Lower Eocene (Wasatchian-aged) strata of the Margaret Formation, Eureka Sound Group on Ellesmere Island, Nunavut preserve evidence of lush mixed conifer-broadleaf rainforests inhabited by Alligators, Turtles, Birds, and at least 25 Mammalian genera. First discovered in 1975 near the head of Strathcona Fiord, the vertebrate fossils in the Eureka Sound Group are few, fragmentary and weather-worn, which can make it challenging to identify them. By far the most abundant and diverse Vertebrate fossils have come from the Margaret Formation cropping out on the Matthew peninsula between Bay and Strathcona fiords. Nearly a century before the first discovery of Eocene Vertebrate fossils on Ellesmere, Eocene Arctic forests were first documented when Sergeant David Brainard, a survivor of the ill-fated Greely Expedition of 1881–1883, discovered petrified logs on northeastern Ellesmere Island. Among the best preserved, extensive, and photogenic of the Eocene Arctic fossil forests is the Strathcona Fiord Fossil Forest, which preserves permineralised in situ tree stumps protruding from a prominent coal seam. The tree stumps are large, with diameters ranging from 40 cm to over a meter, and closely spaced, indicating a dense forest comparable to today’s Cypress swamps in the southern United States. Eocene Arctic palaeotemperature estimates using multiple proxies suggest a mean annual temperature of 5–17˚C, with winters above freezing and summer temperatures above 20˚C. Further, the Eocene Arctic rainforests had high mean annual precipitation and humidity, comparable with today’s temperate rainforests along the North American west coast.
In the summers of 2010 and 2018, Jaelyn Eberle of the Museum of Natural History and Department of Geological Sciences at the University of Colorado, David Eberth of the Royal Tyrrell Museum and team recovered tooth fragments belonging to a Mammal and small, undiagnostic bone fragments at the Strathcona Fiord Fossil Forest. As far as they are aware, these are the first Vertebrate fossils documented from this fossil forest. Most fossil Mammals can be identified to genus and often species by dental morphology. However, the tooth fragments from the Strathcona Fiord Fossil Forest are so incomplete as to be undiagnostic by using their external morphology. Therefore, Eberle et al. analysed the tooth enamel microstructure to assist in identifying the fossils. Mammalian prismatic enamel, the hardest and most resistant material in the body, consistently differs in its microstructure among clades of Perissodactyla (odd-toed ungulates), Rodentia, Proboscidea, in a paper published in the journal PLoS One on 23 September 2020, Jaelyn Eberle, Wighart von Koenigswald of the Institute of Geosciences at the Rheinische Friedrich-Wilhelms University, and David Eberth, demonstrate that the tooth fragments from the Strathcona Fiord Fossil Forest can be identified to genus based on their enamel microstructure. In so doing, they (1) document the first fossil Mammal from the Strathcona Fiord Fossil Forest; (2) help refine the age and correlation of this site within the context of the Margaret Formation elsewhere on Ellesmere Island; and (3) underscore the utility of tooth enamel microstructure in identifying Mammalian tooth fragments that cannot be identified by traditional palaeontologic means.
The tooth fragments described below were recovered from Locality P201088-2(NU) on central Ellesmere Island, along the southwestern coastal region of Strathcona Fiord. The general area, referred to as 'Strathcona Fiord', includes the first two fossil vertebrate sites that were discovered in the Canadian Arctic (Ellesmere Island) by Mary Dawson and her colleagues in the 1970s, approximately 10–12 km southeast and northeast of locality P201088-2(NU).
The sharp-based sandstone that hosts locality P201088-2(NU) occurs less than a meter above the top of an approximately 2 m thick coal that preserves an assemblage of in situ permineralised tree stumps and root balls coined the Strathcona Fiord Fossil Forest.
