially transported from nearby dry land trees. Elsik (1978), in a review of the palynology of Texas Paleogene lignites, attempted to differentiate swamp and marsh floras. He thought that marsh vegetation was represented by the palm genera Liliacidites and Calamuspollenites-Arecipites, as well as sedge (Carex). Frederiksen (1981) analyzed samples from a number of Eocene and Oligocene Gulf Coast lignite and shale samples and was able to contrast floras from various general environments. His brackish water sediment samples yielded essentially a freshwater wetland pollen flora. Mancini (1981, 1983) used the absence of Nyssa and Taxodium pollen to support his hypothesis that the Paleocene Oak Hill Lignite originated as a deltaic coastal marsh. Gee and Westgate (1989) hypothesized mangrove swamps on the Middle Eocene lower Texas Gulf Coast from the presence of Nypa in lignite samples.
Frederiksen (1985) also presented a general review of Early Tertiary plant paleoecology in which he outlined communities and hypothesized environmental requirements of selected taxa. He believed that plant successions inferred from lignite seams were probably not autogenic, but were likely caused by changes in water level.
As has been discussed in the introductory section, detailed studies of palynomorphs can yield more detailed and useful information about lignites than the cursory palynological studies often employed. In this study, I used methods pioneered by Quaternary palynologists in order to acquire this detailed information. One of these methods, also used by Traverse (1955) among others, was sequential counts. These counts allow the detection of a sequence of paleofloras at the sample site. Several cores were used to compare these sequences in different areas of the swamp. Pollen concentrations were obtained in order to compare the actual amount of various palynomorphs in the different samples of lignites. Confidence intervals and running averages were computed in order to determine the importance of variation in these sequences. Other statistical techniques, such as multivariate analysis and diversity were "borrowed" from other types of paleontology. These methods sometimes "point out" features in the data not readily discernable by more subjective analyses.
Pollen and spores were analyzed from sections of seven cores through the San Miguel lignite, as well as from a seam in a highwall section (Text-Figures 3-11). Cores SM-66 OB (Sequence A in this study), SM-210 OB (Sequence B), and SM-303 OB (Sequence C) were drilled by Beacon Construction Co. previous to this project. Cores SM-1 (Sequence D), SM-3 (Sequence E), SM-4 (Sequence F), and SM-5 (Sequence G) were drilled during the spring of 1985 using the Texas A&M University Mobile Drill rig. The latter sites were chosen by S. Gowan and traversed a thickening in the B seam. Samples from one high wall, Highwall A of Gowan (1985) (sequence H) were collected in the fall of 1984. The eight sections form a southwest-northeast traverse across the deposit, with Sequences C through G most closely spaced.
A vertical series of samples was removed from the cores and highwall at 10 cm intervals. A continuous sequence of samples, including partings and some overburden was analyzed for cores 210 and SM-3 (Sequence E). Because of poor recovery of palynomorphs from the partings and seams A and B, further processing involved seams C, D, and E only. Although all cores contain the C seams, cores SM-66 0B (Sequence A) and SM-1 (Sequence D) lack the D seam due to poor core retrieval. The E seam is present only in the SM-3 and SM-4 (Sequences E and F) cores.
Sampling from the Highwall involved the C seam only; the base of the D seam was below the base of the pit. Horizontal samples were taken at stations 6m (20 feet, as measured by Gowan, 1985) apart along the base of the seam and along an ash layer within the seam. Ten horizontal and seven vertical samples were taken at ten cm intervals at Station 7.
Text-Figure 3. Locations of cores and highwall section in the San Miguel Lignite.
Text-Figure 4. Location of Sequence A Samples in core SM-66 OB.
Text-Figure 5. Location of Sequence B Samples in core SM-210 OB.
Text-Figure 6. Location of Sequence C Samples in core SM-33 OB.
Text-Figure 7. Location of Sequence D Samples in core SMR-1.
Text-Figure 8. Location of Sequence E Samples in core SMR-3.
Text-Figure 9. Location of Sequence F Samples in core SMR-4.
Text-Figure 10. Location of Sequence G Samples in core SMR-5.
Text-Figure 11. Location of Sequence H and horizontal samples in Highwall A.
Vertical samples were also taken at 10 cm intervals from the bottom two lignite seams cropping out at the Lake Somerville spillway in Washington County, Texas (Text-Figure 12). Within the intervening shales and sands, samples were taken at 1 m intervals; only one of these samples yielded pollen.
