Animals interact with their habitat when feeding. In order to reconstruct the ecology of an animal, tracking its diet is critical. Most soft tissues of a deceased animal are rarely well preserved in the fossil record. Frequently teeth are one of the only well preserved organs of an animal’s digestive system and, as such, can be valuable sources of data for reconstructing the ecology of a fossilized animal. Certain teeth characteristics can roughly indicate what an animal has eaten. These characteristics are incisor size, canine shape, enamel thickness, ratio of the molar to premolar length, and molar hypsodonty (high-crowned teeth with enamel extending far past the gum line) index. However these methods do not reflect what an animal has indeed eaten, but rather what it is able to eat.
Alternative approaches (independent from tooth shape) in paleontology allow species-independent reconstruction of the dietary preferences of extinct (died out) species and environments. Among these techniques is Dental Microwear Textural Analysis (DMTA) . From empirical studies, it is known that microwear textures of tooth enamel depend on the types of food in an animal’s diet (Figure 1).
However, very little is known about the relationships between dental microwear textures and food mechanical properties. The question has been asked if dental microwear textures reflect more the global diet or rather the small amounts of “fallback” foods (when preferred foods are not available anymore), i.e., tough roots, hard seeds, tough and abrasive herbs. Also, very little is still known about the turnover of dental microwear textures due to diet change. These crucial points require better clarification if a more accurate interpretation of randomly versus biased sampling of data from fossils is desired.
The TRIDENT project funded by the ANR (French Agency for Research) aims at exploring all these important points mentioned above using a three-fold approach: i) collecting data from monitored food trials on domestic sheep and developing the initial model; ii) validation and further model development with dental microwear texture data from existing communities of herbivorous, wild mammals; and then iii) application of the validated model to the analysis of antelope fossils. Such an approach allows the relationship between dental microwear textural analysis and food properties to be calibrated. The application to existing, wild mammal populations will either confirm or disprove the hypothesis. Once these two steps are complete, then the interpretation of microwear data in relation to the dietary habits of fossil species will be more conclusive. Even beyond just discovering new information about the ecology from mammals’ dietary habits, these results are essential for exploring past environmental conditions. It is expected that grazing (eating vegetation at or near the ground level, such as grasses and shorter plants) antelopes, which find food in open habitats, would have distinctive types of dental microwear textures different from those which browse (eating leaves, bark, and green stems from taller plants) and find food in forested habitats.
Paleontologists and ecologists are ultimately motivated to identify the mechanisms of evolution and develop models which reconstruct the ecology of a region over certain periods of time. These ecological models can contribute useful data to the study of climate change.
Data collection and analysis
Teeth were first cleaned with acetone to remove any dust on specimens of existing species. Sometimes glue was used in the field or lab during fossil preparation to stabilize specimens and this was also removed with acetone. After cleaning, dental facets were molded with a silicone dental molding material (polyvinyl siloxane Coltène Whaledent, President Regular Body). Rather than producing a transparent resin-based cast from these molds, scans were made directly on the molds using a Leica DCM8 confocal microscope and optical surface profilometer producing inverse (negative) 3D topographic models (point clouds) of wear surfaces. The profilometer was equipped with a 100x objective lens (Numerical Aperture = 0.90, working distance = 0.9 mm) and a 331 × 251 μm area was scanned for each specimen (Figure 2).
Fig. 2: From the fossil of an antelope (bovid family, hollow-horned ruminants, such as oxen, antelopes, sheep, goats, etc.) to a 3D virtual surface. The natural dental textures of grazing, browsing, and mixed feeding antelopes seen on fossil specimens can be discriminated from the more abnormal ones with the LeicaMap software. All topographic surfaces are 333 × 251 μm2 in area.