Vertebrate fossils in the Strathcona Fiord region produce from the Eureka Sound Group, which consists of four formations in this area. In ascending order, these are the Mount Bell, Mount Lawson, Mount Moore, and Margaret formations. The stratigraphic section that Eberle et al. examined and measured at Strathcona Fiord consists of the uppermost exposures of the fine-grained marine Mount Moore Formation, overlain by non-coaly to coaly, paralic-to-non-marine deposits of the lower Margaret Formation. Eberle et al. place locality P201088-2(NU) approximately 316 m above the base of the multi-kilometer-thick Margaret Formation. Based on thickness data Eberle et al. regard this position as occurring in the lower portion of the Margaret Formation. Beds in this portion of the section are exposed along an extensive north-south trending ridge, and dip steeply (25˚–35˚) to the west. Lastly, Eberle et al. note that during stratigraphic measurement and examination in their study area, they identified errors in formation identification on the geologic map. Specifically, previous authors had erroneously mapped the Strathcona Fiord Fossil Forest and its thick coal as part of the Mount Lawson Formation, which instead is a Palaeocene-aged marine mudstone succession in the lower half of the Eureka Sound Group.
Strathcona Fiord fossil vertebrate sites have previously been interpreted as early Eocene (Wasatchian) in age. Based on these interpretations and stratigraphic patterns, Eberle et al. suggest the age of the Strathcona Fiord Fossil Forest and locality P201088-2(NU) is also Wasatchian, equivalent in time to the lower Margaret Formation interval at Bay Fiord (about 25 kilometers to the north-northeast).
Locality P201088-2(NU) occurs just above the stratigraphic horizon where the Margaret Formation paralic succession (complexly and thinly interbedded marine to non-marine strata) transitions up-section into a 50 m thick succession of strictly non-marine, coastal-plain strata dominated by fine sandstones and minor coals. Exposures of the lower Margaret Formation at Bay Fiord record a similar pattern of overall regression and are characterized by an upsection transition from paralic to coaly coastal plain deposits. The host sandstone for locality P201088-2(NU) exhibits ripple laminae and abundant coalified root traces, the latter indicating that a stable and likely subaerial substrate was present for Plant colonisation subsequent to deposition of the sandstone. Sparse occurrences of fossil Vertebrate fragments are present among mudstone and ironstone intraclasts in the sandstone, suggesting that lower portions of the sandstone may have been deposited as a lag deposit during waning flow. Angiosperm and Gymnosperm leaf fragments, and more root traces are common in the uppermost portions of the host sandstone, again suggesting substrate stability between subsequent later-stage sediment accumulation events.
No other intervals were observed lower in section that preserve well-developed associations of ‘fossil forests,’ ironstone and fossil Vertebrate clasts, rooted horizons, fossil leaves, and an absence of marine indicators. Furthermore, the presence of the fossil forest, multiple rooted horizons, leaf fossils, and ironstone and fossil bioclasts (the tooth fragments) all suggest forested conditions, exposed substrates, and incipient soil formation in a well-saturated setting subjected to episodic flooding. Accordingly, Eberle et al. interpret this transitional stratigraphic interval as recording an up-section shift from frequently flooded paralic shoreline settings to a relatively more up-dip coastal-plain setting where non-marine conditions prevailed. This palaeoenvironmental interpretation matches those previous hypothesised for the Margaret Formation that describe upward-coarsening cycles of interbedded cross-bedded sandstone, siltstone, mudstone, and coal. These are interpreted as proximal delta-front to delta-plain palqeoenvironments characterised by abundant shoreline sands, alluvial to estuarine channels, coal swamps, lagoons and bays, and well-forested, low-gradient interfluves.
Nunavut Fossil Vertebrate (i.e., NUFV) 2092B and 2092E, the tooth fragments analysed in Eberle et al.'s study, were collected along with approximately 20 other tooth fragments and several small, weathered bone fragments by Jaelyn Eberle, David Eberth, and team in July 2010 and by Jaelyn Eberle in July 2018 at locality P201088-2(NU). The fossils were collected on Class 2 Nunavut Territory Palaeontologist Permits 2010-003P and 2018-02P issued by the Nunavut Department of Culture and Heritage. Detailed coordinates for locality P201088-2(NU) are on file at the Nunavut Department of Culture and Heritage in Iqaluit and the Canadian Museum of Nature in Ottawa, Canada.
Given their similarity in thickness and external appearance, and the fact that they were recovered near one another, the tooth fragments probably represent the same taxon and individual. Based upon their thickness, NUFV 2092B and 2092E are from a relatively large Mammal. However, their external morphology does not allow us to reliably identify the Mammal to which they belong. Therefore, Eberle et al. decided to study the enamel microstructure. To investigate the microstructure of NUFV 2092B and 2092E, three traditional planes of section were studied, the horizontal (or transverse), vertical, and tangential sections.
