Diegoaelurus vanvalkenburghae, Zack & Poust & Wagner, 2022

Zack, Shawn P., Poust, Ashley W. & Wagner, Hugh, 2022, Diegoaelurus, a new machaeroidine (Oxyaenidae) from the Santiago Formation (late Uintan) of southern California and the relationships of Machaeroidinae, the oldest group of sabertooth mammals, PeerJ 10, pp. 1-27 : 6-15

publication ID

https://doi.org/ 10.7717/peerj.13032

publication LSID




persistent identifier


taxon LSID


treatment provided by


scientific name

Diegoaelurus vanvalkenburghae

sp. nov.



( Figs. 2 5 View Figure 2 View Figure 3 View Figure 4 View Figure 5 , Table 2 View Table 2 )

Holotype. SDSNH 38343 , right dentary preserving i2, c, p3, and m1-2.

Type Locality. SDSNH locality 3276 (Jeff’s Discovery), Oceanside , San Diego County, California . As described by Walsh (1996), the Jeff’s Discovery assemblage comprises several adjacent localities first discovered by Mr. Jeff Dahlgren. The Jeff’s Discovery fauna has produced a distinctive fauna including a number of taxa otherwise rare in the San Diego Eocene. Aside from Diegoaelurus vanvalkenburghae , distinctive elements of the Jeff’s Discovery fauna include the hyaenodontid Limnocyon sp. , carnivoran Tapocyon dawsonae , tapiroid Hesperaletes borineyi , and basal artiodactyl Ibarus sp. ( Walsh, 1996; Wesley & Flynn, 2003; Colbert, 2006).

Stratigraphy and Age. Member ‘‘C’’ of the Santiago Formation, late Uintan (Ui3) North American Land Mammal Age (NALMA), middle Eocene ( Walsh, 1996).

Diagnosis. As for the genus.

Etymology. Named in honor of Dr. Blaire Van Valkenburgh, in recognition of her substantial contributions to our understanding of iterative evolution in carnivorous mammals and saber-tooth paleoecology.

Description. The dentary is intact along most of the horizontal ramus, but is broken posteriorly, with portions of the ascending ramus, including the condylar process and posterior margin of the coronoid process missing ( Fig. 2 View Figure 2 ).

A flange projects ventrally and somewhat laterally ( Figs. 2A View Figure 2 , 3 View Figure 3 ), extending the depth of the anterior dentary by almost a third. The shape of the flange is curved, contrasting with the triangular flange in Machaeroides , and its posterior edge is slightly concave marking it as a very distinct feature and contrasting with a linear posterior margin in Machaeroides . A thin ridge borders it anteriorly, narrowing to a thin raised blade along the anterior surface below the canine. No mental foramina are present immediately adjacent to the symphysis on the anterior aspect of the jaw, unlike in Machaeroides simpsoni , though three are present on the lateral surface ( Fig. 2A View Figure 2 ), one directly below the p2 alveolus, one placed anteroventrally just above the flange, and a third, smaller foramen about mid-height at the level of the distal margin of the canine. The two larger foramina are directed laterally, while the third is directed anterodorsally, at a very shallow angle to the dentary surface. The placement of these foramina differs from Machaeroides eothen where one foramen is more anterior, very close to the front of the jaw. This difference may relate to the expansion of the flange itself in SDNHS 38343.

The symphysis is well developed and vertically oriented, with a straight anterior margin and a bilobate posterior ( Fig. 2D View Figure 2 ). The lineations on the symphyseal surface trend in two different directions. In the lower lobe along the ventral portion of the flange and opposite the upper portion of the flange, the lineations run strongly anterior/posteriorly, with even the lowest few angled down anteriorly below the plane of the horizontal. The upper lobe by contrast has linear features that splay outward and upward, primarily at about a 60 degree angle from horizontal spanning from the midline to the alveoli of the incisors. The plane of the symphysis is only slightly angled away from the long axis of the dentary, meaning the jaw lacks the wide anterior expansion seen in M. simpsoni .

The horizontal ramus posterior to the flange is rectangular and transversely compressed. Between the p3 and the m1 the dorsal edge of the mandible dips down, a feature which may represent the effects of long finished bone resorption and filling of the alveolus and possibly support the idea that the adult p4 tooth was lost during early life. A furrow visible on the medioventral edge represents a break that exposed the interior cortical bone ( Fig. 2D View Figure 2 ).

Interestingly, this broken edge is filled with sediment; the break was not sustained during final erosion or collection, rather during biostratinomy, or perhaps a prior episode of erosion and redeposition. Given the rather complete nature of the specimen, presence of teeth, and lack of surface weathering this is an oddly contradictory taphonomic feature.

Though slightly damaged the angular process is present giving us the length of the jaw, but unfortunately the condyle and much of the coronoid process are broken ( Figs. 2A, 2D View Figure 2 ).

