- UW Department of Geology and Geophysics
The Hanna Basin is a relatively small foreland basin in south-central Wyoming containing a combined thickness of roughly 38,000 ft (11.5 km) of Upper Cretaceous and Palecene strata. Amber occurs in the Hanna Basin in carbonaceous to lignitic strata, representing fluvial and paludal episodes bounded by incursions of epicontinental seas. Amber occurs, in decreasing age, in the Upper Cretaceous Allen Ridge, Medicine Bow, and Ferris formations (parts of the last straddle the Cretaceous–Tertiary boundary), as well as in the Paleocene Hanna Formation. Because of the extraordinary thickness, unequivocal stratigraphic superposition, and long-lived deposition of Upper Cretaceous and Paleocene amber-bearing strata in the Hanna Basin, a unique opportunity has been provided for integrated study of taxonomic sources, deposition, and taphonomic alteration of ancient resins.
In all relevant Cretaceous and some Paleocene outcrops the amber is preserved mostly as small (4–8 mm diameter) droplets, often highly weathered and oxidized. One site in the Hanna Formation has yielded abundant, large pieces of transparent amber. Composition of samples analyzed by pyrolysis/gas chromatography-mass spectroscopy (PyGC-MS) indicates a common taxonomic source for amber from the Allen Ridge, Medicine Bow, and Hanna formations. The taxonomic source of amber from one part of the Ferris Formation, in contrast, is unique among the sites sampled; its chemical signature probably reflects a distinctive paleoenvironment and flora, originally recognized through palynomorphs. The characteristic PyGC-MS profile from that site is highly indicative of the Dipterocarpaceae, which would imply a rare but expected Mesozoic record of amber from a dicotyledonous tree.
In the Hanna Basin a stratigraphic interval of more than 5 mi (> 8 km) and a time gap of approximately 20 million years separate the lowest and highest occurrences of amber. Such a range in one stratigraphic sequence is unprecedented among known deposits of amber. Of particular interest is that most of these samples apparently were formed by one or several closely related species of trees. The amber is chemically and physically mature, no doubt due to deep burial. Nevertheless, despite dramatic differences in age and depth of burial, only minor chemical changes from diagenetic causes were detected among the samples. Inclusions in well-preserved pieces of amber from the Hanna Formation are fairly abundant, but typically they are distorted or were partially destroyed by effects of compaction and/or microscopic-scale deformation. Sparse wood and plant fragments and spores/pollen grains are present, but only one insect (a thrips: Order Thysanoptera) has been recognized.
Distinctive scales of conifer cones occur in the Allen Ridge Formation. The scales contain radiating vessels of resin, and they represent the taxonomically equivocal genus “Dammara.” PyGC-MS analysis of the vessel resin indicates that the same kind of tree that produced these cone scales also produced the amber in the Allen Ridge, Medicine Bow, and Hanna formations. Moreover, chemical composition of these samples closely matches that from vessels of “Dammara” cone scales from Upper Cretaceous (Turonian) strata in eastern North America. Circumstantial association of “Dammara” cone scales with several types of fossilized foliage suggests Taxodiaceae as the common source, although wood anatomy and amber chemistry also suggest Pinaceae. In spite of this taxonomic uncertainty, it is probable that 30 million years of amber production during the Late Cretaceous and Paleocene in northern North America, and probably much of Holarctica, was the result of a genus of tree that produced “Dammara” cone scales. These new data cast serious doubt upon recent proposals that all Cretaceous ambers were formed by members of the Araucariaceae. Wax residues were chemically discerned in one specimen of cone scale.
Amber, the polymerized and fossilized remains of tree resin, is a substance of scientific and aesthetic renown (Grimaldi, 1996). Inclusions of animal and plant parts within amber have provided unique insights to paleoenvironmental settings and ancient biological diversity; gross anatomical, cellular, subcellular, and even ultrastructural detail can be preserved. The intrinsic chemistry of amber, as well as associations of amber with other forms of plant fossils, can provide critical evidence about the taxonomic sources of ancient resins. Most previous studies of amber, however, have been based on samples from sites in which relative positions within a thick sedimentary column have not been of concern, or were of only minor consideration. Thick, more-or-less continuously deposited sedimentary columns have not been sampled in ways that could document: (1) evolutionary change of amber-forming trees within a single geographic area through geologic time; or (2) the taphonomic effects of progressive burial on amber chemistry in relation to compaction, almost microscopicscale deformation, and positions within an increasing geothermal gradient.
Upper Cretaceous and Paleocene strata within the Hanna Basin of south-central Wyoming (see Lillegraven and Snoke, 1996, fig. 2 for reference map) provide a consistently amber-rich, mostly nonmarine setting in which the evolution of resin-producers and taphonomic effects of deep burial on resins can be studied. The Phanerozoic, compacted stratigraphic column of the Hanna Basin (Fig. 1) totals more than 41,000 ft (12.5 km), of which almost 38,000 ft (11.5 km) were deposited during Late Cretaceous and Paleocene time. Detailed field mapping unequivocally shows superpositional relationships of the named rock units. With exception of lower reaches of the Allen Ridge Formation (Mesaverde Group), roughly the lower 14,000 ft (3.2 km) of the local Upper Cretaceous section were deposited within epicontinental seas. The remaining 8,500 ft (2.5 km) of the Cretaceous section plus more than 15,500 ft (4.8 km) of lower Cenozoic strata, however, were deposited within fluvial, paludal, and lacustrine conditions, never far above contemporary sea-levels. Highly carbonaceous to lignitic strata common throughout nonmarine parts of the column have yielded well-preserved amber, including the eight sample-groups considered here (Fig. 2). Heretofore, amber of Paleocene age has been considered rare worldwide.
Amber-bearing strata reported here represent deposition during an interval totaling almost 20 million years that traversed the Mesozoic-Cenozoic boundary (Fig. 3). Temporal controls on relevant parts of the section in the Hanna Basin are wholly biostratigraphic, using botanical, invertebrate, and vertebrate fossils variously derived from marine and terrestrial strata. The most important biostratigraphic constraints are provided by the following primary sources: Gill et al., 1970 (Upper Cretaceous marine rocks); Fox, 1971 (base of Medicine Bow Fm.); Lillegraven and Eberle, 1999 (uppermost Cretaceous of Ferris Fm.); Eberle and Lillegraven, 1998a, 1998b (lower Paleocene of Ferris Fm.); Nichols, 1998 (in Eberle and Lillegraven, 1998a; pollen near Cretaceous-Tertiary boundary); Higgins, 2000 (lower reaches of Paleocene of Hanna Fm.); and Lillegraven, 1994 (vicinity of Paleocene-Eocene boundary, near top of Hanna Fm.).
Because no reliable radiometric dates or magnetostratigraphic data have been published from the Phanerozoic column of the Hanna Basin, our biostratigraphic correlation (Fig. 3) has been linked to geochronometry determined elsewhere. All relevant strata were deposited and subsequently deformed during the Laramide orogeny. In vicinity of the Hanna Basin, orogenesis began during the Late Cretaceous and culminated in late Paleocene and early Eocene time (see Lillegraven and Ostresh, 1988 and Lillegraven, 1993 for general paleogeographic and biostratigraphic information). Our sampling was done along margins of the Hanna Basin where the modern cycle of erosion is rapidly exposing deformed strata.
HISTORIAL ASPECTS OF AMBER RESEARCH
Amber of Eocene through Oligocene age washed up on modern shores of the Baltic Sea has been collected for at least 13 millennia, and it has been used for amulets, adornment, and later for objets d'art. Serious study of myriad small organisms fossilized in Baltic amber began in the eighteenth century, and comprehensive monographs on fossil plants and animals in Baltic amber were produced in the nineteenth century (Conwentz, 1890; Klebs, 1910). Early European appreciation for the exquisite delicacy of life-forms in Baltic amber probably resulted from large-scale excavations of primary deposits on the Samland Peninsula and refinement of German optics. Amber is the only fossiliferous medium that preserves subcellular and ultrastructural details so consistently (Grimaldi et al., 1994). Even macromolecules can be preserved, although existence of ancient DNA in amber has been disputed (Austin et al., 1997).
In the twentieth century came discovery of other abundant deposits of amber. Examples include: Miocene amber from the Dominican Republic (Sanderson and Farr, 1960; Iturralde-Vinent and MacPhee, 1996); Oligocene amber from Chiapas, Mexico (Hurd et al., 1962); Eocene amber from Arkansas (Saunders et al., 1974); and even Cretaceous ambers from Lebanon (Schlee and Dietrich, 1970), northern Siberia (Zherikhin and Sukascheva, 1973), Canada (Carpenter et al., 1937; McAlpine and Martin, 1969), and New Jersey (Wilson et al., 1967; Grimaldi et al., 1989, 2000). Dozens of amber-bearing deposits occur around the world, varying in age from Triassic to Pliocene. The ancient resins formed from various coniferous and angiospermous trees (the former, until recently thought to have been almost the exclusive sources for Mesozoic ambers; Langenheim, 1969, 1990). The southern Baltic rim has the world's largest known deposits of amber. This fact, plus the extensive history of Baltic exploitation, has made the study of amber fossils a largely eastern European tradition. Paleontological studies on amber in North America, by comparison, have lagged far behind. Here we report new discoveries of stratigraphically extensive, locally abundant outcrops of Upper Cretaceous through Paleocene amber from south-central Wyoming. This now ranks among the largest known deposits of amber from North America.