NUFV 2092B and 2092E were studied under a light microscope to orient the specimens and determine the direction of the occlusal surface. The specimens were subsequently embedded in epoxy resin, and oriented so that the desired sections (horizontal, vertical, or tangential) were placed parallel to the resin surface. The embedded specimens were left for 48 hours at room temperature under a fume hood to allow the epoxy to harden. NUFV 2092B was cut using an Isomet© low-speed saw with 0.3 mm blade thickness, to produce horizontal and vertical sections, whereas NUFV 2092E was used for the tangential section. The specimens were ground in three steps. First, they were ground down to the enamel surface using a grinding wheel (grit 240). Next, handgrinding was done on fine wet sandpaper (grit 800) placed over a glass sheet, and finally the specimens were ground with fine powder grit (grit 1000) mixed with water on a glass plate. Between each of these steps, the specimens were analysed under a light microscope to ensure that grinding was not too extensive. The ground surfaces were rinsed with water, cleaned in an ultrasonic cleaner, blown dry, and etched with 10% hydrochloric acid for approximately three seconds.
Prior to scanning electron microscope analysis, the specimens were sputter-coated with gold or palladium. Scanning electron microscope analysis was conducted on a Camscan MV 2300 instrument in the Palaeontology Section of the Institute of Geosciences and a Cambridge Stereoscan 200 scanning electron microscope in the Institut für Biodiversität der Pflanzen at the Rheinische Friedrich-Wilhelms University in Bonn, Germany.
In describing the enamel microstructure of NUFV 2092B and 2092E, Eberle et al. define the units by their size and level of complexity. First, they describe the crystallites, the smallest units of enamel that are comprised of fine needles of hydroxyapatite with a diameter of less than 0.5 μm and length of more than 100 μm. Crystallites are visible under high magnification (1500x and higher). Next, Eberle et al. describe the prisms that are made up of bundles of crystallites surrounded by a prism sheath. The size and shape of the prisms differ among clades of Mammals. Prisms form at the enamel-dentine junction and grow almost to the outer enamel surface. They are typically arranged in groups or bands with the same prism orientation. Their orientation defines the various enamel types. The prismatic enamels of Mammals often have two or more different enamel types. The most primitive enamel type among placental Mammals is radial enamel, in which the prisms’ long axes parallel one another and extend radially from near the enamel-dentine junction towards the outer enamel surface. More complex enamel types occur in which bands of prisms change their orientations from the enamel-dentine junction to the outer enamel surface. Among the most often described enamel types are Hunter-Schreger bands, which are light and dark stripes often seen under a light microscope. Hunter-Schreger bands are an optical phenomenon caused by the different prism orientation in alternating bands, forming decussations. Hunter-Schreger bands occur in the enamel of most large Mammals and often are arranged horizontally. However, a few taxa, including Rhinocerotoids, have vertical Hunter-Schreger bands. Hunter-Schreger bands function as a crack-stopping device. The level above the enamel type is the schmelzmuster, the three-dimensional distribution of enamel types within the enamel that has both biomechanical and phylogenetic controls. The number of possible combinations of enamel types or schmelzmusters is very large.
By studying the horizontal, vertical, and tangential sections of NUFV 2092B and 2092E, Eberle et al. discovered a complex enamel microstructure that was challenging to interpret solely through scanning electron microscope analysis. Consequently, they also studied NUFV 2092B and 2092E at lower magnification under a light microscope using the light-guide effect. If light hits an enamel prism approximately perpendicular to its axis, the light is reflected and the prism appears light in color. If, however, the light hits a prism parallel to its long axis, it disappears into the prism, and the prism appears dark. When the source of illumination is changed from one direction to another, bands that were dark in one will be light in the other, and vice versa. The light and dark bands, each comprised of many prisms with the same orientation, are the Hunter-Schreger bands. Eberle et al. used a combination of the light-guide effect and scanning electron microscope analysis to study the enamel microstructure of NUFV 2092B and 2092E.
