Abstracts

SVP 2015

Efficient quantitative measurements of enamel occlusal wear with a watershed segmentation algorithm

Tooth wear has become an integral part of understanding the natural history and paleoecology of ancient mammals, providing insights into diet, feeding habits, and mastication. The current best-practice for measuring tooth wear is highly manual and subjective, leading to problems with use on large datasets, replication of results, and comparison across studies. To address these systematic issues we present a new, semiautomated program to measure tooth wear with both high efficiency and fidelity. Measuring Enamel Occlusal Wear (MEOW) treats photographs of fossil teeth as topographic relief maps, and then employs a watershed segmentation algorithm to automatically separate the image into regions of enamel, dentin, and background/other based on sample regions identified by the user. The ratio of enamel to dentin is then automatically calculated and saved for later analysis. MEOW presents a simple user interface, and implements several keyboard shortcuts for rapid analysis. It is also capable of analyzing multiple teeth in a single image, as well as iterating through multiple images in a single run of the program. To illustrate the power of MEOW, we performed a case study using mammals from the Eocene of Wyoming to examine the possible effects of volcanic and volcaniclastic sediment on tooth wear. In order to make a statistically valid statement about volcanic influence on tooth wear, we gathered over 350 samples of teeth from museum collections. Quantitative measurements produced with MEOW show statistically significant differences in wear before and after the onset of volcanism in the area. MEOW grants the ability to dramatically expand the dataset of available mammal tooth measurements, enabling for the first time meaningful, large-scale comparisons across time, space, and habitats

NAPC 2014

The nitty gritty of the US Western Plains: paleontological and geological implications for the evolution of grazers and grasslands

Long has a debate raged about the circumstances of the coevolutionary processes that led to the incredibly interconnected grazer and grassland ecosystem. Although new discoveries further our understanding of these processes, a critical piece of the puzzle is overlooked. Grass macrofossils, pollen, mammal teeth, and body fossils are the foundation of most arguments, but geological processes that might affect organism evolution are not always considered. Plants grow in—and mammals live and eat on—sediment, and any significant changes in surface sediment would impact the organisms involved. However, fossil organisms are too often examined out of geologic context. In this particular debate, it has been suggested that changes in mammal teeth were likely caused by some sort of ingested grit rather than the grasses themselves; unfortunately the grit has not been well studied. The Macrostrat Database and its sedimentological and stratigraphic data allow examination of large-scale changes in sediment throughout the Cenozoic of North America. Analysis of rock-type distributions show the quantity of volcanic rock dramatically increased across the early Cenozoic leading to a sustained high during the Eocene. It is hypothesized that increased volcanic output in the form of ash and particulates may have greatly contributed to the coevolutionary cycle of the North American grassland ecosystems. Active volcanoes and volcanic dust in the sediment will change the soil chemistry for plants and could be the grit previously described that drove tooth evolution in grazing mammals. Fieldwork has begun that will test this hypothesis at sites in the western Wind River Basin of Wyoming, including distinguishable pre- and post-volcanic contributions to the stratigraphic and sedimentologic records.

Grad Symposium 2013

Horses, grasses, and dirt: the compelling new story of the evolution of North America’s Great Plains.

A debate has long waged between proponents of two competing hypotheses regarding the evolution of the grazing mammals and expansive grasslands of North America: did grazers allow for grasslands by trampling and eating forests saplings or did the expansion of grasslands force the evolution of long-toothed (known as hypsodont) grazers? The studies exploring these questions often focus on the plants and animals involved: phytoliths, small silica particles in grass that make them coarse to eat; megaflora, whole plant parts; mammal tooth wear, an immediate response to grit; mammal tooth height increases (hypsodonty), an evolutionary reaction to grit; and isotopes, from both plants and animals. But what about the sediment these plants were going growing from and these animals were eating off? Large scale changes in sediment make up can be seen using the expansive rock database, Macrostrat (http://macrostrat.org). Volcanic sediments show a large spike in the Eocene, right around the advent of hypsodonty in multiple groups, including the grazers. The presence of volcanic sediment has the potential to greatly influence these dynamics, as volcanic dust would both introduce excess grit to mammal diets and teeth (provoking hypsodonty) as well as cause forest disturbance (opening areas for grass to thrive). The marked increase in volcaniclastic sediments presents an interesting direction for future research regarding the evolution of both hypsodont grazers and grasslands in North America, including field study this summer.

