Supporting Information

Files in this Data Supplement:

Fig. 6. Undescribed skeleton of Plesiadapis cookei (UM 87990) from University of Michigan locality SC-117 (late Paleocene), Clarks Fork Basin, WY. (Scale bar: 9 cm.)

Fig. 7. Ignacius clarkforkensis skull and jaws in articulation, right lateral view (UM 108210; holotype). Scale is in millimeters.

Fig. 8. Ignacius clarkforkensis palate (UM 108210; holotype). Note the presence of a small, single-rooted P². Scale is in millimeters.

Fig. 9. Ignacius clarkforkensis, UM 108210 (holotype). (A-C) Left dentary in buccal, lingual, and occlusal views. (D-F) Right dentary in lingual, buccal, and occlusal views. Occlusal views of both dentaries are steriophotographs. (Scale bar: 5 mm.)

Fig. 10. Dryomomys szalayi, UM 41870 (holotype). Skull with left P²-M³ and right C¹-M³ in occlusal view (A) and left premaxilla with I^1-2 in buccal (B) and lingual (C) views. Scale (in A) is in millimeters.

Fig. 11. Dryomomys szalayi, UM 41870 (holotype). Left dentary with I₁-M₁, M₃ in buccal (A) lingual (B), and occlusal (C) views. (D) Right dentary with I₁, P₃-M₃ and right premaxilla (Pmx) with I^1-2 in buccal view (specimens in limestone). Scale: 5 mm.

Fig. 12. Diagram of the positions of bones (and especially phalanges) of Acidomomys hebeticus (in gray or indicated by an arrow) in upper (layer 1) and lower (layer 2) layers of Limestone Block M. (A-D) Boxed areas that are also illustrated at a higher magnification. Bones not attributed to A. hebeticus in layer 1 belong to two or more individuals of paramyid rodents. In A, note an articulated radius and ulna to the right of phalanges 5 and 6, below the crushed skull. Bones not attributed to A. hebeticus in layer 2 belong to a large rodent, an erinaceomorph insectivore, a carpolestid plesiadapiform and a marsupial. A. hebeticus was distinguished from other animals in the accumulation on the basis of associations, size, age and taphonomic state. A. hebeticus is a juvenile with porous bone, whereas the other taxa are in later stages of development. Lengths and length estimates are given for phalanges. Estimates for broken elements were made by assuming proportional constancy for all phalanges.

Fig. 13. Documentation of dental-postcranial associations for Ingacius clarkforkensis. (A) Bones of UM 108210 (Holotype) of Ignacius clarkforkensis in original positions. Note semiarticulation. The block that contained this specimen was collected from locality SC-62. The skull has been published in Bloch and Silcox (2001). (B) Bones of UM 82606 of Ignacius clarkforkensis in original positions. Note semiarticulation. The block that contained this specimen was also collected from locality SC-62; however, the relative location of UM 108210 to UM 82606 is undocumented. In any case, UM 108210 cannot represent the same individual as UM 82606 because it has an astragalus that is identical to that preserved with UM 82606, and it is from the same side of the body (right side). It does not preserve the dentition. However, its astragalus is identical to that preserved with UM 108210, thus allowing the identification of this specimen as another individual of Ignacius clarkforkensis. Analysis of this specimen combined with UM 108210 allowed estimates of interlimb and interbody segment proportions. As additional evidence supporting the validity of considering these two skeletons together in such a way, body mass estimates from elements of UM 82606 were consistent with those generated from the elements preserved in UM 108210. (C) Composite layout (UM 108210 and UM 82606) of Ignacius clarkforkensis in anatomical position. See A and B for original positional relationships of bones included in these specimens. (Scale bar: 3 cm.) Ast, astragalus; C#, cervical vertebra; C, capitate; c#, claw; Ca#, caudal vertebra; Cc, calcaneum; Cub, cuboid; Inmt, innominate; L#, lumbar vertebra; Mc, metacarpal; Mscn, mesocuneiform; Mt, metatarsal; Nv, navicular; S, sacrum; Sc ?, scapula?; T#, thoracic vertebra; R#, rib. Proximal phalanges = 4, 6, 8, **. Intermediate phalanges = 1, 2, 3, 7, ***. Asterisks indicate that the position of an element was not documented during preparation.