Based upon thickness, NUVF 2092B and 2092E belong to a relatively large Mammal. The shiny outer surface of the enamel is covered by ridges and crenulations that extend vertically and diagonally at a steep angle. Of the diverse Mammalian fauna known from the early Eocene Arctic, the Pantodont Coryphodon is among the largest and best represented by fossils, and the enamel on its teeth has vertical ridges and wrinkles. However, these characters are not unique to Coryphodon. Many Mammals, including Brontotheres that also are known from the Eocene Arctic, show varying amounts of rugosity and crenulations on the external surface of the enamel.
Others have noted the appearance of vertical stripes in the tangential section of the enamel of Coryphodon as well as ridges on the shearing crests of its molars, indicating the presence of vertical structures within the enamel. On the occlusal surface of NUFV 2092B, there are weak ridges, and when the tangential section of the enamel is illuminated from one side, faint vertical light and dark stripes are visible. However, these characters also are not restricted to Coryphodon, but occur in Rhinocerotoids and some extinct South American Mammals including Pyrotheres and Astrapotheres.
Coryphodon, however, is characterised by a unique and complex enamel microstructure coined Coryphodon-enamel. NUFV 2092B and 2092E were analysed at low and high magnifications under both a light microscope and scanning electron microscope to determine whether Coryphodon-enamel is present. Eberle et al. describe the crystallites, prisms, and schmelzmuster of the NUFV specimens, and compare them to the enamel microstructure of Coryphodon, Rhinocerotoids and other large Mammals known from the Eocene Arctic.
The cross sections of prisms are best seen in the tangential section. Individual crystallites are visible at high magnification, where they appear as fine needles making up the prisms and interprismatic matrix. The crystallites in the prisms appear approximately parallel to those comprising the interprismatic matrix, which is found in the enamel of Coryphodon. Most of the prisms in NUFV 2092E are rounded and have an open prism sheath, although there is some variability in shape, with some prisms being more oblong and narrowing towards the top. The prisms range in diameter from approximately 4–7 μm and the prism sheaths are open towards the base of the tooth or to one side. A horseshoe-shaped prism sheath was noted for Coryphodon, although it is not restricted to the genus. Although close together, the prism sheaths do not appear to touch one another. Rather, a thin (roughly 1–2 μm) layer of interprismatic matrix surrounds the prism sheaths and appears as a ‘tail’ below each prism that tapers towards the base of the tooth or to the side.
The combined interpretation of the transverse, vertical, and tangential sections of NUFV 2092B and 2092E indicates that the enamel microstructure is dominated by elongate, steeply dipping or near vertical bodies of prisms that penetrate the enamel almost from the enamel-dentine junction towards the outer enamel surface, leaving thin inner and outermost zones.
Next to the enamel-dentine junction, there is a thin inner zone of radial enamel in some places. However, in other areas next to the enamel-dentine junction the prisms do not appear to run parallel to one another, and it is difficult to discern a pattern.
The middle zone, which comprises the greatest thickness of enamel, is made up of a complex enamel wherein elongate bodies of variable thickness extend outward from the enamel-dentine junction towards the outer enamel surface. At higher magnification, the elongate bodies are Hunter-Schreger band-like in that they are comprised of steeply-dipping or rising prisms, and are separated from one another by transitional zones of horizontal (or nearly so) prisms. The nearly vertical orientation of the elongate bodies becomes obvious in the tangential section (discussed below). However, in contrast to Hunter-Schreger bands that are identified in many large Mammals, the elongate bodies and intervening transitional zones are not of uniform thickness or spacing from one another in NUFV 2092B.
Based on the tangential section, Eberle et al. predict that the nested chevrons should appear in cross section as vertical elongate bodies with variable thicknesses and branches, depending upon where the horizontal section transects them. Further, Eberle et al. predict that they should be concentrated in the inner region of the middle zone of enamel. Whereas, in the outer region of the middle zone (towards the outer enamel surface) in horizontal section, Eberle et al. hypothesise that the elongate bodies should be broader in appearance, and the nested chevron structures should all but disappear in the outermost zone of enamel. In fact, the horizontal section appears to show just that, the outer region of the middle enamel zone shows broader bodies that transition in places into radial enamel near the outer enamel surface.
