Ecomorphology of Mammalian Fossil Lineages: Identifying Morphotypes in a Case Study of Endemic South American Ungulates

Abstract

The notoungulates are one major clade of extinct native South American ungulate mammals. Several notoungulate morphotypes have been proposed by comparison with mammals from other large land masses, representing instances of convergent evolution in continental isolation. As with other entirely fossil lineages, the reconstruction of their functional biology is challenging, one major obstacle being the fragmentary and distorted nature of the fossil remains for which application of conventional quantitative data analysis techniques is compromised. Here we explored the patterns in morphospace of representatives of all notoungulate families on the basis of the qualitative scoring of 31 cranial characters that provided a basic description of shape and an approximation to sensory and mechanical functions of the skull. We calculated pairwise distances between all taxa with an index that proportionally corrected overall similarity by both character number and dissimilarities. The distances were used to generate a morphospace with a multivariate technique, Principal Coordinates Analysis. We recovered a tight group of basal Paleocene-Eocene notoungulates and a scattered group of Oligocene and Neogene forms. Lines drawn in that space connecting successive sister terminals clearly showed a departure from basal genera toward four diverging, previously proposed morphotypes, namely rabbit-, large-rodent-, small-horse-, and rhino-like forms. An extensive analysis identified the character basis of these morphotypes. Mapping additional information such as body size and degree of hypsodonty effectively aided in the interpretation of notoungulate morphofunctional evolution. We briefly discuss the utility of the qualitative morphospace as a tool for exploring the ecomorphology of fossil lineages and the evolutionary implications of this reconstruction.

Appendices

Appendix 1 Specimens Examined in the Present Study

Institutional abbreviations: AMNH: American Museum of Natural History, New York; MACN: Museo Argentino de Ciencias naturales Bernardino Rivadavia, Buenos Aires, Argentina; MLP: Museo de La Plata, Buenos Aires, Argentina; MMMP: Museo Municipal de Historia Natural “Lorenzo Scaglia”, Mar del Plata, Argentina; PVL: Colección Paleontología de Vertebrados Lillo, Tucumán, Argentina.

Specimen list (alphabetical):

Archaeohyrax MACN-A 52-617; Boreastylops MLP 78-V-6-5, PVL 4261; Brachystephanus MLP 68-X-17-2; Campanorco MLP 79-VI-16-1, PVL 6225; Colbertia PVL 4183, PVL 4184 (1-2), PVL 4293, PVL 4300, PVL 4608, PVL 6218, PVL 6227; Coquenia PVL 5853, PVL 6397; Griphotherion PVL 5903; Hegetotherium PVL 91; Interatherium MACN-A 3444; Leontinia MACN-A 52-573 (photographs provided by Virginia Deraco); Mesotherium MACN 2036, MACN 8690; Notopithecus MACN-A 10790, MACN-A 10787; Notostylops MACN-A 10499, MACN-A 10466; Oldfieldthomasia MACN-A 10376, MACN-A 10772; Pachyrukhos PVL 84, MACN-A 3313; Paedotherium MACN-A 1184, MACN-PV 1081, MACN-PV 7214, MACN-PV 10178, MMMP 226-S, MMMP 698-S, MMMP 710-S, PVL 503, PVL 2272, PVL 2274, PVL 2377, PVL 3386; Pampahippus PVL S-4192; Protypotherium MACN-A 3882, MACN-A 3920, MLP 84-III-9-12; Scarrittia AMNH 29581, AMNH 29583, AMNH 29584 (photographs provided by Virginia Deraco); Simpsonotus MLP 73-VII-3-11; Toxodon MLP 12-II-26, PVL uncatalogued specimen; Trachytherus MACN-A 52-490.

Appendix 2 Commented List of Characters Used in this Study

See Fig. 1 for illustration of characters.

• 1. Basicranial angle: less than 180° (0); or approximately 180° (1); or greater than 180° (2). The basicranial angle is defined as the angle between the horizontal plane of the basicranium with respect to the rostrum of the skull. The former is defined by the ventral aspect of the presphenoid, basisphenoid, and basioccipital, and the latter by the upper alveolar line, specifically the plane including the points at the base of left and right canines and last molars. The angle is assessed on a lateral view of the skull.

• 2. Zygomatic plate: absent (0); or incipient (1); or well developed (2). The term “zygomatic plate” refers here to a lateral widening of the anterior root of the zygomatic arch, associated with a flat or concave smooth ventral area for the origin of enlarged anterior tracts of the masseter muscle. The skulls with incipient plates present a moderately widened ventral edge of the anterior root of the arch and a comparatively small flattened area. In taxa with a well-developed plate (state 2), the anterior root of the arch is greatly widened and a large, concave, and smooth area is present.