Fig. 14. Dryomomys szalayi, UM 41870 (holotype). (A) Skull and skeleton partially prepared from fossiliferous limestone, University of Michigan Locality SC-327, late Clarkforkian (cf-3) North American Land Mammal Age. (B) Composite map of the bones recovered. (Scale bar: 1 cm.)

Fig. 15. Dryomomys szalayi, UM 41870 (holotype). (A) Composite drawing of skull and skeleton with numbers on bones corresponding to those of the anatomical layout in B. (Scale bar: 1 cm.) (B) Skull and skeleton laid out in anatomical position with bones attributed to regions based on positional information. Note that B was made before all of the bones were prepared from the limestone and that not all bones depicted in A are in B. (Scale bar: 3 cm.)

Fig. 16. Documentation of postcranial associations for cf. Tinimomys. (A) Original positions of bones of a Wasatchian micromomyid (cf. Tinimomys). Although it lacks craniodental remains, this specimen does preserve many of the same elements as in the specimen in UM 41870 (Figs. 9-10), so we can confidently identify it to the familial level. Stratigraphic position and body size suggest that this skeleton belongs to Tinimomys. This specimen preserves complete tibiae, fibulae, and innominates as well as a complete radius. Subdivisions on scale: 1 mm. B, Some of the fully removed and prepared bones from this specimen laid out in anatomical position. Note that most of the right wrist, the left hand, many vertebrae, and the left distal tarsal row are preserved. (Scale bar: 3 cm.)

Fig. 17. Intermediate phalanx vs. proximal phalanx length. (A) Reduced Major Axis (RMA) regression of intermediate phalanx on proximal phalanx length (solid line; 2) for select nongliding mammals (small gray circles). Light gray area encompasses the 95% confidence limits for the relationship (y = 0.98 x -0.32, R² = 0.98, n = 18 taxa). The slope is indistinguishable from that expected for an isometric relationship between the two variables (1.0). Mitten-gliding dermopterans (filled black triangles) have relatively much longer intermediate phalanges than do paromomyids (large gray diamonds), other plesiadapiforms (open circles), and non-mitten-gliding gliders (open triangles). The latter three groups plot closer to nongliding mammals. More important than how closely all taxa follow the regression line is the fact that all are below the x = y line (1), except for the dermopteran. Only bats and clawed suspensory sloths share this position with Cynocephalus. (B) Box plots of logged interphalangeal ratios for a sample extant mammals and plesiadapiforms. Boxes encompass 50% of data points, with medians depicted as a horizontal line within them. Whiskers encompass all data. Numbers below boxes represent the sample size for data points comprising it. A, Acidomomys hebeticus; Ig, Ignacius clarkforkensis; Mcm, micromomyid, Dryomomys szalayi; Pl, Plesiadapis cookei. Model I ANOVA shows significant among-group variance. Paired comparisons of samples, grouped as depicted, show Cynocephalus to be significantly higher than micromomyids and paromomyids and these groups to be higher than other plesiadapiforms, arborealists, and terrestrialists at P < 0.05. However, consideration of individual points shows the significance of these differences to be more complicated (see discussion in text): note that the dashed oval encompassing Plesiadapis data is at the same level as that encompassing Ignacius data.

Fig. 18. Comparison of phalangeal proportions in the manual third digit ray normalized to the length of the proximal phalanx. Taxa include the dermopteran Cynocephalus, paromomyid (Ignacius), callitrichine Cebuella, and micromomyid Dryomomys. Rows from bottom to top are: metacarpal III, proximal phalanx, intermediate phalanx, and distal phalanx, respectively. Note that the euprimate Cebuella and both the paromomyid and micromomyid plesiadapiforms have much shorter intermediate relative to proximal phalanges than those of dermopterans. (Scale bars: 3 mm.)

Fig. 19. Elongation index of intermediate phalanx. Box plots of logged elongation indices for a sample extant mammals and plesiadapiforms. Boxes encompass 50% of data points with medians depicted as a horizontal line within them. Whiskers encompass all data. Numbers below boxes represent the sample size for data points comprising it. Model I ANOVA reveals significant among-group variance. Paired comparisons of these groups show terrestrial taxa, arboreal taxa, and Cynocephalus to be significantly different from each other at P < 0.05. Nonparomomyid plesiadapiforms, paromomyids, and micromomyids are significantly different from terrestrialists and Cynocephalus at P < 0.05 but not from one another or arborealists.