Evident in vertical sections of NUFV 2092B, an outer zone of radial enamel in which the prisms run parallel to each other lies near the outer enamel surface. In some areas, it may be up to 300 μm in thickness, whereas in other areas it is much thinner, and the microstructures of the thick middle zone extend nearly to the outer enamel surface. In some areas adjacent to the outer enamel surface, the prism sheath vanishes between the parallel-oriented prisms, so that these merge into a prismless outer enamel zone, a feature observed in a number of Mammalian taxa. In other areas, the enamel contains large vacuities, probably the result of diagenesis.
In summary, the analysis of NUFV 2092B and 2092E in horizontal, tangential, and vertical sections at both low and high magnification indicates that the enamel is characterized by elongate, vertical bodies extending from near the enamel-dentine junction towards the outer enamel surface that are made up of steeply-angled prisms that decussate with adjacent bodies. The inner region of the middle enamel zone contains nested chevron or treelike structures that are evident in horizontal and tangential sections, whereas in the outer region, the vertical bands become broader and the nested chevron structures are lost. An outer zone of radial enamel or prismless outer enamel zone occurs next to the outer enamel surface.
The enamel microstructure that occurs in NUFV 2092B and 2092E is comparable to that of the Pantodont Coryphodon. Coined 'Coryphodon-enamel', this enamel type is characterised by vertical or oblique structures that in tangential section appear as light-colored bands of nested chevrons or treelike structures separated from one another by similar, though narrower, dark-colored bands. The difference in width between the light and dark stripes when light is reflected from one direction indicates that the prism orientation is not symmetrical. Also like Coryphodon-enamel (although not restricted to it), NUFV 2092B and 2092E have an outermost zone of radial enamel, as well as rounded prisms that open to one side and are surrounded by interprismatic matrix that is nearly parallel to the prisms. Given the suite of characters shared with Coryphodon-enamel, but predominantly the vertical bodies that manifest as lines of nested chevrons or treelike structures in tangential view, NUFV 2092B and 2092E possess Coryphodon-enamel. However, does Coryphodon-enamel occur in any of the other large Mammals known from the Arctic localities?
In the Canadian Arctic, Rhinocerotoids are known from early Miocene sediments of the Haughton Formation on Devon Island and from the Yukon. Rhinocerotoids have vertical Hunter-Schreger bands in their tooth enamel that look somewhat like the vertical elements in Coryphodon-enamel. However, the vertical bands in Rhinocerotoid enamel contrast with those in Coryphodon-enamel in that they are of consistent thickness and spacing from one another and separated by thin transitional zones of 2–3 prisms wide. In tangential section, Rhinocerotoid enamel shows light and dark vertical lines that bifurcate in a regular pattern and lack the nested chevron or treelike structures in Coryphodon-enamel.
Although not nearly as regular in pattern as the vertical Hunter-Schreger bands in Rhinocerotoids, Coryphodon-enamel nevertheless cannot be considered irregular enamel. This enamel type, which occurs in Proboscideans (Elephants and their extinct relatives) and some Rodents, is characterised by prisms that twist irregularly in bundles or as single prisms around each other, and bundles of prisms show a range of angles and attitudes with no consistent pattern.
The Brontotheres cf. Eotitanops and Palaeosyops are documented from Early-Middle Eocene rocks of the Margaret Formation on Ellesmere Island, and tooth fragments from a larger, younger Brontothere were recovered from middle Eocene strata of the Buchanan Lake Formation on nearby Axel Heiberg Island. However, Brontothere tooth enamel differs from Coryphodon-enamel in having U-shaped Hunter-Schreger bands, an intermediate condition between the horizontal Hunter-Schreger bands found in most large Mammals and the vertical Hunter-Schreger bands of Rhinocerotoids. Tapiroids occur at Early Eocene localities in the Margaret Formation. However, Tapiroids have horizontal to curved Hunter-Schreger bands.
The nested chevron pattern seen in the tangential section of Coryphodon-enamel is reminiscent of the zigzag Hunter-Schreger bands found in advanced Carnivorans, particularly Crocuta crocuta (Hyaena) that correlate with ossiphagous (or bone-eating) habits. However, the enamel of Crocuta crocuta differs from Coryphodon-enamel in that the vertical structures show a symmetrical pattern when illuminated from opposing directions in tangential section, and the vertical light and dark bands are of similar thickness.