Ruth Dickie Seminar Competition

What sediment can tell us about the grazer/grassland evolution story of North America

The story of grazer and grassland evolution in the North American West is a complicated 'chicken and egg' problem. Living grazer impacts on grasslands include eating and trampling forest saplings, allowing the expansion of grasses; inducing the creation of phytoliths, siliceous particles that cause grass roughness; and aiding growth of grasses with fertilizer. Grasslands, on the other hand, likely caused the general grazer body plan, with long legs for long distances, etc., and changes in grazer tooth morphology, from low-crowned to high-crowned and eventually to ever-growing, due to the gritty phytoliths. Onset of high-crowned teeth and forest opening earlier than definitive grasslands complicate the picture further. Perhaps unsurprisingly, then, thistopic has inspired decades of debate amongst scientists. Using geologic information held in the Macrostrat Database about North America through the Cenozoic, I noticed a large spike in the amount of volcanic sediment around the time when some of these earliest changes occurred. Could it be that volcanoes are actually at the root of this issue? Volcanoes would cause wanton destruction of forests, through disturbances and possibly the initiation of fire regimes. Volcanic dust would also increase the amount of grit in mammals’ diets, which would affect all mammals, not just the grazing ungulates. For these reasons it seems possible that volcanoes may have started the feedback loop leading to modern grazers and grasslands. Perhaps it was not the chicken or the egg, but rather the dirt the chicken and egg were sitting on that really tells the story.

SVP 2012

The importance of sediment in the grazer/grassland story of North America

A debate has long waged between two competing hypotheses regarding the evolution of grazing ungulates and expansive grasslands of North America. Did grazers cause the sweeping grasslands we see today or did the expansion of grasslands force the evolution of hypsodont grazers? The plants and animals involved are generally the focus of studies involving these questions, from phytoliths to megaflora to tooth wear and isotopes. The presence of volcanic sediment has the potential to greatly influence these dynamics, as volcanic dust would introduce excess grit to mammal diets and teeth, provoking hypsodonty, as well as cause forest disturbances, leading to openings for grass take over. The Macrostrat Database is used to examine changes in lithology through the Cenozoic of North America. Volcanic sediments show a large spike in the Eocene, right around the advent of hypsodonty in multiple groups, including the grazing ungulates. The marked increase in volcaniclastic sediments presents an interesting direction for future research regarding the evolution of both hypsodont grazers and grasslands in North America.

SVP 2011

Contrasting patterns of rock and biotic diversity in the marine and terrestrial fossil records of North America

Macrostratigraphy examines the relationship between sedimentary dynamics and biotic diversity in the marine realm. Using packages of continuous sedimentation bound by hiatuses of non-deposition, erosion, or alternations between marine and non-marine sediments, we are able to quantify large scale patterns of sedimentation through the rock record. We linked macrostratigraphic data in the Macrostrat database to fossil collection data in the Paleobiology Database (PaleoDB), which connects fossils to the rocks in which they were originally found. Using the information from these two linked databases, we have shown that the geologic completeness of paleontological sampling, a measure of the available rock with at least one recognized fossil occurrence, is similar for the marine and the non-marine. Here, we test the hypothesis that the temporal distribution of lithologies (in the form of evenness) has a relationship to genus-level taxonomic diversity in both the marine and non-marine realms. Our evenness metric is calculated similarly to ecological community evenness, but with number of units (as abundance) with a lithology type (as species) within each time interval (as the community) therefore, dominance of one or a few lithologies within a time period would constitute unevenness. Uneven sampling of lithologies and their corresponding depositional environments can bias our perception of taxonomic diversity if taxa in poorly sampled environments are underrepresented in the fossil record. Our preliminary results show a positive correlation between and lithologic evenness and non-marine generic richness – more even sampling of lithologies corresponds to a higher observed taxonomic richness. There is no strong relationship between lithologic evenness and marine diversity. This demonstrates a distinct difference in the relationship between biologic and sedimentary processes at work in the marine and non-marine realms, and possibly differences in the magnitude of the bias imposed by the rock record on the underlying biologic patterns.