On a worldwide basis, adequately studied deposits of amber are principally from lacustrine or deltaic paleoenvironmental settings. The resins were buried in extensive beds of clay or sand now associated with lignite or higher grades of coal, commonly with pyritic remains associated with iron sulfides (Iturralde-Vinent and MacPhee, 1996; Grimaldi, 1996; Grimaldi et al., 2000). Consistency of occurrence in such environments appears to be based on two primary phenomena. First, the buoyancy of resin in water allows it to be wafted and concentrated into bays and onto levees. Second, dense, largely anoxic sediments protect resins from oxidative degradation which, particularly in exposure to ultraviolet light, can occur within a few years.
RARITY OF PALEOCENE AMBER
Later parts of the Tertiary are relatively well-represented among major amber deposits, and the same is true for parts of the Cretaceous. Strata representing Paleocene time, however, contain very few known deposits of amber. Prior to our work in Wyoming, the only significant accumulations of amber of similar age were reported from lowermost Eocene strata of the Paris Basin (Nel et al., 1998), the Paleocene of Sakhalin Island of the Russian Far East (Grimaldi, 1996), and five upper Paleocene-lower Eocene localities in Switzerland (Soom, 1984). A Paleocene age for amber from Sakhalin, however, is inferred only on the basis of nearby strata and phylogenetic placement of insect inclusions. The amber in question was transported and redeposited along modern river shores; its source-deposits remain unknown.
PITFALLS IN DETERMINATION OF BOTANICAL ORIGINS
The complex array of terpenoids, their derivatives, and other compounds found in amber have been studied extensively using diverse analytical techniques (Beck et al., 1964; Beck, 1986; Grimalt et al., 1988; Shedrinsky et al., 1991; Anderson and LePage, 1995). Because molecular composition is unique to each resin or amber having a botanically distinct origin, this information has been used to detect the authenticity (Shedrinsky et al., 1991), provenance (Beck, 1986; Shedrinsky et al., 1991), and taxonomic origin (Langenheim, 1969) of specimens. Molecular composition also has been used to infer age (Lambert et al., 1996), although Grimaldi (1995) provided a critique of attempts at “chemical dating” amber.
Interpreting molecular composition of fossil resins requires caution. For example, inferring taxonomic source of amber on the basis of molecular composition alone can be difficult because of progressive biochemical changes of resin incurred by evolution of the plants themselves. Molecular composition also can be misleading due to diagenetic effects caused by geothermal heating in combination with millions of years of polymerization. The loose term “maturity” (or the process of “maturation”) often is applied in these diagenetic situations, not only to amber but also to many organic substances of geological interest (Brooks, 1981). Although differences in levels of maturity can be manifest physically (e.g., through reflectance [Douglas and Williams, 1981] or brittleness), maturity generally is quantified on the basis of chemical characteristics (MacKenzie and Maxwell, 1981; Clifford and Hatcher, 1995; Murae et al., 1995).
In truth, very few ambers, particularly those from Cretaceous strata, have had their taxonomic origins unequivocally identified. Occasionally, identification of taxonomic sources of amber has been attempted by way of plant remnants composing the lignite in which it is embedded. Anatomy of Cretaceous wood or leaf impressions, however, is rarely uniquely diagnostic of plant genera. Wood of the two living araucarian genera (Agathis and Araucaria), for example, cannot be differentiated reliably (Stockey, 1982). Species of Araucariaceae have been implicated as the source of all known Cretaceous ambers on the basis of a study exclusively using C13 NMR analysis (Lambert et al., 1996). Very importantly, anatomical fine-structure of Araucarioxylon and other Mesozoic woods originally believed to represent the Araucariaceae is ambiguous with respect to assignment to modern families of conifers (Stewart, 1983; Miller, 1988).
Botanical inclusions within amber also have been used to identify the resin-producers. Abundant stamens, petals, and even whole flowers in Dominican amber, for example, have corroborated chemical identification of this amber as formed from Hymenaea (Leguminosae), extant throughout the neotropics and eastern Africa (Langenheim, 1969; Hueber and Langenheim, 1986). But taxonomic origins can become obscured even in the presence of abundant data on molecular composition and botanical inclusions. For example, and ironically, despite a century of study, the origins of Baltic amber remain controversial (Langenheim, 1995). Anatomical structure of wood and morphology of cone inclusions indicate Pinaceae (Berendt, 1856; Conwentz, 1890), but amber chemistry suggests Araucariaceae (although several classes of compounds diagnostic of this family are lacking; Gough and Mills, 1972). Even though the Araucariaceae presently is the most commonly inferred source for Baltic amber, macrofossil remains of that family (such as cone scales) are actually rare in it.
Perhaps the best examples of confirmed taxonomic sources for disseminated ambers involve material from Eocene deposits of Axel Heiberg and Ellesmere islands. Samples have been matched chemically with amber found within fossilized cones and wood from the same deposits (Anderson and LePage, 1995). The cones and wood represent extinct species of extant Pseudolarix (Pinaceae) and Metasequoia (Taxodiaceae). Similarly, Grimaldi et al. (2000) identified a possible taxodiaceous origin of amber from Cretaceous strata of New Jersey on the basis of resins preserved in fossil wood and cone scales, although some compounds were suggestive of Pinaceae. A match in chemistry between disseminated amber and amber from an identifiable plant structure (preferably a reproductive one) provides ideal evidence for identifying a taxonomic source of the precursor resin. In addition to discovery of new amber-bearing deposits of Cretaceous and Paleocene age in Wyoming, our study reports a similar combination of preservation that links molecular structure of disseminated amber with direct evidence from associated amber-bearing macrofossils. This provides greater constraint in identification of taxonomic sources for amber in strata of the Hanna Basin.
DEPOSITIONAL, TEMPORAL, AND BOTANICAL EFFECTS ON AMBER CHEMISTRY
Amber occurs through nearly 20 million years of sedimentary accumulation within the Hanna Basin. Because no other amber deposits have been found with such extensive and nearly continuous stratigraphic superposition, this Laramide basin affords unique opportunities to examine not only the longevity of taxonomic sources but also the effects of progressive burial on chemical maturation. We suggest that molecular variation derived from differing botanical origins can be distinguished from variation due to post-depositional chemical maturation. The former should be revealed by discrete, qualitative chemical differences. The latter should be represented by more gradually varying, quantitative differences that correlate with progressive increases in age and depth of burial.
Prior studies on chemical maturation of amber have depended upon comparison of samples from deposits that are geographically distant from one another and/or geologically and botanically distinct. Therefore, effects due specifically to diagenesis have been virtually impossible to recognize with certainty. Nevertheless, tentative conclusions about maturation of amber have been offered through earlier work. For example, Grimalt et al. (1988, p. 686) stated: “…as a general rule, the proportion of hydrocarbons increases with the age of the samples….” They provided several examples in which compounds in amber were believed to represent diagenetic derivatives of original precursors. Likewise, Grimalt et al. (1988, p. 687) stated that for Cretaceous ambers a “…lack of preservation of the unsaturated components at increasing age is also observed….” along with presence of a few functional groups. Fenchone, camphor, and epicamphor, for example, are among the few oxygenated compounds in the Cretaceous ambers they studied.
Elsewhere, Murae et al. (1995) concluded that decarboxylation was the main process behind geothermal alteration of fossil resins. Clifford and Hatcher (1995) suggested that increased maturation of resin also is manifest by depletion of exomethylene radicals and an increase of naphthalenes “…relative to compounds indicative of labdatriene precursors…” Polymerization also is indicated as one of the processes involved in maturation, but the extent of polymerization is also very much a result of composition of the original resin (Langenheim, 1995). Conclusions such as these are testable using fossilized resins recovered in stratigraphic context from the Hanna Basin.
GEOGRAPHICAL AND GEOLOGICAL SETTINGS
The Hanna Basin is a relatively small but unusually deep foreland basin in south-central Wyoming (see Lillegraven and Snoke, 1996, figs. 1 and 2 for geographical overviews). As structurally defined, the Hanna Basin of today is bounded on all sides by areas of major deformation that occurred during the Laramide orogeny. The topography is mountainous to the north and south of the basin, where Precambrian basement rocks commonly are exposed at the surface. Although modern topography at western (i.e., Rawlins uplift) and eastern limits (i.e., Simpson Ridge and Flat Top anticline) of the basin is less dramatic, the structural elements in these areas are of huge proportions and also are basement-involved. The deepest part of the Hanna Basin is along its northern margin, where the original cumulative thickness of Phanerozoic strata probably exceeded eight miles (∼13 km; Fig. 1). Contractional faulting late in the Laramide orogeny elevated rocks of the Precambrian basement in excess of 40,000 ft (> 12 km) relative to the adjacent, deepest parts of the sedimentary basin (Blackstone, 1993).