Several Carnivoromorpha are known from the Margaret Formation, but their enamel shows significant differences from Coryphodon-enamel. Miacis, cf. Vulpavus, and Viverravus are small-bodied members of Carnivoromorpha whose tooth enamel is much thinner (and smoother) than that of NUFV 2092B and 2092E. Further, the tooth enamel of these early Eocene Carnivores contains undulating Hunter-Schreger bands, which are essentially horizontal, slightly wavy Hunter-Schreger bands. The Oxyaenid Palaeonictis and Mesonychid Pachyaena are the largest Carnivores in the Margaret Formation, although their fossils are rare, with each taxon represented by a single fossil in the Arctic. The enamel of these taxa has undulating Hunter-Schreger bands in the lower half of the tooth that transitions into zigzag Hunter-Schreger bands in the upper one-half to one-third of the tooth. Coryphodon-enamel altogether lacks undulating Hunter-Schreger bands. In addition to its namesake, Coryphodon-enamel is known to occur in middle Eocene Uintatherium and late Eocene Entelodon, neither of which is known from the early Eocene nor from the polar region. Therefore, it is unlikely that the Coryphodon-enamel reported here from the Strathcona Fiord Fossil Forest belongs to any other Mammal besides Coryphodon.
The tooth enamel fragments reported here, along with some poorly preserved bone fragments, thus far are the only documented vertebrate fossils from the Strathcona Fiord Fossil Forest. However, Coryphodon fossils occur elsewhere in the Margaret Formation on Ellesmere Island, so its presence at the Strathcona Fiord Fossil Forest is not surprising. What is novel in Eberle et al.'s study is the way in which they identify the fossils NUFV 2092B and 2092E to Coryphodon by way of their enamel microstructure. Complete Mammalian teeth and jaws are morphologically diagnostic and readily identified to genus and even species. However, in the Arctic, Eocene Vertebrate fossils are rare, and many are fragmentary. Eberle et al. provide an example of how enamel microstructure can be used to identify the Mammal.
The presence of Coryphodon suggests that the strata containing the Strathcona Fiord Fossil Forest are temporally correlative with the early Eocene (Wasatchian) fossil-bearing strata of the Margaret Formation approximately 25 km further north at Bay Fiord, a conclusion reached independently by lithologic correlation. At Bay Fiord, Perissodactyls, Hyaenodontid Creodonts, Miacis, and cf. Vulpavus, all of which first appear at mid-latitudes in the Wasatchian, as well as the Wasatchian index taxon Pachyaena and the archaic Ungulate Anacodon, which last appears in the Wasatchian, occur in the lower faunal level of the Margaret Formation. Although Coryphodon occurs at early middle Eocene (Bridgerian) localities at mid-latitudes, it is restricted to the early Eocene (Wasatchian) faunal assemblage in the Arctic. The strata containing the Strathcona Fiord Fossil Forest were initially mapped as the Late Paleocene Mount Lawson Formation of the Eureka Sound Group. However, Eberle et al.'s study indicates that they should be re-mapped as the Margaret Formation on the basis of lithology, palaeoenvironmental interpretations, and the presence of Coryphodon.
Vertical elements such as those found in the enamel of the Carnivore Crocuta (Hyaena) and Entelodonts are hypothesised to be an adaptation to deal with high stresses during mastication/ Crocuta is carnivorous and ossiphagous, whereas Entelodonts have been interpreted as omnivores and Pig-like in their diet, ingesting a variety of food items, including plants, meat, and bones. In contrast, Coryphodon, based on its transverse shearing lophs and carbon isotope values, is interpreted as an herbivore, and probably semiaquatic. Based on its oxygen and carbon isotope values, Coryphodon is inferred to be a year-round resident above the Arctic Circle and therefore experienced months of darkness and the shutdown of photosynthesis during the polar winter. Seasonal isotopic variations, and specifically high winter proportional carbon¹³ values in the enamel of Coryphodon teeth from Ellesmere Island, suggest a varied diet during the dark winter that probably included wood, leaf litter, and evergreen Conifers. The presence of vertical elements in the enamel microstructure of Coryphodon may have pre-adapted these large Mammals to ingesting tough, poorer quality food items during the long dark winters above the Arctic Circle. This could partly explain why Coryphodon is the most abundant herbivore in the Eocene Arctic.
See also...
Online courses in Palaeontology.
Follow Sciency Thoughts on Facebook.
Follow Sciency Thoughts on Twitter.