Samples were processed using standard techniques (Doher, 1980). Core 210 was processed by EXXON in Houston, TX. Samples were treated with Hcl to remove carbonates, HF to remove quartz, HNO3 to oxidize lignin, NH3OH to oxidize humic compounds, sonication using darvan to remove the remaining organic material, and heavy liquid separation using ZnBr to remove the remaining mineral material. Slides were mounted in castolite, a rigid mounting medium. The SMR-3 (Sequence E) core was processed in the same manner in the Palynology Laboratory at Texas A&M, and samples were mounted, as were samples from succeeding cores, in silicone oil. The use of silicone oil allowed the palynomorphs to be rolled and examined from various angles, making possible the identification of a higher proportion of the palynomorphs than did the use of a rigid medium. Samples from other cores consisted of lignite only and heavy liquid separation was deemed unnecessary and therefore omitted. Most samples from these cores were processed in the Paleontology Prep Room, Geology Dept., at Texas A&M.
For most samples, a minimum of 200 pollen and spore (excluding Lycopodium) grains were counted. Spores were included in the count because of their importance in wetland ecology; the history of the wetland that formed the San Miguel lignite was the object of study, not the history of the regional flora. If one slide yielded less than 25 grains, counting was discontinued. If,
Text-Figure 12. Location of samples in the Lake Somerville outcrop.
after three slides, a count of 200 was not attained, aminimum of 150 grains was counted. "Indeterminable" grains could not be identified because of poor condition and were not included in the palynomorph sum. Unidentifiable prolate, tricolporate grains with a long axis measuring approximately 15 µm were very common in some samples from core SMR-3 (sequence E). These grains were designated as "degraded small tricolporates" and were included in the pollen sum. "Unknown" grains were in good condition, but could not be assigned to previously named taxa; these grains were not included in the pollen sum. These grains are described in Appendix 1. Fungal spores were also counted outside of the palynomorph sum.
Controversy exists as to the ideal number of pollen grains to count. Rull (1987), after reviewing the literature, stated: "The general opinion is that each situation requires counting a different number of grains and no standard size can be fixed....an arbitrary lower limit of 200-300 grains is the most common criterion." (p. 471) Rull used the point where the slope, (i.e. differences between adjacent percentages) in the width of the 95% confidence interval approaches zero as an indicator of reproducibility. He considered a 200 grain count adequate above this number there is no significant shift in the width of the confidence intervals. This relatively low figure is not adequate, however, to accurately represent true percentages of uncommon grains; for taxa whose true percentage is 1% of the sample, it would be necessary to count 1620 grains. A total of 36,221 pollen grains were tallied from the San Miguel lignites and 4086 from the Lake Somerville lignites.
Moore and Webb (1978) demonstrated that a count of 150 grains is accurate for taxa comprising 20% or more of the pollen sum.
Text-Figures 13-19 show the effect of the number of grains counted on the size of the 95% confidence interval of various counted percentages. For example, if a count of 50 grains is made, the true value of a 5% counted value for a specific taxon has a 95% likelihood of being between 1.6% and 14.8% (Text-Figure 14). If 200 grains are counted, then this interval becomes 2.8%
and 9.0%; if 400 grains are counted, then these values become 3.2% and 8.0%. In this case, a count twice as large would be necessary to narrow the 95% confidence interval by 2%. There is little available information on the actual meaning of the percentage values of Early Tertiary palynomorph counts, and consequently the precision attained in a count of 200 grains was deemed adequate in this study.
A tablet of Lycopodium "spike" was added to the final two cores to be processed, SM-4 (Sequence F) and SM 303-OB (Sequence C). The "spike" was added in order to determine pollen concentration, an estimate of the actual number of palynomorphs in a given volume of sediment. Additional interpretations can be made from these estimates, especially if a time frame is available for sediment deposition. Lycopodium was used because it is easy to identify and distinct among the Tertiary pollen and spores in these samples. These Lycopodium spores were counted outside of the palynomorph sum.
Text-Figure 13. 95% confidence intervals for counted percentages of 1% to 3%.
Text-Figure 14. 95% confidence intervals for counted percentages of 4% to 6%.
Text-Figure 15. 95% confidence intervals for counted percentages of 1%, 5% and 20%.
Text-Figure 16. 95% confidence intervals for counted percentage of 0%.
Text-Figure 17. 95% confidence intervals for 1% to 20% counting 200 grains.
Text-Figure 18. 95% confidence intervals for 21% to 40% counting 200 grains.
Text-Figure 19. 95% confidence intervals for 41% to 60% counting 200 grains.