• 3. External auditory meatus orientation: nearly horizontal (0); or dorsolateral (1). In the groups considered here, the external auditory meatus is a lateral tubular projection of the ectotympanic. Orientation of the tube is independent of the length of the meatus. In state 1, the tube extends laterally, dorsally, and usually posteriorly, reaching the level of the skull roof.

• 4. Development of the epitympanic sinus (squamosal): absent or inconspicuous (0); or well developed (1); or prominent (2). The epitympanic sinus of the squamosal is a singular feature of notoungulates and pyrotheres (see Billet 2010). This structure is a hollow space developed inside the squamosal bone. The sinus is located at the posterolateral corner of the skull, and connected to the tympanic cavity ventrally. This feature is present in all notoungulates (Cifelli 1993) and in Pyrotherium; however, in some Paleogene genera the internal space of the squamosal is not associated with an external inflation and the sinus is not evident in intact skulls (state 0). Thus, the first state also contemplates the possibility that the sinus is absent. Nevertheless, the presence of the sinus was confirmed in some genera assigned 0, such as Colbertia, even in the absence of the external inflation (see García-López 2011).

• 5. Development of the tympanic bulla: small (0); or well developed, prominent (1). All notoungulates exhibit a tympanic bulla that is part of the ectotympanic bone (see Gabbert 2004). In the first state, the bulla is small with respect to the area of the basicranium in ventral view, and not particularly prominent ventrally. In the second state, the bulla is a large, balloon-like structure, usually occupying a large surface of the basicranial area in ventral view.

• 6. Height of the rostrum: low (0); or high (1). This character is considered as the relation between the length of the rostrum in lateral view (from the rostral edge of the orbit to the tip of the premaxilla) and the height of the rostrum immediately rostral to the orbit. In skulls with a high rostrum, both measures are similar. In a low rostrum, the first measure is clearly greater than the second.

• 7. Length of the rostrum: short (0); or long (1). This character is measured as the proportion of the length of the palate (from the edge of the choanae to the rostral edge of the premaxilla), relative to the total skull length. A long rostrum is one comprising more than 50 % of the total length of the skull in ventral view.

• 8. Shape of the mandibular (= glenoid) fossa: shallow, rostral border not well defined (0); or transversal, rostral border well defined (1). The first state of this character is referred to a shallow fossa where width is subequal to the length and where rostral limit is not marked by a well-developed process on the ventral edge of the zygomatic arch. This is related to the shape of the articular process of the mandible, and, therefore, this last feature was not included in the analysis to avoid redundancies.

• 9. Shape of the anterior palate (on the premaxilla): semicircular (0); or subrectangular (1); or Pyrotherium-like (2). Most genera show state 0, with a roughly semicircular dental arcade. Some genera (e.g., Toxodon) exhibit a truncated rostrum with a transversally straight arcade (state 1). The condition observed in Pyrotherium is of irregular shape due to the great development of the incisor alveoli (state 2). This character has been used to assess dietary adaptations of extinct ruminants (Solounias and Moelleken 1993).

• 10. Shape of the upper toothrow: straight (0); or bowed, toothrow medially concave (1). State 0 applies to either parallel or divergent toothrows as the character refers to one side of the upper dental arcade.

• 11. Posterior width of the palate: very narrow (0); or narrow (1); or wide (2). State 0 refers to the singular condition observed in Pyrotherium, which presents an unusual narrow palate. State 1 is scored when the posterior width of the palate represents less than two times the buccolingual length of the molars. In state 2 the posterior width exceeds that measure.

• 12. Position of the nasal aperture: terminal (0); or slightly retracted (1); or retracted (2). The lateral edges of the nares and the rostral end of the nasals can be located at the level of the first incisors (state 0), slightly behind this position (state 1), or located at the level of the postcanine teeth (state 2, only present in Pyrotherium).

• 13. Diastema: absent (0); or present, small (1); or present, well developed (2). A diastema is any space in the dental arcade between the canines and cheekteeth, existing as a consequence of evolutionary loss of dental pieces or separation between them (usually associated with some reduction of the teeth).

• 14. Postorbital constriction: narrow (0); or wide (1). The confluence of the temporal lines just posterior to the postorbital processes may determinate a strong postorbital constriction (state 0).

• 15. Temporal fossa: small (0); or large (1). The extent of the temporal fossa is determined dorsally by the temporal lines and the sagittal crest, which mark the perimeter of the area for the origin of the temporal muscle in the skull. Small fossae (state 0) are shallow and retracted to the surface of the parietal, with sagittal crests weak or absent. In large fossae (state 1), the temporal lines begin immediately posterior to the postorbital processes and converge in a strong sagittal crest that joins the lambdoid or nuchal crest.