Work by Lillegraven and Ostresh (1988), Lillegraven (1994), and studies in progress by Lillegraven, Snoke, and their students has led to new understanding of the structural and depositional history of the Hanna Basin. What is now the Hanna Basin was not a discrete depositional basin prior to origin and westward, basement-involved thrusting of the Rawlins uplift late in Paleocene time. Prior to the middle Paleocene, the area of today's Hanna Basin was an uninterrupted, structural and depositional eastward extension of a relatively enormous Green River Basin.
Late Cretaceous and Paleocene strata of the Hanna Basin rank among the thickest known and biostratigraphically best constrained for that interval of Earth history (Figs. 1 and 3). Regular interdigitations of marine and terrestrial strata prior to about 69 Ma allow invaluable tie-points between paleontologically based time scales developed independently in marine and nonmarine depositional realms. Terrestrial strata are being zoned using pre-existing definitions of North American Land Mammal “Ages” (Wood et al., 1941; Lillegraven and McKenna, 1986; Woodburne, 1987), and the marine strata have been zoned most importantly using associations of the largely endemic ammonites characteristic of the Cretaceous Western Interior Seaway (Obradovich and Cobban, 1975). The integrative work by Hicks et al. (1999) in marine strata of the Powder River Basin allows closer correlation among parts of time scales based on geochronology, magnetostratigraphy, and paleontology.
FIELD METHODS, SAMPLE PREPARATIONS, AND GEOLOGICAL CONTEXT OF SAMPLING AREAS
Local Amber and Its Occurrence
Figures 1–3 indicate, respectively, the stratigraphic, geographic, and chronologic distributions of amber-bearing strata sampled in the present study. Although the figures suggest at least seven sampling levels, each of these represents multiple, discrete amber-bearing outcrops. The amberiferous levels involve 27,000 ft (> 8 km) of the composite stratigraphic section, deposited through almost 20 million years. The four rock units sampled include (from oldest to youngest) the Allen Ridge Formation (fluviatile unit; Upper Cretaceous); Medicine Bow Formation (Upper Cretaceous); Ferris Formation (uppermost Cretaceous part); and Hanna Formation (lower and upper levels; Paleocene).
Prospecting for amber was done by checking eroded surfaces of highly carbonaceous to lignitic mudstone, shale, and fine-grained sandstone. Commonly, amber within such strata in the Hanna Basin is sufficiently abundant that it can be seen in natural outcrop without excavation. With exception of a few amber nuggets found on weathered surfaces, however, almost all collecting of intact pieces involved shallow excavations. Contrary to well-studied amber-bearing strata elsewhere, we have discovered local amber neither in definitively lacustrine nor deltaic paleoenvironmental settings. Nevertheless, virtually all of the occurrences represent watery paleoenvironments, especially in swampy areas of shallow depth on paleotopographic surfaces of exceedingly low relief. Productive levels tend to be rich with woody debris, and typically they include partial leaf impressions. Deeper-water, more clearly lacustrine settings have failed to yield amber.
Rock samples with amber were placed with water into tight-sealing plastic bags in the field and transported to New York for sorting, cleaning, microscopic examination, photography, and chemical analysis. Following cleaning, transparent pieces were searched for inclusions under 10–35x magnification with backlighting. All samples of amber for this study are deposited in the amber collection of the Department of Entomology, AMNH. Vertebrate fossils found in vicinity of the amber-bearing levels are housed in the Collection of Fossil Vertebrates, Department of Geology and Geophysics, UW. Physical characteristics of the samples taken for chemical analysis are described below, with sample numbers prefixed by “Wy-.”
Amber from Allen Ridge Formation (Sampling Area 1)
The Allen Ridge Formation (Figs. 1, 3–5, 9A–C) is the second of four superposed formations of the locally recognized Mesaverde Group (Lillegraven and Snoke, 1996, fig. 8). As summarized by Gill et al. (1970), the formation in the northeastern Hanna Basin has three unnamed parts: (1) fluviatile unit at the base; (2) marine unit; and (3) brackish-water unit at the top. To date, amber has been discovered only in the fluviatile unit, and this represents the oldest fossil resins known from the Hanna Basin.
Two prime outcrop areas for the fluviatile unit of the Allen Ridge Formation exist along the northeastern margin of the Hanna Basin. In both areas the strata dip rather steeply toward the center of the basin. Gill et al. (1970, fig. 7) measured and described the fluviatile unit in what they called “Pine Tree Gulch” (designated “Pine Draw” on the Como East 7.5 min USGS topographic quadrangle). That section is well exposed in the eastern half of sec. 19, T. 23 N., R. 79 W., and its thickness was reported as 685 ft (209 m) by Gill et al. (1970). Although Lillegraven discovered amber in some abundance in 1998 throughout the Pine Draw section, we have not thoroughly sampled those outcrops.
The second area of excellent exposure of the fluviatile unit (roughly three miles northwest of the Pine Draw section) is shown in Figure 4. The indicated area is immediately east of an area mapped geologically by Lillegraven and Snoke (1996, fig. 14). Measured sections A–A′ (complete for the rock unit) and B–B′ (representing only the lower 265 ft; 81 m) are described in Figure 5 (insert). From exploration done in 1998 and 1999, we report a total of 23 significant amber-bearing levels from section A–A′ and 14 from section B–B′. Although our enumeration of amberiferous levels is independent between sections A–A′ and B–B′, several can be traced directly between the two outcrops (e.g., bed 9 of section A–A′ = 11 of B–B′). Indeed, local topography at the time of deposition was extremely low, as several well-developed levels of carbonaceous shale can be traced essentially uninterrupted through the length of the outcrop area.
The fluviatile unit of the Allen Ridge Formation is dominated by fine-grained, carbonaceous to lignitic strata interspersed with much less abundant sandstone. Several levels of lacustrine strata exist, most of which exhibit isolated stromatolitic structures; unionid clams are present, but they are rare and generally poorly preserved. Occurrences of amber seem to be restricted to highly carbonaceous to lignitic mudstone and shale, in places grading into siltstone or even very fine-grained sandstone. Those layers tend to be rich with fossilized plant debris (Figs. 6 and 7), and some of them yield waterworn bone fragments. The amber-bearing strata represent deposition in laterally extensive, shallow-water swamps. Although major plant accumulations characterized this watery paleoenvironment, coal beds of commercial magnitude did not develop.
This generally paludal to lacustrine setting existed only a short distance from the western shoreline of the Cretaceous Western Interior Seaway. Indeed, as discussed by Gill et al. (1970), the type section of the Rock River Formation, a mostly shallow marine, lateral equivalent of the fluviatile unit of the Allen Ridge Formation, is only 30 miles (48 km) southeast of the outcrops shown in Figure 4. In terms of age (Fig. 3), the Allen Ridge Formation is approximately correlative with the mostly non-marine, highly fossiliferous Judith River Formation of Montana and northernmost Wyoming.
Amber and Plant Fossils
The richest deposits of Cretaceous amber from the Hanna Basin were found in the Allen Ridge Formation. Much of its amber is weathered and oxidized (Fig. 10A), but less so than material from the Medicine Bow and Ferris formations. Amber in the Allen Ridge Formation is dispersed within layers of carbonaceous siltstone and lignitic peat. It is preserved as irregular to rounded or droplet-shaped pieces 4 to 8 mm in length. Amber from the Allen Ridge Formation is never completely transparent, although the cores of pieces are often deep, translucent red with glassy, highly fractured surfaces; any given surface is comprised of a superficial rind of opaque, oxidized amber. Under the scanning electron microscope, freshly broken surfaces reveal glassy, conchoidal fracturing, indicating little or no interior deterioration (Fig. 8A, B).
An outcrop of fine-grained, consolidated sandstone is present between amber outcrop localities 14 and 15 of measured section A–A′ (Fig. 5). Bedding planes of this sandstone exhibit dissociated conifer stems and needles as well as droplets of amber (Fig. 6). The transparent yellow droplets are approximately 6 by 4 mm in greatest dimension. Although typically well preserved, they are highly fractile and difficult to remove intact from the hard matrix. Most of the plant remains are carbonized, with some preserved mainly as compressions having a relief of leaf-scale attachment sites on the axis (Fig. 7). The dominant plant fossil is Cunninghamites pulchellus Knowlton (1905), which has distinctively long, narrow, lanceolate leaves (needles; Fig. 7, top); this plant was originally described from the Judith River Formation of Montana. Another, less common plant, identified as Sequoia reichenbachi (resembling Araucaria longifolia sensu Brown, 1962), also occurs in this sandstone (Fig. 7, middle and bottom). These are the only close foliar associations found with amber in the Hanna Basin.