Fig. 20. Principal Components (PC) Analysis (PCA) of manual intermediate phalangeal morphology of extant vertical clingers (yellow) and fossils. Data are derived from nine different measurements taken on 147 intermediate phalanges representing 20 genera and 22 species. Taxa are color coded by functional group: pink, gliders; yellow, vertical clinging; blue, clawed, pronograde quadrupeds; green, nonclawed, specialized grasping; gray, fossil. We controlled for body size by running the analysis on variables that were the ratio of the value each raw measurement to the value of the geometric mean of all measurements on each specimen. Much of the comparative extant sample was provided by M. W. Hamrick (Medical College of Georgia, Augusta, GA) and was originally used in Hamrick et al. (1999). Eight plesiadapiform species are plotted, including the following: Carpolestes simpsoni, Nannodectes intermedius, N. gidleyi, Plesiadapis cookei, Acidomomys hebeticus, Ignacius clarkforkensis, and Dryomomys szalayi. For descriptions and illustrations of measurements taken, methods of data transformation, and analysis, see Hamrick et al. (1999). PC 1 represents 33.744% of the variance in the data set, PC 2 represents 27.056%, and PC 3 represents 12.93%. PC 1 is most strongly correlated to increasing mediolateral breadth of the distal articular surface, decreasing dorsopalmar depth of the proximal articular surface, and decreasing total length. Thus, narrow, deep, long phalanges have high-negative PC 1 scores. PC 2 is most strongly correlated to increasing shaft and proximal end dimensions, decreasing breadth of dorsal margin of distal articular surface and decreasing length. Thus, phalanges that are relatively elongate with gracile shafts and ends have high-negative PC 2 scores. Taxa that use their digits for clinging and/or climbing on large-diameter vertical supports have phalanges with low PC 1 scores (i.e., the callitrichine primate, Cebuella pygmaea), unlike arboreal primates that predominately grasp small-diameter supports, or terrestrial taxa that do not subject their phalanges to tensile forces, but instead load them in compression. Whether or not a taxon glides has little bearing on its phalangeal morphology, in general; however, note that gliders Anomalurus and Cynocephalus occupy a unique "glider space." (i.e., the lower left quadrant of the morphospace plot), characterized by low PC1 and PC2 scores. Paromomyids (Ignacius and Acidomomys, UM 82616 and UM 108210) have manual phalanges that plot with other plesiadapiforms.

Fig. 21. Intermediate phalanx elongation. (A) Intermediate phalanx length vs. cross-section at midshaft. Solid lines are RMA regressions of the natural log of intermediate phalanx length on the natural log of the midshaft area for select mammalian gliders [unfilled and solid black triangles (1) with black line], nongliding primates [small green circles (2) with green line], plesiadapiforms [large red and pink circles (3) with red line], and other extant nongliding mammals [small blue circles (4) with blue line]. A composite nongliding mammals regression is illustrated by a shaded gray area that encompasses the 95% confidence limits for the relationship (y = 0.54 x + 1.88, R² = 0.83). Gliders (1), including dermopterans (closed triangles), have more elongate intermediate phalanges than nongliders, generally falling outside the 95% confidence limits. Among nongliders, euprimates (2), and plesiadapiforms (3) have more elongate intermediate phalanges than do other nongliding mammals (4). Note that paromomyids (pink circles) lack the elongation characteristic of mitten-gliding dermopterans and are in the range of other plesiadapiforms (3). Equations for the regressions are the following: (1) y = 0.63x + 2.53, R² = 0.97, n = 3 taxa; (2) y = 0.43x + 2.07, R² = 0.71, n = 8 taxa; (3) y = 0.41x + 1.98, R² = 0.91, n = 6 taxa; and (4) y = 0.48x + 1.645, R² = 0.94, n = 10 taxa. Note that only Eq. 1, representing gliders, has a slope higher than that for isometry (0.50). All other groups are slightly negatively allometric with larger members having stumpier phalanges.