• 16. Development of the postorbital process: small (0); or large (1). A small postorbital process (state 0) is either a tubercle-like structure with a limited base, or a sharp, needle-like process. In either case the tip of the process does not reach far into the orbit. A large process (state 1) exhibits either a wide base, or it is long, thus approaching the zygomatic arch.

• 17. Length of the upper molar row relative to upper premolar row: similar length (0); or molar row longer (1). This difference could be due to a lesser mesiodistal development of the premolars (being all present) or to the evolutionary loss of one or more of the anterior premolars.

• 18. Development of the jugular (= paracondylar) process in occipital view: short, not exceeding the extent of the tympanic bulla ventrally (0); or long, greatly exceeding the tympanic bulla ventrally (1).

• 19. Development of the zygomatic arch in lateral view: slender (0); or intermediate (1); or robust (2). Slender arches (state 0) are thin bars of bone. Arches are scored as robust (state 0) when its depth is comparable to the height of the orbit.

• 20. Orientation of the mandibular symphysis: sub-vertical (0); or procumbent (1). In the procumbent state, the mental surface of the symphysis forms an angle of 45° or less with the horizontal (ventral edge of the mandible).

• 21. Position of the articular process (= mandibular condyle) relative to the lower molar row: slightly dorsal to the molar row level (0); or much higher than the molar row level (1). In all the studied genera the articular process is located above the level of the lower molar row. The character is scored as 0 when this difference in height is less than twice the height of lower molars. In Campanorco the mandible is unknown. We scored this taxon on the basis of the relative position of the glenoid fossa with respect to the occlusal surface of the upper molars.

• 22. Mandibular depth at the level of the lower premolars: shallow (0); or deep (1). This character is scored as 1 when the depth of the mandibular body is greater than three times the heigh of the lower premolar crowns.

• 23. Development of the angular process: not expanded (0); or expanded (1). In most Paleogene genera the angular process is a rounded prominence on the mandibular ramus, which is more developed caudally than ventrally. Expanded processes (state 1) describe a wide, continuous curve from the articular process (= mandibular condyle) to the rostral end of the masseteric line, on the ventral edge of the mandibular body.

• 24. Anterior extent of the insertion of the masseteric muscle on the mandible: at the level of the distal end of the lower molar row (0); or anterior to the level of the caudal end of the lower molar row (1).

• 25. Height of the coronoid process: low, not well developed (0); or high, well developed (1). The top of the coronoid process is the area of insertion of the temporal muscle, and its size is associated with the development of this element. Low processes do not exceed dorsally the height of the articular process (state 0). High processes (state 1) greatly exceed the height of the articular process.

• 26. Size of upper incisors: I1 similar to I2 (0); or I1 larger than I2 (1). In this character only the size of the first upper incisor is considered and not its morphology. The character is scored as state 1 when the mesiodistal length or the crown of the I1 greatly exceed that of I2.

• 27. Aspect of the occlusal surface of the upper molars at average condition of wear: dominated by the ectoloph (0); or roughly flat (1). The cusps of the ectoloph (paracone and metacone) are usually well developed in most notoungulates. This character considered the height of these cusps. In the state 0, the paracone and the metacone are greatly exceed in height the protocone and the hypocone, which are the other two main cusps of the notoungulate upper molar; so the ectoloph is the most prominent structure of the occlusal surface. In state 1, height is comparable across all the cusps, and the occlusal surface appears roughly flat.

• 28. Upper canine: absent (0); or moderately developed (1); or hypertrophied (2). The upper canine can be absent (state 0) or present with two degrees of development as compared with the size of upper incisors. The canine can be of comparable size relative to the incisors (state 1) or be significantly larger than the incisors (state 2; present in only one of the included genera, Thomashuxleya).

• 29. Morphology of the upper molars: protoloph and metaloph moderately developed (0); or wide lophs (1); or occlusal surface without fossettes (2); or bilophodont (3). Although this character is easily influenced by wear, the development of the lophs can be discriminated by other features, like the orientation of the walls of the lophs. State 0 refers to slender lophs and state 1 to strong lophs, which occupy a large occlusal area. State 2 refers to a continuous occlusal surface, without fossettes. State 3 is proper of Pyrotherium, which exhibits two lophs likely homologous to a protoloph and metaloph (see Billet 2010).

• 30. Differentiation of the upper cheekteeth: heterodont, premolars and molar well differentiated (0); or virtually homodont (1). Despite the different degrees of molarization in notoungulate premolars, most representatives of this group bear well-differentiated premolars and molars. Pyrotherium was scored 1 given that all but the first cheektooth are bilophodont with just minor differences in size across the row.

• 31. Skull size: small (0); or medium (1); or large (2). Assessment of skull sizes in this study is only approximate and relied on visual comparisons of skulls rather than on metric estimates given the degree of fragmentation and deformation of several key representatives of notoungulate families.