Beds of distinctive, dissociated cone scales (Fig. 11A–E) are present at several levels in measured sections A–A′ and B–B′ of the Allen Ridge Formation (Fig. 5). The beds also contain dissociated pieces of amber, but no other identifiable plant macrofossils were discerned. The coniferous scales are identifiable as “Dammara,” an invalid generic name originally applied to the tree genus Agathis, and first used for fossil cone scales from Cretaceous strata in Greenland (see historical discussion in Grimaldi et al., 2000). It is significant that these cone scales from the Allen Ridge Formation contain thick vessels of fossilized resin. Samples of amber from the vessels were taken for chemical analysis (samples Wy-1 and -2; Table 1); these were compared to analyses of disseminated amber collected from surrounding sediments (samples Wy-3, -4) and fossil resin removed from geographically distant Cretaceous “Dammara” cone scales (e.g., Fig. 11E; see Chemical Analyses, below).
Exact sizes and number of vessels in the cone scales are difficult to determine because individual scales are invariably incompletely preserved. The most complete specimens of scales are estimated to be 13 to 15 mm in total width across 12 to 14 vessels. Extracting complete scales from enclosing rock was difficult because the scales are delicate and the amber vessels are extremely friable. Thickest vessels occur toward the midline of the scale, and thinnest vessels are at their margins. Width of each vessel is also much thicker apically, and sample measurements of maximum/minimum widths of representative, well-preserved vessels are 1.05/0.44, 1.14/0.30, and 0.88/0.40 mm. Vessels are contiguous at the base; they fan out apically and then curve mediad toward the apex. Preserved adjacent to individual amber-filled vessels is a black, shiny, jet-like substance; clearly this is the carbonized remains of non-resinous woody tissue. Amber within the vessels typically has an oxidized, opaque, yellowish-white to light orange appearance. Where the fossilized resin retains some translucency, it is reddish.
“Dammara” cone scales are known from various Cretaceous deposits in the Northern Hemisphere (reviewed by Grimaldi et al., 2000). The only other reports from the western interior of North America, however, are: (1) “Dammara” acicularis Knowlton, from the Cretaceous Judith River Formation of Montana; and (2) cone scales like “D.” acicularis from the Paleocene Fort Union Formation (Brown, 1962). Interestingly, Brown (1962) mentioned that the cone scales also were found in carbonaceous shale with amber. “Dammara” acicularis differs from the scales we collected by having a distinctive, acuminate tip.
Definite cone scales from the Hanna Basin were found only within outcrops of the Allen Ridge Formation; they were found neither still attached as an assembled cone nor associated with identifiable foliage. Similarly resinous cone scales are known from Paleogene strata in Europe, specifically from Eocene deposits of Chechnya and the famous locality at Messel (Wilde, 1989; Schaal and Ziegler, 1992). The European Tertiary cone scales have been identified as Doliostrobus, a genus originally defined on the basis of vegetative remains. It is unclear on what basis the European cone scales were associated with the foliar genus. We are also unsure why this genus was attributed to the Taxodiaceae or possibly Araucariaceae (e.g., Wilde, 1989) and what the distinction is between cone scales identified as Doliostrobus and “Dammara.” A similar, convergent structure of parallel resinous vessels in a plant reproductive structure exists in the fossil Eomastixia (Dicotyledonae: Cornaceae), known from Messel (Collinson, 1988) and lower Eocene strata of Dorset, England (Chandler, 1962).
Amber from Medicine Bow Formation (Sampling Area 2)
Marine conditions returned to vicinity of the Hanna Basin after deposition of the Pine Ridge Formation of the Mesaverde Group (Fig. 1). Estuarine settings dominated during deposition of the Almond Formation, with progressively deeper water represented by the overlying Lewis Shale (Perman, 1990). The Lewis Shale records the last major marine incursion into this part of Wyoming. Final shallowing of the local sea is represented by various facies of the Fox Hills Formation plus roughly the lower third of the Medicine Bow Formation (see Fox, 1971 and Gill et al., 1970 for discussions).
Parts of the lower Medicine Bow Formation above the marine beds are similar lithologically in many respects to the fluviatile unit of the Allen Ridge Formation. In contrast, coal beds (of marginal commercial interest) are prominent in the lower half of the Medicine Bow Formation, and sandstone forms a substantially greater proportion of strata than it does in the Allen Ridge fluvial unit. We sampled amber in 1998 south of the Shirley Mountains on the northern margin of the Hanna Basin (Fig. 2; Schneider Ridge 7.5 min USGS topographic quadrangle) from: (1) coal in an abandoned, handdug, surface lignite mine near the eastern end of Schneider Ridge (Lillegraven and Snoke, 1996, figs. 28–30; east of cross section C–C′ and southwest of “Card-tricks Hill”); and (2) highly carbonaceous, paludal shale just west of the primitive mine. We discovered traces of amber in several beds of highly carbonaceous, nonmarine shale of the Medicine Bow Formation in that vicinity. Strata at the points of sampling dip about 75° to the southeast. The fossiliferous strata occur just south of a series of major out-of-the-basin thrust faults; the originally almost horizontal fault planes were in some cases structurally overturned by subsequent, nearby contractional faulting.
Sampled parts of the Medicine Bow Formation correlate (Fig. 3) in east-central Wyoming with the nearshore Fox Hills Sandstone and/or lower parts of the nonmarine, and locally highly fossiliferous, type Lance Formation. The western edge of the Cretaceous Western Interior Seaway by this time had regressed toward the eastern boundary of Wyoming. As with the fluvial member of the Allen Ridge Formation, local paleotopography was low, the depositional setting remained close to sea-level, and the paleoenvironmental conditions remained very watery. Sandstone in lower parts of the Medicine Bow Formation is composed almost entirely of quartz, with no recognized detrital feldspar that would indicate erosional exposure of Precambrian basement rocks on the rising Sweetwater arch (Granite Mountains of Love, 1970) to the north. Feldspathic detrital sand grains are, however, present in sandstone of upper reaches of the Medicine Bow Formation on the south flank of Schneider Ridge.
In the Medicine Bow Formation, amber was found in greatest abundance near the center of the SE 1/4 of sec. 8, T. 24 N., R. 82 W. Amber occurs largely as trace quantities of ovoid and spherical pieces ranging from less than 1 mm to no more than 8 mm diameter (Fig. 8C, D). It is entirely degraded to a light yellow, opaque, powdery composition. Where collected, the amber was embedded in heavily compacted lignite, as well as in fine-grained, brown, silty mudstone containing microveins and dispersed granules of carbonized plant remains. Two samples used for chemical analysis (Wy-5, -6, Table 1) exemplify the range of preservation, from largely decomposed (Wy-5) to reddish with some transparency still intact (Wy-6).
Amber from Ferris Formation (Sampling Areas 3 and 4)
The Ferris Formation rests conformably on the Medicine Bow Formation; lithologic criteria for distinction of the two rock units near their type areas were summarized and discussed by Eberle and Lillegraven (1998a). The Ferris Formation is an almost completely nonmarine unit that has its type area in west-central parts of the Hanna Basin. Deformed, but lithologically virtually identical strata exist at the eastern margin of the Green River Basin on western flanks of the Rawlins uplift, there mapped as Lance Formation. General features of the Ferris Formation and its contained vertebrate faunas, as seen near its type area, were summarized by Eberle and Lillegraven (1998a, 1998b) and Lillegraven and Eberle (1999).
The Ferris Formation is of unusual interest because rapid deposition of its thick, locally fossiliferous strata continued across the Mesozoic-Cenozoic boundary without significant interruption. This rock unit, therefore, has yielded important new information about histories of lower vertebrate and mammalian faunas through latest Cretaceous (Lancian) and early Paleocene (Puercan) time. Plant fossils (especially leaf impressions and palynomorphs) also are locally abundant in the Ferris Formation. Here we report discoveries of amber in latest Cretaceous parts of the Ferris Formation as they occur (Fig. 2): (1) in vicinity of its type section (sampling area 3); and (2) south of the junction between the Shirley Mountains and Freezeout Hills on the northern structural border of the Hanna Basin, north of Dry Creek (sampling area 4).