Fig. 22. Plesiadapis cookei (UM 87990) phalanx. (A) Lateral. (B) Proximal. (C) volar view. P. cookei clearly has a proximal articular surface that is dorsoventrally taller than it is mediolaterally wide. Also, the shaft is straight, or slightly dorsally convex (depending on how the central axis is defined) with no dorsal recurvature at the distal end. These are features claimed by Beard (1993b) to uniquely characterize Cynocephalus volans and paromomyid plesiadapforms. Plesiadapis seems to be too robust-limbed to have been a glider (Beard, 1989), is known from a specimen with soft tissue preservation that does not show a gliding membrane (P. insignis from Berru) and also lacks the unique interphalangeal proportions of C. volans. This brings into question the utility of dorsoventrally deep shafts and lack of dorsal recurvature of the phalanx as indicators of Cynocephalus-like functions in phalanges of smaller forms. (Scale bar: 5 mm.)

Fig. 23. Manual intermediate and proximal phalanges. (A) Ignacius clarkforkensis. (B) Cebuella pygmaea. (C) Micromomyid, Dryomomys szalayi.(D) Pteropus pumillio. (E) Cynocephalus volans. (F) Choloepus hoffmani. Ventrolateral view of intermediate phalanx (on the left) and proximal phalanx (on the right). Phalanges are normalized to length of the proximal phalanx to show variation in intermediate phalanx length, except for F, which is normalized to the length of intermediate phalanx in E. Note that suspensory taxa (D-F) differ from fossils (A and C) and vertical clinger and climber (B) in having intermediate phalanges that are relatively longer, with more ventrally oriented proximal ends and/or articular surfaces and less distinct flexor sheath tubercles. The proximal phalanges of the suspensory taxa have more deeply trochleated distal articular surfaces. (Scale bars: 3 mm.)

Fig. 24. Comparison of relative lengths of metacarpals. Right metacarpals V-III (left to right) in palmar view. (A) Cynocephalus volans. (B) Ignacius clarkforkensis. (C) Cebuella pygmaea. (D) Micromomyid, Dryomomys szalayi. Elements are standardized to the length of the third metacarpal. Note that A is unique in having a fourth and fifth metacarpal that extend distally beyond the third metacarpal.

Fig. 25. Relative lengths of metacarpals and metatarsals for Cynocephalus and Dryomomys. Third digit rays of hands and feet of Cynocephalus (Left) and micromomyid, Dryomomys szalayi UM 41870 (Right). The rays are standardized to the lengths of the metatarsals. Note that manual elements are longer in Cynocephalus, whereas the pedal elements are longer in the micromomyid.

Fig. 26. Forelimb elements of Ignacius clarkforkensis, (A) Right humerus in (1) anterior and (2) posterior views. (B) right radius in (1) posterior and (2) lateral views. (C) Right ulna in (1) medial, and (2) lateral views. Note that the proximal-most part of humerus is not preserved, nor are the distal tips of the radius and ulna. (Scale bar: 5 mm.)

Fig. 27. Proportions of select elements among different gliding and nongliding taxa. For each lettered specimen, there are three numbered elements, or sets of elements: (i) right ulna and radius in medial view, (ii) right humerus in anterior view, (iii) articulated sacrum and innominates in ventral view. (A) Cynocephalus volans (colugo). (B) Glaucomys (flying squirrel). (C) Ignacius clarkforkensis. (D) Sciurus niger (gray squirrel). Features present in A and B can be taken to be convergently acquired for gliding. Some of these features include a narrow ulna with deep trochlear notch, which is fused to the radius distally; a radius that is substantially longer than the humerus; a distal humerus that is mediolaterally narrow, with large capitular area, and that is substantially longer than the innominate; an innominate with a large distance between the inferior end of the auricular surfaces and the acetabulae, narrow iliac crest, long ilium relative to ischium, and short, caudally situated pubic symphysis. Ignacius (C) shares more features in common with Sciurus (D) than it does with either of the two gliding taxa. (Scale bar: 3 cm.)

Fig. 28. Comparison of distal radii of Ptilocercus lowii and micromomyid. (1) distal, (2) ventral, and (3) dorsal views of the radii of A, primitive euarchontan mammal, Ptilocercus lowii, and (B) cf. Tinimomys (unnumbered, semiarticulated specimen from SC-26). Note that they are similar in having a ventrally oriented distal articular surface for the proximal carpal row that is also deeply cupped and marked by a prominent ridge projecting from the extensor surface. Note also that the tubercle separating the 1st from the 2nd extensor compartments (A1, projection to the lower right) is much larger in P. lowii. Thus, unlike in sciurids, hyperdevelopment of this tubercle is not clearly associated with gliding. (Scale bar: 3 mm.)

