Supplemental Data
Total Page:16
File Type:pdf, Size:1020Kb
Supplementary Information Diversity Data The Paleobiology Database [1] was used as the principal source for taxonomic ranges of the early mammalian genera. Much of the data in the Paleobiology Database came from Kielan- Jaworowska et al. [2], Lofgren [3], Clemens [4], Archibald [5], Hartman [6], Fox [7], Eaton and Cifelli [8], Montellano [9], Pearson et al. [10], and Lillegraven and McKenna [11]. Dental Functional Types For our purposes, triconodont refers to a molar with three primary cusps in a linear or obtuse- angled arrangement, such as in eutriconodontans, amphilestids, and morganucodontans. Symmetrodont molars are those with three primary cusps in a triangular arrangement and no talonid. Tribosphenic molars have dual-function trigonids and talonids such as in true tribosphenic eutherians, metatherians, and stem boreosphenidans, as well as pseudo-tribosphenic australosphenidans. Upper molars of plagiaulacoid multituberculates possess two rows of cusps and upper molars of cimolodont multituberculates have three rows, although the two dental types are also distinguished by number of incisors and premolars [2]. See Kielan-Jaworowska et al. [2] for descriptions of gondwanatherians, haramiyids, eupantotherians, and docodonts. One of the most contentious assignments for dental functional type involves the triconodonts (in the most general usage of the term) (see [71]). It should be noted, however, that even if alternative dental functional types were chosen for triconodont genera (e.g., if amphilestids were included with symmetrodonts instead of triconodontans), the overall diversity patterns of tribosphenic mammals and multituberculates relative to other dental types in the Cretaceous (the main focus of this portion of the study) would not be altered (figure 1). Lower Jaw Images: Extant Species Mandibles from 39 mammalian species belonging to 32 families and 18 orders were studied. When choosing extant species to include in this study, three traits were taken into consideration: 1) diet type, 2) similarity to Mesozoic mammals, and 3) taxonomic diversity. Since one goal of the study was to use dietary information from extant species as analogs for the extinct species, it was important to include a full range of modern dietary types, and therefore some species were included simply based on their specific diets (e.g., the grass-eating capybara). Some species were chosen based on overall ecomorphological similarities to Mesozoic mammals. For example, many Mesozoic mammals (especially multituberculates) appear rodent-like (i.e., their jaws possess a diastema, a small coronoid process, and procumbent incisors); therefore, eleven species of Order Rodentia were chosen. Other early mammals appear to be shrew-like insectivores (i.e., they have sharp cusped teeth and a pronounced coronoid process), and so, two shrew species and many other small insectivores were included (a total of 15 insectivores). Kielan-Jaworowska et al. [2] describe eutriconodonts, which include the confirmed carnivore Repenomamus [12], as animalivorous and relatively large, indicating a carnivorous diet of more than just insects. Therefore, several modern, small, extant flesh-eating carnivores were included in the study. Jaw Lengths of Extant Insectivores and Carnivores/Omnivores The distinction between insectivory and carnivory is more strongly related to body size in mammals than it is to dental or jaw specializations. To help distinguish insectivorous and carnivorous morphotypes, we tested for differences in jaw size between extant mammals belonging to these diet types. Jaw lengths of the extant taxa are recorded in table S2. Sources for these data are the same as those for dietary information and are listed in table S2. Jaw length was chosen as a proxy for body size 1 because jaws of extinct mammals were used as a proxy for body size and a measure of disparity in this study (figure S2), making the results more consistent with and comparable to the results of extinct mammals. The average jaw length of the extant insectivores (n = 15) is 39.7 mm, and the average jaw length of the extant carnivores/flesh-eating omnivores (n = 10) is 82.9 mm. The log values of the jaw lengths were used for statistical comparison of the two dietary types. A two-tailed t-test of the log values of the jaw lengths demonstrates that the mean insectivore jaw length is significantly shorter than the mean carnivore/flesh-eating omnivore jaw length (t = 3.36, p = 0.003). Since normal distribution of the data could not be confirmed, the jaw lengths of the two dietary types were also compared using a Mann- Whitney-Wilcoxon sum rank test, the results of which corroborated the significant difference between the two samples (U = 19, p < 0.05). These results support the assumption that an increase in average mammalian jaw length (i.e., body size) reflects the increased likelihood of carnivorous species within the mammalian fauna. Geometric Morphometric Landmarks Seven landmarks were chosen for the lower jaws (figure S1). The great variety of jaw shapes limited the landmarks to only those that were