Nichols (in Eberle and Lillegraven, 1998a) described several samples of palynomorphs from below and above the Lancian-Puercan boundary. Among the most important were samples “D8105-A, -B, -C,” taken respectively from within, just below, and just above a coal bed located about 121 ft (37 m) below the base of a 26 ft- (8 m-) thick “zone of uncertainty” that defines the stratigraphic position of the Lancian-Puercan boundary (approximately the Cretaceous–Tertiary boundary; Lillegraven and Eberle, 1999, fig. 7). As discussed by Nichols (p. 17), “The sample from the coal (D8105-B) is an unusual deposit, dominated by cysts of fossil algae, constituting a boghead coal.” Lillegraven discovered amber in this lignitic deposit in 1997, and a full sampling was done by Grimaldi's team in 1998 (Fig. 9D, E). In terms of stratigraphy, the boghead coal is 1,925 ft (587 m) above the base of the type Ferris Formation (Lillegraven and Eberle, 1999, figs. 3, 4, and 7). Although certainly a persistent pond, the paleoecological and paleochemical settings of this most unusual site are otherwise unknown. Strata bearing typical Lancian vertebrate assemblages, including a partially articulated ceratopsian dinosaur skeleton, exist up-section from the boghead coal.
The second area of amber-bearing sites (Fig. 2) studied in the Ferris Formation is located in south-central parts of sec. 12, T. 24 N., R. 81 W. (TE Ranch 7.5 min USGS topographic quadrangle). The strata are tectonically overturned, within a faulted wedge of dinosaur- and Lancian mammal-bearing strata characteristic of the Ferris Formation (see geologic maps by Lillegraven and Snoke, 1996, figs. 4 and 22). Amber was discovered by Lillegraven in 1998 within several discrete horizons of highly carbonaceous, slightly lignitic shale. Extensive samples were collected by Grimaldi's team in 1999 from a layer of carbonaceous shale at the northern edge of the outcrop area. Most of the strata seem to represent ephemeral swampy facies within more general floodplain overbank deposits. There can be no doubt, on the basis of lithostratigraphy and vertebrate fossils collected nearby, that these strata accumulated during Lancian time, after deposition of the Medicine Bow Formation (Fig. 3). The overturned, amber-bearing wedge of strata is surrounded on all sides by faults or other structural complications, however, making it impossible to gain reliable measurements and to determine its true stratigraphic position within the local column.
Excavations at sampling area 3 (palynological sample sites HAN-92-1 and HAN-92-2) in the Ferris Formation revealed amber embedded in carbonized/coal-like plant material. Larger pieces of amber (up to 1 cm diameter) are lens-shaped, indicative either of subcortical formation or post-depositional compaction. Smaller pieces are rounded but too degraded to retain original drop shapes (Fig. 10B). Most pieces have irregular, asymmetrical shapes; they are slightly “marbleized,” showing swirls of darker and lighter colors. The larger, lens-shaped pieces also are microlaminated, usually disintegrating into small flakes (as in Fig. 8D, although not as powdery as amber from Medicine Bow Formation). Two samples used for chemical analysis represent the range of physical preservation: Wy-7 was a sample comprised of minute reddish fragments from the same handful of matrix; and Wy-8 was a single, transparent reddish piece (Table 1).
Unfortunately, the small samples collected from the Ferris Formation north of Dry Creek (sampling area 4) proved unsatisfactory for analysis.
Amber from Hanna Formation (Sampling Areas 5–8)
The Hanna Formation is the youngest rock unit that was involved in Laramide orogenesis of the Hanna and Carbon basins (Figs. 1 and 3). Around parts of the southern, western, and northern margins of the Hanna Basin, deformed Upper Cretaceous and older strata are overlain unconformably by nearly flatlying, almost universally poorly exposed patches of post-Laramide rocks (Ingle, 1977; Love and Christiansen, 1985). The post-Laramide sequence is dominated by fine-grained, generally poorly indurated volcaniclastic debris and often is mapped as the Browns Park Formation (e.g., Montagne, 1991). Biostratigraphically informative fossils from these patches in the Hanna Basin, however, have yet to be discovered, and reliable isotopic dates have not been established. Nowhere do erosional remnants of these post-Laramide strata directly overlie rocks younger than lower parts of the Hanna Formation. All contacts between Laramide and post-Laramide strata involve deposition upon erosion surfaces cut deeply as paleoslopes into the older rocks. Just when the Laramide orogeny ended in vicinity of the Hanna Basin, therefore, can not be determined directly from preserved stratigraphic-structural relationships.
The stratigraphically highest level of the Hanna Formation (Lillegraven, 1994) is seen north of the Medicine Bow River, along southwestern flanks of the Freezeout Hills (see Fig. 2, west-southwest of the tip of the arrow indicating “Beer Mug anticline”). There, the formation is truncated by the present erosional surface, so the rock unit's original Paleogene thickness cannot be known. Surviving remnants of the formation, however, exceed two miles (3.5 km) in thickness along northern parts of the Hanna Basin; in all probability that is close to its original thickness. As also was the case for the Medicine Bow and Ferris formations, the Hanna Formation is a syntectonic rock unit. All three formations developed through basinal subsidence in combination with multiple sources of sediment provided by progressively rising uplands to the north and south, as well as by major eastward drainages coming from the relatively enormous catchment of the Paleocene Green River Basin (see paleogeographic maps in Lillegraven and Ostresh, 1988 and Lillegraven, 1994).
Relationships of the base of the Hanna Formation to older rock units are variable from one part of the basin to another. For example, south and northwest of the town of Hanna (Fig. 2), the Hanna Formation is conformable upon the underlying Ferris Formation. Eberle and Lillegraven (1998a) discussed the essential lithologic similarities and distinctions between these two formations as seen northwest of Hanna (where the Hanna Formation was originally defined). In contrast, along eastern parts of the northern rim of the Hanna Basin (Fig. 2, vicinity of “The Breaks”), localized early Paleocene deformation of the southern Freezeout Hills (after deposition of the Ferris Formation) led to profound erosion prior to advent of deposition of the Hanna Formation. Indeed, at least 19,000 ft (5.9 km) of lowermost Paleocene plus Upper Cretaceous strata were eroded from southwestern parts of the Freezeout Mountain anticline (Lillegraven and Snoke, 1996, figs. 7 and 9) before initial deposition of the Hanna Formation. Finally, the entire eastern margin of the Hanna Basin is characterized by thrusting of the upper half of the Hanna Formation eastward (i.e., out-of-the-basin faulting) directly upon greatly deformed strata of the Ferris and Medicine Bow formations.
Lillegraven and Snoke (1996, fig. 13) provided a composite, summary description of the Hanna Formation in vicinity of The Breaks (Como West and Difficulty 7.5 min USGS topographic quadrangles). Their generalized measured section was based upon description of 20 superposed legs, positions for the lower 17 of which are shown in figure 14 of Lillegraven and Snoke (1996). Expanded description of the section at The Breaks is in progress by them, so only a few salient features relevant to the present study will be presented here. To date, samples of amber from the Hanna Formation have come exclusively from the northern Hanna Basin, from four general horizons (nos. 5–8 of Figs. 1–3) that range from about 500 ft (152 m) above the local formational base to about the 8,500 ft (2.6 km) level.
For the most part, the Hanna Formation in The Breaks is magnificently exposed. Above the 9,700 ft (ca. 3 km) stratigraphic level, however, outcrops become generally poor. Structural aspects complicate the section at The Breaks; nowhere in outcrop does there exist a complete section of the Hanna Formation uninterrupted by important faults. Within the lower 2,500 ft (762 m) of the measured section exist three consequential thrust faults (between legs 4–5, 5–6, and 9–10, respectively). In each of the three cases, the principal direction of thrusting was out-of-the-basin (i.e., broadly northward), and detailed mapping shows that each fault had the effect of placing younger parts of the stratigraphic column directly upon older parts. The precise amount of structural separation at any of these faults remains unknown. The estimate of total thickness (Fig. 1) of the Hanna Formation in The Breaks is somewhat low because of the faulting; it is not artificially thickened through structural repetition of strata.
The Hanna Formation below its 9,700 ft stratigraphic level is mainly a fine-grained section (Fig. 15), dominated by mudstone and carbonaceous to very carbonaceous shale (see Lillegraven and Snoke, 1996, fig. 12 for typical appearance in outcrop). Although all size-grades of sandstone as well as pebble-and cobble-conglomerate do exist at many levels in the column, their combined relative volumes are only a small fraction of the volume comprised by fine-grained rocks. As a result, the Hanna Formation is a structurally incompetent unit that was highly vulnerable to all manner of Laramide deformation, including folding, faulting, and widespread interstratal sliding, all of which are discernible at diverse scales (centimeters to kilometers). Deformation is observable even at microscopic levels within pieces of amber, as interpreted below.
In part because of the ductile nature of this section of the Hanna Formation, exact positions of faults are difficult to map in the field. Unlike faults in brittle strata, where discrete fault planes are clearly visible, most major faults in the Hanna Formation of The Breaks are distributed through several to many tens of meters of soft strata. The faults usually are observable only as broad, ductilely deformed, near-chaotic bands of outcrop.