Fig. 29. Selected trunk indices comparing Ignacius and Cynocephalus. Vertebral column and limb indices for a subset of the comparative sample of extant taxa used in this study are presented with individuals placed in order of increasing index to aid in identification of functional trends. Ignacius clarkforkensis (UM 108210 and UM 82606) is represented by a black bar in each plot, whereas Cynocephalus volans is represented by an open bar. Ignacius clarkforkensis has indices that, in general, are separated from those of gliders and suspensory taxa by intermediate index values of agile arborealists and scansorialists. *, One parameter in the index has been estimated; **, both parameters were estimated.

Fig. 30. Vertebral profiles comparing Ignacius and Cynocephalus. Comparison of vertebral proportions using the method of Gingerich, (1998). The y axis depicts the logged value of two separate measurements, (1) vertebral body height and (2) vertebral body length normalized to the average height of the first six anterior thoracic centra. Thus, if the boundary of a bar (vertebral body measurement) is positive, then it is greater than the average height of the first six thoracics, whereas if it is negative, the reverse is true. In white bars, the upper boundary represents the length (superoinferior) of the body, whereas the base of the bar represents the height (anteroposterior). For black bars, the reverse is true (note that in all taxa depicted it is only the atlas that has such proportions). Thus, the shorter the bar, the closer the vertebral body is to being square in lateral view. The gray areas depict those vertebrae most closely associated with pectoral (anterior thoracics) and pelvic (sacral) girdles. The gray bar represents the anticlinal vertebra, the boundary between vertebrae with caudally projecting spinous processes (superior vertebral positions) and cranially projecting ones (inferior vertebral positions). In Ignacius clarkforkensis, (UM 108210 and UM 82606) the neck is short, trunk vertebrae increase in height and length posteriorly, the sacrum is robust, and the tail is long and robust. Such features suggest a relatively posteriorly shifted center of mass of the axial skeleton. Arborealists, Sciurus and Saguinus, depicted on the right have proportions similar to each other and Ignacius. Cynocephalus volans, on the other hand, exhibits a different pattern of vertebral proportions. In C. volans, the neck is long, trunk vertebrae remain roughly constant in size throughout the column, the sacrum is gracile, and the tail is shorter and more slender. Such features suggest a more anteriorly positioned center of gravity. We interpret proportional features in the paromomyid to be reflective of hindlimb dominance in forward locomotion, whereas those of C. volans reflect a need for maneuverability while gliding and equal emphasis on the fore and hindlimbs in suspensory locomotion.

Fig. 31. Comparison of antepenultimate lumbar vertebra of Ignacius and Cynocephalus. Antepenultimate lumbar vertebrae in lateral view of Saguinus mystax (A), Ignacius clarkforkensis UM 82606 (B), and Cynocephalus volans (C). A and B have longer, more ventrally canted transverse processes, more cranially extended zygapophyses, and narrower, more cranially angled spinous processes than C. volans. The suite of features characterizing A and B reflects use of pronograde postures with a habitually ventriflexed back and allows a large range of powerful flexion and extension in the lumbus relative to that possible in the back of C. volans.

Fig. 32. Vertebral elements from different regions of Ignacius. Representative vertebrae of Ignacius clarkforkensis (UM 82606). Note the cranioventrally oriented transverse processes in Ap lumbar vertebra; the lack of a well developed spinous process on the first sacral vertebra (which is unbroken) and large vertebral canal in the sacrum; the large vertebral canal in caudal II; and the robusticity and length of caudal VI. These features are not expected for an animal predominantly using suspensory or gliding behaviors. (Scale bar: 5 mm.)