Grad Symposium 2011

Geological completeness of paleontological sampling in the non-marine rock record of North America

Large-scale, long term diversity patterns in the fossil record have been linked to temporal changes in the sedimentary rock record. Here we use Macrostrat and the Paleobiology Database (PaleoDB) to measure the geological completeness of paleontological sampling in the non-marine rock record of North America. Sedimentary rock units in Macrostrat may be unmatched to fossil collections in the PaleoDB because those units have no recognizable fossils, because the fossils have not been published in the scientific literature, or because the literature describing its fossils has not yet been entered into the PaleoDB. To differentiate between possible reasons for absences in the PaleoDB, we randomly sampled 50 of the stratigraphic names in Macrostrat that are not represented in the PaleoDB. Fossils in the published literature were found for 50% of these stratigraphic names, but there is taxonomic disparity. Of the 50 randomly sampled units, 34% had documented animal fossils in the published literature that were not entered into the PaleoDB, 24% had documented vertebrate fossils, and only 16% had documented mammal fossils. With this information, we will first be able to add new literature to the PaleoDB and then do pointed collection to determine if fossils missing from the published literature are available, or if certain rock units are barren. We also find no correlation between completeness and sampled terrestrial biodiversity. Thus, we conclude that this particular aspect of rock record-related sampling variability in the terrestrial fossil record of North America has imparted no bias in estimates of fossil diversity estimates. This understanding of the completeness of paleontological sampling has opened many avenues to further study of macroevolutionary and macrostratigraphic patterns, such as changes in the mammal lineage of North America over the Cenozoic, testing the common cause hypothesis on land, and using macrostratigraphy for enhancing phylogenetic studies.

SVP 2010

What are we missing?: Geological completeness of paleontological sampling in the terrestrial Cenozoic of North America

Paleontologists study evolution based on physical evidence preserved in the fossil record, which varies in its fidelity due to unequal sampling and the vagaries of preservation. Recent analyses of the geological completeness of the North American fossil record documented a long-term increase in completeness during the Phanerozoic. The extent to which marine vs. terrestrial completeness differs and varies over time remains, however, unknown. Using the Macrostrat Database (MD; a comprehensive geologic compilation for the entire continent, including lithologic and depositional information for most known lithostratigraphic units) and the Paleobiology Database (PBDB; fossil collections with taxonomic and geologic data), we measured the geologic completeness of paleontological completeness in the Cenozoic of North America. The total number of lithostratigraphic units and gap-bound terrestrial sediment packages remains largely stable during the Cenozoic until the Pliocene, when the total number increases substantially. The contrast between the marine and terrestrial Cenozoic sedimentary records suggests that much of the volatility in total geologic completeness is due to variability in the marine record. Geological completeness of paleontological sampling in the terrestrial fossil record is approximately 35% for lithostratigraphic units and ~45% for gap-bound sediment packages. For comparison, mean Phanerozoic completeness for the entire fossil record is approximately 22%. To determine the cause of geological incompleteness in the terrestrial fossil record, we generated a list of all the named formations in MD that do not have any PBDB collections matched to them, and then performed a literature search for fossil data on a random sample of those formations. Preliminary analysis indicates that between 47% and 68% of the named formations without any fossil collections in the PBDB have literature sources that have not yet been entered into the database. Our results have important implications for understanding the history of terrestrial evolution in North America and for helping to accelerate the pace at which the PBDB can acquire truly new paleontological information.

Grad Symposium 2010

What are we missing: Geological completeness of paleontological sampling in the terrestrial Cenozoic of North America

Paleontologists study evolution based on the physical evidence preserved in the fossil record. Because the completeness of the fossil record varies in both time and space, patterns of evolution in the fossil record can rarely be interpreted at face-value. Several groups in the paleontological community are working toward a system that will integrate geological and paleontological data for the purposes of testing evolutionary hypotheses without the confounding effects of variably complete sampling. This information is being stored in two main places- the Paleobiology Database (PBDB) and the Macrostrat Database. The PBDB is the home of fossil collections and their age, location, and geologic context, while the Macrostrat Database houses comprehensive geologic information for the entire continent, including lithologic and depositional information for all known formations. These sources can be used to determine the geological completeness of paleontological sampling in the terrestrial rock record of North America, and to possibly determine what causes incompleteness in our knowledge of the fossil record. Using the Macrostrat database, and focusing on the Cenozoic of North America, we measured the proportion of terrestrial sedimentary and volcaniclastic units in North America that are known to contain at least some body fossils. A random sample of formations that currently lack fossils will be assessed to determine if they are truly barren intervals or whether the PBDB has not yet recorded the relevant literature. Our results will have important implications for understanding the history of terrestrial evolution in North America.