Most of the Hanna Formation as seen in the northern basin represents deposition in fluvial (channels to floodplains) and paludal settings (Hansen, 1986). Immature paleosols can be observed via color-banding through much of the section. Distal reaches of alluvial fans persist within The Breaks, composed of debris shed southward from various eastern parts of the Sweetwater arch. Amber in our sampling area 5 (of Figs. 1–3) is in very carbonaceous shale only a few hundred feet above the base of the Hanna Formation. Samples from area 5 were collected approximately a half mile northwest of the measured section at a strati-graphic level equivalent to leg 4 of the measured section.
Strata of the Hanna Formation approximately between the 6,500 and 9,750 ft (1,981 and 2,972 m) stratigraphic levels in The Breaks are dominated by two especially important cycles of coal-swamp development. In both cases, the swamps became drowned by expansion of lakes (Lillegraven and Snoke, 1996, fig. 13). These two pulses of unusually watery conditions were separated by an intervening interval more typical of fluvial to paludal depositional settings. The intervening fluvial sequence constitutes most of the strata between the 7,500 and 8,200 ft (2,286 and 2,499 m) horizons. Amber sampling area 7 (on leg 15 of the measured section) is in a thick section of very carbonaceous shale (almost a lignite) below the lower series of lakebeds. Sampling area 8 (several miles west of The Breaks; Fig. 2) has a similar lithologic setting, stratigraphically below the upper sequence of lakebeds, in the lateral equivalent of leg 17 of the measured section.
Sampling area 6 (Fig. 2) is close to the Shirley Mountains, within heavily deformed parts of the Hanna Formation that are geographically separated from The Breaks. The site is north of Dry Creek, not far southeast of sampling area 4 of the Ferris Formation. Geographically juxtaposed strata of sampling areas 4 and 6, however, are separated by a major out-of-the-basin thrust, which makes accurate measurement of the stratigraphic position of amber-bearing horizons in this part of the Hanna Formation impossible (Figs. 1 and 3). Sampling area 6 is dominated by carbonaceous mudstone and shale, similar to strata typical of the lower half of the Hanna Formation in The Breaks. That similarity is the sole basis for our stratigraphic placement of the locality indicators for sampling area 6 as tentatively shown in Figures 1 and 3. We collected the amber at several points along a single, laterally extensive, highly carbonaceous layer that exhibits restricted streaks of lignite and contains locally abundant, fragmentary, dicotyledonous leaf impressions. Collecting required removal of significant overburden because these soft sedimentary rocks have been deeply weathered by the modern climate. Strata in the immediate area of collection dip more than 60 degrees to the southeast.
Amber and Inclusions
Sampling area 8 (Figs. 1–3) was the only site discovered in the Hanna Basin that yielded significant quantities of high-quality amber. At the bottom of a stratified sequence of compacted clay, fine-grained lignitic peat, and thin layers of sand, amber was found in localized pockets with proportions of approximately 1 kg amber per 100 kg of rocks. At this site, dubbed the “Glory Hole” (Fig. 9F–H), amber ranges from small (4–10 mm diameter), drop-shaped pieces (Fig. 10H) to more common pieces 2 cm or larger in diameter (e.g., sample Wy-9, Fig. 10C, D). Larger pieces typically are of irregular oval shapes and somewhat flattened. Amber pieces also were found scattered on the surface of fine, disintegrated lignite.
Amber occurs in strata at this site upward from the Glory Hole to just below the upper lacustrine unit (ULU) of the Hanna Formation. The ULU here (Fig. 10H) is some 8,695–9,680 ft. (2,650–2,950 m) above the base of the formation in the section by Lillegraven and Snoke (1996), measured in vicinity of The Breaks. Amber from strata above the Glory Hole is much smaller (mean of 4.2 mm diameter, 2–8 mm range [N = 20 pieces]) than that from the Glory Hole, and commonly it is in the form of opaque droplets. The larger pieces of amber from the Glory Hole are transparent deep orange to red and permeated with fine, parallel cracks and fissures (Fig. 10E). Upon exposure to air for several hours, moisture in the cracks evaporates and these cracks typically develop an opaque patina, rendering the piece an opaque yellowish-orange. Therefore, pieces were kept wet from the time of collection through laboratory preparation. Intact, non-fractured material is quite hard (2.5–3.0 on Mohs scale versus 2.2 for Dominican and Mexican amber); it accepts a glassy polish, much like Eocene Baltic amber and Burmese amber (the latter of probable Cretaceous age). Amber from the Hanna Formation, however, fractures easily during excavation and processing. For example, when pieces were processed in a vibratory “tumbler” to remove the dark rind of impressed lignite (a gentle technique used routinely for fragile Cretaceous ambers; Nascimbene and Silverstein, 2000), it fragmented extensively.
The good transparency and large sizes of amber pieces from The Glory Hole made it the only viable material from the Hanna Basin for preservation of biological inclusions. To scan a piece for inclusions, however, opposite sides had to be polished to produce parallel “windows.” In some cases, the polishing reduced specimens to sections only a few millimeters thick due to depth of occluding fractures. Because of fractility of the amber, pieces with inclusions were either embedded in highly stable epoxy under vacuum (Nascimbene and Silverstein, 2000; Fig. 10F) or mounted between glass microscope slides and coverslips in a synthetic resin (Permount). This method of preparation should prevent or at least significantly retard long-term degradation due to desiccation or oxidation. Preparations with inclusions (e.g., Figs. 12 and 13) were given unique sequential numbers. For chemical analysis, four pieces were selected that represented a range of physical characteristics and stratigraphic position within sampling area 8 (Table 1). These samples consisted of two pieces taken from the Glory Hole (Wy-9, Wy-10), one from immediately below the upper lacustrine unit (Wy-12), and another from between these two levels (Wy-11). The two samples from the Glory Hole were from large, transparent pieces; the others were from small, drop-shaped pieces.
Most pieces of amber from the Glory Hole contain organic and inorganic inclusions (Figs. 12–14), although few of the organic inclusions are identifiable. Organic material is common but almost always preserved as “smears” of particulate debris (Fig. 12A, B) of uncertain origin. Other inclusions involve whitish, opaque, cloudy formations (Fig. 12C, D), which are rare. These whitish formations vary from dense (Fig. 13C) to wispy and diffuse (Fig. 12D). Inspection of the interior of such formations under 300–500x magnification with a scanning electron microscope (Fig. 14) revealed an amorphous and highly porous fine-structure having no recog-nizable biological form. Thus, it is uncertain if these opaque formations represent highly decomposed remains of biological origin. In amber, particularly in Baltic amber, decomposed insects and plant parts commonly are engulfed in milky shrouds of froth composed of microscopic bubbles. Bubbles are common in almost all ambers, and that is also the case at the Glory Hole. But in samples from those sites, the bubbles are compressed into masses of flat ovals, preserved with their long axes aligned in the same direction (Fig. 12E, F). Presumably, this uniformity in orientation of bubbles is an almost microscopic reflection of a common strain axis. One piece (Fig. 12G, H) of amber from the Glory Hole contains fine grains of sand that somehow became entrapped within the original resin.
Some biological inclusions, typically fragments of wood or other plant parts, were not deformed beyond recognition (Fig. 13A–C). Larger fragments of wood (Fig. 13B, C) revealed little or no microscopic anatomy even when exposed sections were observed under the SEM. That is unfortunate, because if tracheids and pitting were observed they could provide evidence for taxonomic source of the amber. Other plant remains include rare, unidentified grains of pollen or spores (Fig. 13E–G). One piece of amber contains a group of approximately 80 such grains; their preservation on a single plane indicates virtually simultaneous capture. Details of isolated grains from this group, as well as other pieces, indicate the natural shapes to be ovoid, with several irregular, longitudinal ridges and furrows (Fig. 13F, G). The longitudinal furrows and ridges do not appear to be artifactual (i.e., due to compression or shrinkage).
Among several hundred pieces of amber screened for inclusions, only one contains a recognizable insect. The animal is very small (1.1 mm length), and its identity is obscured by extensive compression and deformation (Fig. 13H). However, the opisthognathous, cone-shaped beak, short antennae, short legs with apparently few tarsal segments, and an abdomen relatively large compared to the thorax are highly indicative of Thysanoptera (thrips). The apparent absence of wings may be preservational or indicative that the insect is a nymph. Thrips are generally uncommon but occur in virtually all ambers known from the Lower Cretaceous into upper Tertiary rocks. Insect fecal pellets (frass) found in one piece (Fig. 13D) indicate ancient proximity of insect activity and feeding.
Some amber from the Hanna Formation is preserved in original droplet shapes (Fig. 10H). Much of it, however, particularly from the Glory Hole, shows deformation due to compaction. Whereas the paucity of organismal inclusions could be attributable to subcortical formation of resin, that is doubtful in this case. Shapes of pieces from the Glory Hole amber are not lens-like, as commonly seen in resinous masses formed beneath bark or within wood. Also, capture of spores/pollen is improbable in unexposed resin. Most probably, compaction and tectonically induced microdeformation of the amber transformed biological inclusions into the smears of organic matter that are so common in amber from the Hanna Formation.