Fig. 33. Regression of lumbar vertebrae against body mass proxies. (A) Natural log femoral TA vs. average depth of the posterior three lumbar vertebrae in bounding (open diamonds) and nonbounding/ambulatory (gray diamonds) taxa. The depth represented is the dorsoventral distance between the tip of the mammillary process on the prezygapophysis (dorsal) to the tip of the transverse process (ventral). The leverage of the erector spinae muscles in the sagittal plane is determined by their dorsoventral dimensions and the dorsoventral distance of their attachments from the vertebral body. These dimensions are captured by the measurements described above. Femoral TA (cross-sectional area at midshaft) is a body-size proxy. Because TA is an area and the y value is a length, isometry is represented by a slope of 0.5. Thus, both lines show slight positive allometry. Taxa that leap and bound incorporate forceful flexion and extension of the back into their gaits and require better leverage and more force from their erector spinae muscles than taxa that do not. Thus, it is not surprising that bounders scale differently from nonbounders, with a deeper vertebra for a given femur area. Ignacius (large black circle) falls with bounding taxa, yet again. Likewise, Cynocephalus (black triangle) falls with the other group. Nonbounder line: y = 0.512x + 0.12; R² = 0.73. Bounder line: y = 0.524x + 0.74; R² = 0.93. Natural log average spinous process length vs. depth for the posterior three lumbar vertebrae of bounding (open diamonds) and nonbounding/ambulatory (gray diamonds) taxa. The bounder line is lower because the spinous process of the posterior three lumber vertebrae are dorsoventrally longer and craniocaudally narrower than those of nonbounders. Narrower spinous processes result in more sagittally mobile backs, required for a bounding gait, which utilizes substantial flexion and extension of the vertebral column to increase the stride length. Ignacius (large black circle) clearly falls with bounding taxa in having very narrow lumbar spinous processes, whereas Cynocephalus (black triangle) falls with the other group. Nonbounder line: y = 0.79x + 0.6; R² = 0.77. Bounder line: y = 0.91x - 0.45; R² = 0.92.

Fig. 34. Comparison of innominates of Ignacius and Cynocephalus. (A) Left innominates in lateral view: Saguinus mystax; Ignacius clarkforkensis; Cynocephalus volans. Elements are standardized to ischium length. (B) Left innominate in lateral view (on left) and medial view (on right) Ignacius clarkforkensis (UM 82606). [Scale bar: 5 mm (B).] In Ignacius and Saguinus, note the relatively shorter and more flaring ilia, the longer and/or more superiorly extended pubic symphysis, the larger ischial tuberosity, and the inferiorly positioned ischial spines.

Fig. 35. Functional trends in intermembral index in different groups of mammals. Intermembral indices of some of the taxa used in the comparative sample of this study, presented in order of increasing index to aid in identification of functional trends. Taxa from a given higher level clade are designated by a unique letter above the bars representing them. R, Rodentia; C, Callitrichinae; T, (tree shrew) Scandentia; M, marsupial. Cynocephalus is represented by an open bar, whereas Ignacius is in black to highlight its position. An asterisk in front of a taxon name indicates that it is extinct. Note that in each clade of mammals, the most terrestrially adept members (wavy lines or diagonal) have the lowest intermembral indices, whereas the arboreal ones that spend time on large-diameter vertical supports have higher indices (diagonal to cross-hatched) and the gliders (and suspensory taxa) have the highest (closely packed horizontal lines). For instance, among Callitrichinae, Saguinus is more terrestrially adept than Cebuella, which, in turn, is more committed to foraging on vertical supports (e.g., Youlatos, 1999). Furthermore, although Callimico goeldii is not plotted, it has been noted that it has the shortest intermembral index of any callitrichine; C. goeldii is not known to use large-diameter vertical supports or the exudate resources procurable there. Note that the ground squirrel Citellus has shorter forelimbs than the tree squirrel Sciurus. Both have shorter forelimbs than the gliding squirrels Glaucomys sabrinus and volans. This trend holds for marsupials as well, with the locomotor generalist Trichosurus exhibiting shorter forelimbs than the glider, Petaurus. Behavioral overlap occurs in the region of high indices, such that a high intermembral index is not evidence of gliding by itself. There does, however, seem to be a lower limit to the intermembral indices exhibited by gliders. That limit appears to be somewhere ≈80 [although out of eight species of flying squirrels, one (Eoglaucomys) has an index <80 (Thorington and Heaney, 1981)]. Thus, although a strong case for gliding in a fossil taxon cannot be made on the basis of a high index alone, a strong argument against it can be made on the basis of a low index. Almost no extant gliders have an index as low as that of Ignacius clarkforkensis This fact, combined with other information discussed above, makes gliding an extremely unlikely locomotor mode for Ignacius.