nearly universal for all mammals. The one exception to universality was the angular process (landmark 7), which was lost in many Mesozoic species (e.g., most multituberculates and ‘symmetrodonts’). When the angular process was not present or clear, the most posteroventral point of the jaw ramus was chosen. The landmarks were designed to capture three major characteristics of the jaw: overall length and depth, temporalis attachment area, and pterygoids/masseter attachment area (figure S1). A large temporalis muscle is indicative of a carnivorous diet and this muscle attaches to the coronoid process, the tip of which is landmark 4. Therefore, the coronoid process is expected to be largest (i.e., landmark 4 is furthest from landmarks 2 and 5) in carnivores. Herbivores are more likely to have strong pterygoids and masseters, ideal for producing grinding movements of the jaw, and therefore a more pronounced angular process and smaller coronoid process is expected. A major difference between carnivore and herbivore jaws appears to be the distance between the condylar process (landmark 6) and the angular process (landmark 7). The results of this study indicate that this holds true for extant species (figure 2). Descriptions of the landmarks are given here: 1. Anterodorsal-most tip of the body of the lower jaw. (Teeth were ignored). 2. Point at which the base of the coronoid process meets the horizontal body of the jaw. 3. A straight line is drawn between landmark 1 and landmark 2. An additional straight line is drawn perpendicular to the line between landmark 1 and landmark 2. Landmark 3 is defined as the point at which this line meets the ventral edge of the jaw. 4. Dorsal-most point of the coronoid process. 5. Anteroventral-most point of the gap between the coronoid process and condylar process. 6. Posterodorsal-most point of the condylar process. 7. Posteroventral-most point of the angular process. If the posteroventral edge of the angular process was broad, then a central point of this edge was chosen, as in figure S1. If the species did not have a well-developed angular process, then the most posteroventral point of the jaw ramus was selected. 2 Mean Shape 4 Temporalis+ 6 A-achment++ 5 1 2 Jaw+Length/Depth+ Pterygoid/Masseter+ A-achment++ 7 3 Figure S1. The mandible of an eastern mole (Scalopus aquaticus) illustrated using the seven landmarks of this study. The three triangles represent three target areas that could help distinguish different dietary types. The outline represents the consensus shape, produced by geometric morphometric analyses of the 39 extant mammals. Jaw Lengths Geometric morphometric techniques, as performed in this study, are valuable in analyzing morphology but disregard the size of specimens. The average size of non-multituberculates was desired as a means of differentiating insectivores from potential carnivores. Therefore, jaw and tooth lengths were recorded as proxies for body size. This information provided the added benefit of allowing supplementary disparity curves for non-multituberculates to be created (figures 5 and S2). Jaw images for determining length were obtained from the same sources as the jaws used for geometric morphometrics (see table S3), with any additional fragments used for length estimation taken from Kielan-Jaworowska [2]. Using the image and scale bar, lower jaw length was measured from the posterior-most end of the condylar process to the anterior-most point of the horizontal body of the jaw. All non-multituberculate jaws used in the geometric morphometric analysis were measured. To increase sample sizes in the time bins, an additional 25 jaw lengths were estimated using jaw fragments, for a total of 88 values. Since twelve genera spanned more than one time bin, the data set included a total of 100 occurrences considered in the time scale. The primary fragments used for estimations were ones in which teeth were attached and for which there existed intact jaws of closely taxonomically related species, which allowed for comparison in size (via tooth measurements) and jaw shape (via estimation of the placement of missing landmarks). Tooth Lengths Though tooth lengths can vary considerably relative to body size and may be a less accurate indicator of size than jaw lengths, more tooth fossils are available than jaw fossils, allowing for greater temporal resolution of results. Primary literature was searched for images that included intact first lower molars (m1). Since lower molars of one species can vary significantly in terms of size and shape, only confirmed first lower molars were recorded. In an attempt to bolster sample sizes in older time bins, lower molar holotype lengths were included for nine early lineage and docodont species of the Late 3 Triassic and Jurassic. Anteroposterior m1 lengths were either obtained from measurements presented in the literature or calculated using the image and scale bar. Measurements were recorded at the genera level. If multiple samples of the same species or multiple species of one genus were discovered, then the largest measurement was used in order to minimize the chances of using lengths from a juvenile.