Using the software Leica Map, raw topographic surface data are processed to erase abnormal peaks, level the surface, and produce 2D photo-simulations. Among the different textural parameters, the area-scale fractal complexity (denoted simply as “complexity”) and the length-scale fractal anisotropy (denoted simply as “anisotropy”) are known to be the best parameters for discriminating between ungulate (hoofed mammals) species with different feeding habits (Figure 3) [1, 2].
Fig. 3: Anisotropy and complexity are the most efficient parameters to use for discrimination between ungulate species based on their feeding preferences. Anisotropy versus complexity for specimens of the grazing African hartebeest (Alcelaphus buselaphus), the leaf eating giraffe (Giraffa camelopardalis) and camel (Camelus dromedarius), the mixed feeding wild donkey (Equus africanus asinus), and the yellow-backed duiker (Cephalophus silvicultor) which is a fruit browsing antelope inhabiting central African forests.
Monitored food trials
Although early studies made important contributions to the field of dental microwear textural analysis, they all failed to pinpoint to which extent each food item contributed to the different types of textures. Such issues can only be resolved by the development of monitored food testing. Up to now, only one study has been done with DMTA which involved monitored food testing on 32 rabbits . The TRIDENT project conducted a monitored food study on several dozen domestic sheep clustered into different dietary classes. The study was carried out at an experimental farm under the supervision of the Centre Interrégional d’Information et de Recherche en Production Ovine (CIIRPO) and the Institut de l'Elevage (Idele). The analyses were performed on domestic sheep using only cull ewes, meaning sheep no longer suitable for breeding and sold for meat. None of the experiments required the sheep to be handled. Sheep had full access to foods with which they were familiar. The sheep were kept inside a covered hold and fed during a minimum period of 70 days. The sheep were not kept on hay, which normally they would have eaten, but rather on dust-free wood shavings. Feeding troughs were covered with a plastic film and cleaned out daily to avoid contamination.
There were multiple food groups corresponding to different ungulate diets:
- Clover silage (livestock food preserved via fermentation in a silo) alone to simulate soft leaf browsing;
- Clover silage with chestnuts, barley, or corn, i.e., grains of different size and hardness, to simulate different types of fruit browsing;
- Ray grass silage to simulate grazing;
- Several groups with different proportions of silage to evaluate the effects of various types of mixed diet.
This unique dual dataset (diet composition and dental microwear textures) constitutes the raw data which is used to identify the relationship between diet properties (toughness, silica content, hardness) and dental microwear textures (Figure 4). The turnover rate of dental microwear textures, the textural parameters, sample size, sampling methods, and representation of a given scan size or dental facet or tooth are explored via analysis of this unique dual dataset.
Model validation with an existing population of ungulates (hoofed mammals)
It is essential to test the developed model from the DMTA and monitored food trials data using existing, wild mammal communities to determine its validity. In the course of the TRIDENT, several examples have been explored. Here are shown the results from the Białowieża case study, i.e., ruminant communities which include the European bison. The Białowieża Forest in Eastern Poland is a unique ecosystem (Figure 5). It is the only place in Europe where red deer, roe deer, moose and bison can be studied in their native environment. In addition, the inner forest is a prehistoric mixed forest free of human activity. The moose is a leaf browser, the roe deer a leaf/fruit browser, the red deer a mixed feeder, and the European bison is assumed to feed mostly on grasses. A large bone collection from the forest gathered over the last century includes 2,000 skulls and jaws.
The analysis of data obtained from specimens of 3 Białowieża cervids (animals of the deer family, namely deer, caribou, elk, and moose, characterized by bearing antlers) shows them to have a feeding habit of browsing. No significant differences in dental microwear textures are seen between browsers (moose and roe deer) and mixed feeders (red deer). The results from the European bison are unequivocal: its diet is not dominated by grasses as initially thought. The European bison is a mixed feeder including grazing (grasses) and browsing (taller plants) in its dietary bolus (soft mass of chewed food in an animal’s digestive tract). Furthermore, through 3D dental microwear texture analysis, the high adaptability in feeding behavior of the European bison to environmental conditions due to changes of the seasons was tracked .