SVP 2009

Lower jaw of Alvugena (Early Paleocene: Eutheria), stratigraphic debt, and the origin of the Taeniodonts

Taeniodonta, an enigmatic group of eutherian mammals from the Paleogene of North America, have had their origins and relationships questioned for well over a hundred years. Recent discoveries have provided conflicting evidence on taeniodont origins, either supporting an origin in the early Paleocene from didelphodont insectivores (palaeoryctoids) or an origin in the Late Cretaceous from some unknown eutherian group. Alvugena, a possible palaeoryctoid-taeniodont intermediate of Puercan (Pu2) age from the Hanna Basin, Wyoming, has been known from a single upper dentition. A well-preserved lower dentition that we refer to Alvugena of Pu1 age from the Williston Basin of North Dakota, adds to our knowledge of this taxon. Because the oldest, most basal taeniodont occurs in the Late Cretaceous, the interpretation of Paleocene Alvugena itself as ancestral to the taeniodonts is problematic. The temporal range of Alvugena now extends into the early Puercan (Pu1), and so the actual time separating these two taxa is not great. Stratocladistic analysis of 37 morphological characters and a stratigraphic character using StrataPhy supports the results of the analysis of morphological characters alone: monophyly of Taeniodonta, placement of Schowalteria as the most basal taeniodont, and Alvugena as a sister-taxon to the entire in-group. Furthermore, there are no known autapomorphies that would preclude Alvugena from being the ancestor to the taeniodonts including Schowalteria. The order of appearance Schowalteria-Alvugena-other taeniodonts does not precisely match the branching order of the phylogeny, implying that an Alvugena ghost lineage may extend into the latest Cretaceous. The temporal gap between Alvugena and Schowalteria is small, much less than the average duration of a mammalian species in either the Cenozoic (~2.5 my) or Cretaceous (~4 my), minimizing the assumed duration of the hypothetical ghost lineage. There is mismatch here between branching order and appearance of fossils even though the phylogeny exhibits great congruence overall. This suggests that more can be learned about the origin and early evolution of taeniodonts by sampling in the uppermost Cretaceous.

SVP 2008

Preliminary phylogeny of Taeniodonta, an enigmatic order of eutherian mammals (Paleogene, North America)

Taeniodonts are an enigmatic group of eutherian mammals from the Paleogene of North America that ranged from small insectivorous mammals to large, pig-like ones. Since Robert Schoch’s monograph describing all then known taeniodonts and their relationships, researchers have made several important discoveries that either complement or contradict his ideas on taeniodont phylogeny and origins. Because there is now a debate about the phylogeny and origins of this group, I here undertake a phylogenetic analysis including all of the taeniodonts. I also examine the relationships of taeniodonts to other early eutherian mammals. This study provides a phylogenetic analysis of the group using characters and descriptions from the literature. Using a Willi Hennig version of TNT, a phylogenetic analysis was performed using the characters from the literature as well as personal observations of taeniodont specimens at the American Museum of Natural History. The phylogenetic analysis shows that the Conoryctidae and Stylinodontidae family divisions of Schoch are supported. The plesiomorphic Onychodectes, however, falls outside of the Conoryctidae, in support of the removal of Onychodectes from this family. The monophyly of Taeniodonta was supported with the purported Stylinodontid Schowalteria falling out as the sister taxon to the rest of the Taeniodonta. More analyses with other outgroups and additional characters are still needed. Finding where the Taeniodonta are rooted in the phylogeny of early mammals will be able to help to evaluate the hypothesis of the Cimolesta, a higher level grouping of several extinct orders with the extant pangolins. This new phylogeny will also be used to test the claim of increased evolutionary rates in these animals, on the basis of their achieving large body size so early in the Cenozoic mammal record.