Samples were pyrolyzed using a platinum coil filament autosampler (Model 2500 Pyroprobe, CDS Analytical Inc.). Approximately 100 micrograms of sample material were placed into a quartz tube and held in place using quartz wool. The samples were placed on-line with the GC carrier flow, and then heated to 650°C for 10 seconds. The pyrolysis chamber was cleaned to vent at 1000°C for 10 seconds before each run. A Model 6890 Hewlett Packard gas chromatograph was equipped with a HP-5 column (30 m length, 0.25 mm diameter, 0.25 micron film) and operated with a 50:1 split ratio using helium as the carrier gas at constant pressure of 5.9 psi. The oven initial temperature was 40°C for two minutes, programmed at 8°C/minute to 300°C, held for 5 minutes. A mass selective detector (Model 5972A, Hewlett Packard) provided the mass spectra. Scans were taken at a rate of 2.4/second over the mass range of 30 to 550. Only compounds in which accuracy of identification matched 90 percent or more to MS database standards were used. These peaks are numbered on the PyGCs (Figs. 16–23), and their molecular identities are indicated in Figure 24.
Amber Chemistry, Diagenesis, and Possible Taxonomic Sources
Chemical composition of amber from each sampling area (1–8) is remarkably uniform, with probably the greatest within-locality variation being from area 1 (from measured sections A–A′ and B–B′ of the Allen Ridge Formation; Fig. 5). Samples Wy-1 and Wy-3 are virtually identical chemically (Wy-1 from vessels of a “Dammara” cone scale, Wy-3 from a loose piece of amber). Sample Wy-4 (a loose piece of amber) lost most of peak 9 (ionene, a naphthalene), but after 23 minutes it retained the long slope that is characteristic of samples from the Allen Ridge Formation. The most distinctive sample from the Allen Ridge Formation is from another cone scale (Wy-2). For that run the interval from 23 minutes on has a shallower slope, with 8 large peaks in this area that are barely discernible (or not recognizable at all) in other samples. Another distinctive feature of cone scale sample Wy-2 is diminution of peak 5 (a bicycloheptanol) to less than 1/10 the abundance found in other samples. In all other respects, PyGCs of samples from the Allen Ridge Formation indicate a common taxonomic source. In fact, samples from that formation closely match chemical characteristics of loose amber from the Medicine Bow and Hanna formations as well.
The only apparent chemical trend among samples from the oldest and deepest (Allen Ridge Formation) to youngest and shallowest (Hanna Formation) amber is gradual diminution of peak 11 (ionene), its surrounding PyGC shoulder, and the slope after 23 minutes. This trend, if significant, probably can be attributed to diagenetic maturation of resin samples; in all other aspects the PyGCs among these formations are extremely similar, compatible with a common taxonomic source. These results are consistent with conclusions made by Clifford and Hatcher (1995) on the abundance of naphthalenes in older fossil resins. Other naphthalenes in our samples (peaks 6, 7, 13), however, show no trends correlating with progressive age or depth of burial. In agreement with conclusions by Grimalt et al. (1988), few oxygenated compounds were found in amber from the Hanna Basin. Among those available and identifiable in our database of standards is camphor (peak 5) and its non-oxygenated form, camphene (peak 4). The abundance of camphor (the only oxygenated compound identified) differs dramatically among samples, from barely detectable (samples Wy-2, -5-8, -10) to being among the most abundant components in the amber (Wy-1, -3, -4, -11, -12). Similarly, we found no trends for camphor or related compounds that correlate with increasing geologic age or depth of burial.
The distinctive series of peaks observed in sample Wy-2 (Fig. 16, bottom), one of the cone scales analyzed from the Allen Ridge Formation, may be due to residues of natural wax. Waxes, a heterogeneous group of natural products from animals and plants, typically are composed of non-glyceryl esters formed from high molecular weight alcohols and fatty acids. The residue is present as a series of 10 evenly spaced peaks (alternating in size) of hydrocarbons between 22 and 32 minutes. Carbon numbers of the peaks are from C25 to C33, corresponding respectively to cicosane, hexacosane, heptacosane, octadecane, and nonacosane, as determined by PyGC-MS.
The carbon number and abundance of odd C-members in the cone scale residue is strikingly similar to what is seen in bees'wax (Tulloch, 1971, 1972; White, 1978; Mills and White, 1994). Original production of the wax by bees can be safely dismissed, however, because representative bees are exceedingly scarce from Cretaceous strata and modern bees secrete wax only within the nest. Scale insects (Coccoidea) also produce wax, often in considerable abundance, and there exists good documentation that coccoids were diverse and abundant on amber-producing conifers in North America (Koteja, 2000). The chemical content of coccoid wax, however, is distinctively different from our sample. Likewise, the cone scale wax is not ozocerite, a dark hydrocarbon wax of geological origin. In ozocerite the abundance of odd and even hydrocarbon members does not alternate, similar to the situation in the few natural plant waxes that have been examined (e.g., carnauba; White, 1978; Mills and White, 1994). Although to our knowledge waxes from conifer cones have not been analyzed, the residues in question almost certainly are a natural exudate from the cone, probably from the external surfaces where waxes are known to occur in extant conifers. No traces of ester components of the original wax were detected. That is not surprising given that esters are prone to pyrolysis above 300°C, whereas hydrocarbons are stable. Presumably, the temperature generated by deep burial was sufficient to pyrolyze the esters. Some cone scales have an oily surface sheen, which probably results from the waxy residue observed in analysis of sample Wy-2.
Based on the excellent chemical match between cone scale resin and loose amber from the three formations, there is little doubt of a common taxonomic source. Clearly, the kind of tree that produced the cone scales also formed the disseminated amber from the Allen Ridge, Medicine Bow, and Hanna formations. Incredibly, these deposits are separated maximally by 8,000 m of strata and some 20 million years. Amber from the Hanna Basin, overall, is extremely mature and probably was altered somewhat as a result of geothermal energy deriving from thick, overlying sedimentary rock. Nevertheless, these effects are not strongly manifest among deposits from the Hanna Basin, despite their dramatic disparities in stratigraphic position. Remarkably little chemical variation attributable to diagenetic transformation of amber has been found between the highest and lowest deposits, indicating chemical stability of amber through deep burial, geothermal heating, and geologic time.
Composition of resin from “Dammara” cone scales in the Allen Ridge Formation closely matches that of morphologically similar cone scales from Upper Cretaceous strata of Martha's Vineyard, Massachusetts and Perth Amboy, New Jersey (Raritan Formation; Fig. 23 and Grimaldi et al., 2000). This similarity is particularly noticeable for the higher molecular weight compounds (eluting > 12 minutes). Although identity of “Dammara” cone scales has been controversial (the Araucariaceae, Taxodiaceae, and Pinaceae have all been suggested; reviewed by Grimaldi et al., 2000), molecular composition of the vessel resin is suggestive of Pinaceae, as is anatomical structure of the associated wood. Peaks 10 and 12 in the PyGCs are abietane diterpenoids, found most commonly in this family of conifers (Grantham and Douglas, 1980). However, in all samples except two these peaks are minor or even indiscernible. The two samples that have peaks 10 and 12 in abundance (especially the former) are Wy-9 and -10, which were particularly large, clear pieces from the Glory Hole of the Hanna Formation. This might reflect samples that have the least maturity. Additional chemical and macrofossil evidence (below) makes interpretation of a pinaceous origin for most amber from the Hanna Basin equivocal.
Resin from sampling area 3 of the Ferris Formation unquestionably had a taxonomic source distinct from amber of the other three formations sampled in our study. This unique origin is indicated by: (1) abundance of compounds (triterpenoids) eluting at 14–16 minutes in the PyGCs (Fig. 19); and (2) virtual absence of peaks (e.g., peak 10, dihydroabietane) between 17–21 minutes that consistently occur in all the other PyGCs. The PyGC profiles, in fact, closely match that for dammar, a resin commercially harvested in southeast Asia from several genera (e.g., Shorea and Hopea) of large, emergent canopy trees in the Dipterocarpaceae (De La Rie, 1988; Van Aarssen et al., 1991; Mills and White, 1994).
The Dipterocarpaceae produce copious quantities of distinctive resins (Bisset et al., 1966, 1971; Diaz et al., 1966; Bandaranayake et al., 1975). Their commercial varieties have a clarity and hardness that is ideal for varnishes on paintings, a use employed since the early 19th century (De La Rie, 1988). We compared samples from the Ferris Formation to a sample of Kremmer's dammar (Pigmente Malmittel, D-7971, Aichestetten-im-Algau, Germany; a source well-documented to be from Indonesian dipterocarps). The match is extremely close, including identification of several compounds diagnostic for dammar resins (Fig. 20). A major difference among the samples is greatly diminished C30 and triterpenoid fractions in samples Wy-7 and -8 (Fig. 19). Although Van Aarssen et al. (1991) found little diagenetic change between recent and Tertiary fossil dammars, the diminution of certain fractions we observed in samples from the Ferris Formation probably is due to the much greater age and depth of burial than true for any dipterocarp resins examined previously.