Fig. 36. Illustration of hindlimb elements of UM 82606. Right femur of Ignacius clarkforkensis in (1) anterior, (2) posterior, and (3) distal views. View 3 is oriented so that the greater trochanter and fovea capitis femoris form a horizontal line on the page. That line is projected on the figure and labeled FGSp. Note that the condyles face posterolaterally and would have contributed to postures in which the feet were widely spaced (abducted), uncharacteristic of suspensory taxa. Instead, such morphology is seen in scansorialists such as sciurids and tupaiid tree shrews. The right tibia of Ignacius clarkforkensis is depicted in proximal (5), anterior (7) and medial (9) views. The right fibula is shown in proximal (4), anterior (6), and medial (8) views. The broad shelf on the proximal fibula (4), oriented perpendicular to the shaft axis, gives it mobility with respect to the tibia. On the tibia (7), lateral inclination of the astragalar facet presumably accommodates asymmetry of margins of tibial facets on the astragalus during dorsiflexed postures. On 9, note that the patellar tendon groove (Ptg) is distally positioned. FGSp, Plane defined by Fovea capitis femoris-Greater trochanter-Shaft. (Scale bar: 5 mm.)

Fig. 37. Comparison of femoral condyles of Ignacius among pronograde and antipronograde mammals. Distal femora in posterior view with the distal end pointing up. (A) Galagoides demidovii (left). (B) Saguinus sp. (left). (C) Choloepus sp. (left). (D) Rattus sp. (left). (E) Ignacius clarkforkensis (right). (F) Nycticebus sp. (right). All taxa except C and F have a wedge-shaped lateral condyle and distal condylar eminences that are more closely appressed than the proximal condylar eminences. In B and E, as evident from the shallowness of the sulcus between the condyles on the distal margin of the femur, the tibial articular surface is limited to the posterior aspect of the condyles. The unique morphology of C and F reflects the capacity for extreme extension and habitual use of postures that require extended limbs. (Scale bar: 5 mm.)

Fig. 38. Comparison of tibia of Ignacius and Cynocephalus in lateral view. Right tibia in medial view showing angle formed between shaft and medial facet of tibial plateau. From left to right, taxa depicted (and the angle formed) are the following: Nycticebus sp. (73); Cynocephalus volans (69); Cebuella pygmaea (65); Ignacius clarkforkensis (64); Leontopithecus sp. (62); Smilodectes gracilis (60). Nycticebus and Cynocephalus, which frequently use extended limb postures, have medial facets at more of an angle to the shaft (they approach perpendicular). Taxa that are more pronograde, or interpreted to have been leapers, have medial facets at less of an angle to the shaft (closer to parallel). (Scale bar: 1 cm.)

Fig. 39. Patellar groove comparison. From top to bottom are a callitrichine euprimate (Saguinus), Ignacius clarkforkensis (UM 82606), and Cynocephalus volans, in medial and anterior view. Note that, despite it's being an agile arborealist, the distal femur in Saguinus is nearly identical to that in Ignacius. Both these taxa differ only slightly from the condition in C. volans with respect to patellar groove morphology. However, other features discussed and figured in this chapter differentiate the fossil and Saguinus from C. volans, showing the former two to be more agile in pronograde postures than the latter. Images are standardized to distal femur breadth.

Fig. 40. Cladogram based on dental, cranial, and postcranial characters (see Fig. 4) with the form of the I1 optimized. Note the extensive homoplasy, with the evolution of an apical division occurring at least three times in early primate evolution (at the base of the group and independently in Paromomyoidea and Plesiadapoidea). Thus, recent claims that the enlarged, multicuspate incisors of certain "plesiadapiforms" are clear evidence against a plesiadapiform euprimate link (Kirk et al., 2003) are in need of reevaluation. Specific illustrations modified from Clemens (2004) (Purgatorius; Fig. 5), Rose et al. (1994) (Micromomyidae, Microsyopidae, Paromomyoidea, Saxonellidae, Carpolestidae, Plesiadapidae; Fig. 3), Silcox (2001) (Pandemonium; Fig. 3.7C), Beard and Wang (1995) (Chronolestes; Fig. 6), and Szalay and Delson (1979) (Nannopithex; Fig. 130D).

Fig. 41. Results of the Total Distribution Analysis from MacClade 3.08a (Maddison and Maddison, 1992). Patterned branch segments indicate the place of origin of the various clades.

Fig. 42. Results of the First Occurrences Analysis from MacClade 3.08a (Maddison and Maddison, 1992). Patterned branch segments indicate the place of origin of the various clades.