Exploring the past ecology and environments
Regarding fossils, one of the first questions to be addressed is the feeding ecology of the animal under study (Figure 6): what did it feed on and what was its role in the ecosystem? As direct observations are impossible, proxies are necessary to reconstruct the ecology of extinct species. In this context, the study of existing species mostly serves as a baseline for the dietary reconstruction of fossilized species. For example, the paleo-community of antelopes at the 8-million-year-old site Nikiti-2 in northern Greece is dominated by grazers or mixed feeders. Among antelopes, Nisidorcas was apparently the most engaged in grazing, Gazella was a mixed feeder, and Miotragocerus was the closest to the browser end of the spectrum (Figure 6) .
Fig. 6: The results obtained when data from fossilized species are interpreted with a model based on existing species of antelopes grouped into their dietary category: grazers, browsers, and mixed feeders (ellipses encompass 80% of animals in each category). Extinct species are Miotragocerus (M), Nisidorcas (N), and Gazella (G). Mean values (circles) and 95% confidence intervals (horizontal and vertical lines) are shown.
The study of past ungulates is often directed towards paleo-environmental reconstructions [6, 7], as there is a good, though not perfect, correspondence between diet, food availability, and habitat . Browsers are absent from Nikiti-2 and the environment was best described as a savanna with a well-developed herbaceous layer . Once the paleo-environment at one given site has been evaluated, paleontologists apply the same approach on older or younger fossil sites to track possible changes in environmental conditions over geological time scales.
Summary and conclusions
Reconstructing the ecology in which animals have lived is possible via the study of their tooth structure found in fossilized or preserved remains, a sort of “paleo-dietary” investigation.
The technique Dental Microwear Texture Analysis (DMTA) allows the dietary preferences of extinct species in past environments to be determined. Wear of the teeth at the microscopic level (microwear) depends on the physical characteristics of the food in an animal’s diet. The method 3D Dental Microwear Texture Analysis (3D-DMTA) is used to produce a set of dental data from monitored food testing on sheep. With this dataset, dietary reconstruction models are developed.
Paleontologists and ecologists are interested to better understand the evolution of species and develop models which reconstruct the ecology of a region for specific time periods. These ecological models can also help to improve the understanding of climate change.
- Calandra I, and Merceron G: Dental Microwear Texture Analysis in Mammal Ecology. Mammal Rev. – in press (2016).
- Scott JR: Dental microwear texture analysis of extant African Bovidae. Mammalia 76: 157–74 (2012).
- Schulz E, Piotrowski V, Clauss M, Mau M, Merceron G, and Kaiser TM: Dietary abrasiveness determines variability in microwear and dental surface texture in rabbits. PLoS One 8: e56167 (2013).
- Merceron G, Hofman-Kamińska E, and Kowalczyk R: 3D dental microwear texture analysis of feeding habits of sympatric ruminants in the Białowieża Primeval Forest, Poland. For. Ecol. Manag. 328: 262–69 (2014).
- Merceron G, Novello A, and Scott RS: Palaeontology of the Upper Miocene Mammal Localities of Nikiti (Chalkidiki Peninsula, Macedonia, Greece) – Paleoenvironments inferred from Phytoliths and Dental Microwear Texture Analyses of meso-herbivores. Geobios – in press (2016).
- Merceron G, Kaiser TM, Kostopoulos DS, and Schulz E: Ruminant diet and the Miocene extinction of European great apes. Proc. R. Soc. B 277: 3105–12 (2010).
- Ungar PS, Merceron G, and Scott RS: Dental microwear texture analysis of Varswater bovids and Early Pliocene paleoenvironments of Langebaanweg, Western Cape Province, South Africa. J. Mamm. Evol. 14: 163–81 (2007).
- Mac Naughton SJ, and Georgiadis NJ: Ecology of African grazing and browsing animals. Annu. Rev. Ecology Syst. 17: 39–65 (1986).