SUMMARY AND GENERAL CONCLUSIONS
Although occurring in greatly varying quantities and physical conditions, amber is widespread throughout nonmarine Laramide strata of the Hanna Basin. It has been recovered from a stratigraphic interval extending from the Upper Cretaceous Allen Ridge Formation into the Paleocene Hanna Formation, representing sedimentary thickness of more than 5 miles (> 8,000 m) and encompassing 20 million years of geologic history. PyGC-MS chemistry of the fossil resin indicates the same (or a very similar) taxonomic source for amber from the Allen Ridge, Medicine Bow, and Hanna formations. The source almost surely was from a conifer. In contrast, amber from the Ferris Formation of sampling area 3, collected below the Cretaceous–Tertiary boundary (and intermediate in age between the Medicine Bow and Hanna formations), has a different source, probably the dicotyledonous Dipterocarpaceae. This may be due to a unique, persistently boggy local paleoenvironment that, as originally recognized on the basis of palynomorphs, probably had a paleoflora considerably different from the other sampling areas.
Despite presence of more than 27,000 ft (> 8 km) of strata between the youngest (Hanna Formation) and oldest (Allen Ridge Formation) amberbearing outcrops, only a few compounds show even minor trends that would seem to be related to increasing age and/or depth of burial. The greatest diageiietic alteration of amber composition was observed in the fossil dammar samples from the Ferris Formation, which show a considerable loss of higher molecular weight compounds. Diagenetic transformation of the amber probably does account for its overall maturity, however, and the organic inclusions in amber from the Hanna Formation were transformed by major compaction and microscopic-scale deformation.
Chemistry of vessel resin from “Dammara” cone scales, found within the Hanna Basin thus far only in the Allen Ridge Formation, indicates this general variety of plant to have been the source of dispersed ambers from the Allen Ridge, Medicine Bow, and Hanna formations. From no other amber deposit in the world has such an extensive chronological and depositional occurrence of an apparently single primary botanical source been found. Moreover, fossil resin from cone scales in the Allen Ridge Formation is chemically extremely similar to that from “Dammara” cone scales in Upper Cretaceous deposits of the eastern United States. It is probable, therefore, that many deposits of Cretaceous amber from northern North America, and probably much of Holarctica in general, were formed from the same kind of tree that produced these cone scales. No such cone scales have yet been found in the Paleocene Hanna Formation or other areas of Tertiary deposits in North America, but similar scales do occur in Eocene strata of Europe (Wilde, 1989); probably all were from closely related trees. All these results suggest existence of an amber-forming genus or species group of trees that persisted from at least the Turonian (Raritan Formation) into the late Paleocene (Hanna Formation) of North America, some 30 million years. We do not suggest that such a species or species group of tree was the exclusive source of Late Cretaceous and Paleocene amber. It was a dominant or at least common source, however, much as Pinites succinifera was a source of northern European Baltic amber (“succinite”) throughout the Eocene.
Unfortunately, the taxonomic affinities of “Dammara” cone scales remain cloudy. Besides the diagnostic presence of dihydroabietic acid in amber from the Hanna Basin, a pinaceous origin is suggested by chemical match of this amber with fossil resins from the eastern United States. Macrofossil and chemical results provided by Grimaldi et al. (2000) indicate that “Dammara” cone scales and Pityoxylon wood from the Cretaceous of eastern North America derived from the same type of tree. Samples of resin taken in situ from scales and wood match identically. Pityoxylon wood, although with a generalized (“noncommittal”) microscopic anatomy, is nonetheless consistent with Pinaceae or Taxodiaceae, and clearly is not araucariaceous. The close physical association we found between amber and Cunninghamites in the Allen Ridge Formation suggests these remains, as well as “Dammara” cone scales, formed from the same kind of plant.
Additional support for this genetic connection comes from observations by Knowlton (1905), who found that foliage of Cunninghamites and cone scales of “Dammara” acicularis commonly occurred together in the Judith River Formation. Also, Bell (1949) found foliage of Elatocladus albertaensis and cone scales of “Dammara” closely associated in Upper Cretaceous strata of the Belly River Group in Alberta. Elatocladus Bell (1965) is similar to, and perhaps synonymous with, Cunninghamites. Comments by Bell (1949) and Brown (1962) on the araucarian affinities of Cunninghamites notwithstanding, the most recent placement (Miller, 1988) of this genus is in or near the Taxodiaceae. Reconciling the botanical identity of “Dammara” and Cunninghamites may depend simply on better understanding of relationships between the Pinaceae and Taxodiaceae, especially among extinct and phylogenetically basal species.
Attribution of samples from the Ferris Formation of area 3 to the Dipterocarpaceae is compatible with the record of Cretaceous amber from a dicotyledonous tree documented by Meuzelaar et al. (1991). Unlike identities of Cretaceous conifer resins, chemical composition of dipterocarp resin is distinctive, making taxonomic attribution on the basis of chemistry alone far more certain. However, caution is indicated by the work of Van Aarssen et al. (1994), who analyzed resin within vessels of Eocene fruits of Eomastixia (Cornaceae, including present-day dogwoods), unrelated to Dipterocarpaceae. This resin also contains polycadinenes, indicating that presence of this class of compounds is not exclusive to the Dipterocarpaceae. Nevertheless, because extant Cornaceae do not produce the copious resinous exudates characteristic of the Dipterocarpaceae and exhibit other chemical differences, it can be reasonably assumed that dispersed polycadinene-containing fossil resins derived from the latter family.
Although arborescent dicots of considerable diversity are well-documented from macrofloras of Upper Cretaceous strata, including those in the western interior of North America (e.g., Upchurch and Wolfe, 1993), Cretaceous angiosperm resin has been found only rarely. An unusual styrene-like “resin” derived from trees taxonomically close to Liquidambar (Hamamelidaceae) has been reported from uppermost Cretaceous-Paleocene rocks of New Jersey (Grimaldi et al., 1989) and Upper Cretaceous strata of eastern North America (Langenheim, 1969). Dipterocarpaceae today is an Old World tropical family with macrofossils unknown from Cretaceous strata of the western interior. The only definitive macrofossil occurrence (Parashorea) of the family in North America is from Eocene strata of Alaska (Wolfe, 1977). Occurrence of fossil resin containing polycadinene-like compounds in Eocene coal from Utah also suggests presence of Dipterocarpaceae at this time in North America (Crelling et al., 1991; Meuzelaar et al., 1991); this identification, however, based wholly on chemical peculiarities, is not definitive. The same is true for the suggestion that trees of Dipterocarpaceae produced a rich deposit of distinctive fossil resin in the Eocene Claiborne Formation of Arkansas (Saunders et al., 1974). This interpretation was based on infrared spectra, not on strict identification of polycadinenes or other diagnostic compounds.
Putative occurrence of dipterocarps in Cretaceous strata of the Hanna Basin is consistent with conclusions by Upchurch and Wolfe (1993) on the existence of broad-leaved evergreen forests in the western interior during Late Cretaceous time. Generalizations, based exclusively on chemical data, that all Cretaceous ambers derived from the Araucariaceae (Lambert et al., 1996; Poinar et al., 1999) are almost certainly incorrect.
As done within several other studies (e.g., Hurd et al., 1962; Frost and Langenheim, 1974; Van Aarssen et al., 1994; Anderson and LePage, 1995; Alonso et al., 2000; Grimaldi et al., 2000), we promote a multidisciplinary approach to studies of amber. Such investigations should continue to incorporate not only chemistry and examination of inclusions, but also comprehensive stratigraphy and the study of associated plant macrofossils.
We thank the following individuals for their help in land access, field and laboratory assistance, financial generosity, or scientific advice: Victor and Nancy Anderson, Burt and Kay-Lynn Palm, Casey and Nellie Palm, Steve and Cindy Olsen, Mark Miller, Gary and Lynn Vivion, Powd and Mary Ann Boles, Terry Braxton, Rebecca Grimaldi, Dennis P. and Amy Frechette, Bob Korkow, Dick Jarrard, Keith Luzzi, Paul Nascimbene, Tam Nguyen, Steve Swolensky, Henry Silverstein, Sherrie C. Landon, Ruth Stockey, and Arthur W. Snoke. The National Geographic Society generously funded exploration of these and other North American amber deposits through a grant to Grimaldi and Lillegraven. NSF grant EAR-9506462 to Lillegraven and Snoke aided development of the geological context for this research. We are most appreciative for reviews of the manuscript by Drs. Scott Bohle and Jean Langenheim; the latter provided particularly detailed and insightful commentary. Final editorial attention by Dr. Donald W. Boyd greatly improved the manuscript.
- Received February 25, 2000.
- Revision received May 22, 2000.
- Accepted June 19, 2000.