The anterior and ventral borders of the masseteric fossa are clearly defined, showing that it would have been a deeply incised and large area of muscle attachment in life. The anterior margin of the masseteric fossa extends to a level below the back of the trigonid of m2.

The dorsal margin of the fossa is less incised than the other borders and ends only slightly below the preserved portion of the coronoid process. Although the posterior margin of the coronoid process is absent, it appears to represent nearly the full height, extending to a level just above the height of the m2. Such a low coronoid process is typical of saber-tooth carnivores ( Matthew, 1910; Emerson & Radinsky, 1980).

As in other machaeroidines, the dentary preserves evidence of only two lower incisors, the more mesial represented by a root and partial crown and the more distal by an alveolus ( Figs. 2B 2C View Figure 2 , 3 View Figure 3 ). There is a small space mesial to the first incisor, but there is no evidence of an alveolus for a third incisor. Reduction in the number of lower incisors is shared with other machaeroidines in which this region is preserved: Apataelurus kayi and Machaeroides eothen ( Scott, 1938; Gazin, 1946). Denison (1938) hypothesized that i1 was lost in A. kayi , and, in the absence of contradictory evidence, the lower incisors of D. vanvalkenburghae are tentatively considered i2-3.

Much of the apical crown of i2 is broken ( Figs. 2 3 View Figure 2 View Figure 3 ). The preserved portion of the tooth, including the root, is compressed mesiodistally and slightly inclined buccally. The crown appears to have been relatively tall. The buccal face of the crown is gently convex and vertical, while the lingual face is flat and slopes basolingually. The lingual surface is wider than the buccal surface, making the crown wedge-shaped in occlusal view. The preserved mesial and distal margins are vertical and parallel, but not enough of the crown is preserved to determine if it was pointed or more spatulate. There is no development of basal cuspids or cingulids, and the preserved structure of the crown is simple.

The alveolus of i3 is immediately distal to i2 ( Figs. 2 3 View Figure 2 View Figure 3 ). The alveolus indicates that the tooth was somewhat larger than i2, while the shape of the alveolus indicates that i3 was also compressed mesiodistally.

The unworn lower canine is nearly completely preserved, a first for Machaeroidinae ( Figs. 2 3 View Figure 2 View Figure 3 ). The tooth is substantially larger than either incisor and would also have been much taller than i2 if the latter were complete. The canine is implanted vertically, lacking the buccal inclination of i2. As with the incisors, the crown of c1 is mesiodistally compressed, with an oblique buccolingual axis, oriented posteromedially relative to the dentary. The crown is tall and recurved lingually, tapering to a sharp point. The buccal surface is convex, while the lingual surface is formed by mesiolingual and distolingual faces that meet at a distinct angle, forming a peak that runs down the center of the lingual surface of the crown. Where the buccal and lingual surfaces of the crown meet, blunt, mesial and distal carinae are present. Serrations extend the length of the distal carina ( Fig. 4A View Figure 4 ), while the mesial carina lacks serrations. The distal carina is stronger and extends further basally than the mesial carina. There is no development of cingulids or a basal heel, but the thin enamel extends modestly further basally on the buccal and lingual surfaces of the crown than on the mesial or distal surfaces.

A large diastema separates the canine from p3 ( Figs. 2A 2C View Figure 2 ). Within this diastema, near the midpoint but closer to p3 than to the canine, is a single, small alveolus, for a single-rooted p2. The size of the alveolus indicates that p2 was likely the smallest lower tooth, approximately half the size of i2.

The first preserved cheek tooth is a double-rooted p3 ( Fig. 2 View Figure 2 ). The protoconid is broken near its base, but the remainder of the crown is intact. The tooth is approximately 57% the length of m1, relatively larger than in Apataelurus kayi (42%) but substantially reduced relative to Machaeroides eothen (88%). Mesially, the crown has a small but differentiated paraconid, deflected slightly lingually from the protoconid. This contrasts with both M. eothen and A. kayi , which lack paraconids on p3. The protoconid was large, dominating the crown, although its height cannot be determined. In cross-section, the broken base of the protoconid is lenticular. Distal to the protoconid, the talonid bears a single cusp, a large, low hypoconid positioned close to the buccal margin of the crown and aligned with the distal carina of the protoconid. The hypoconid has a steep buccal and more shallow lingual slope. Lingual to the hypoconid, the margin of the talonid is slightly concave at the distolingual corner of the crown, with very faint cuspids evident on either side of the concavity.

Immediately posterior to p3, p4 is absent. Alveoli for p4 are not visible externally nor are they evident in a radiograph of the specimen ( Fig. 5 View Figure 5 ), indicating that the absence of p4 reflects either pathology or in vivo loss well prior to death rather than postmortem damage. The size of the gap between p3 and m1 indicates that p4 was substantially longer than p3, but shorter than m1 and much shorter than m2.

Posterior to the missing p4, m1 is an elongate, narrow tooth ( Fig. 2 View Figure 2 ). The trigonid of m1 is bicuspid, with no trace of a metaconid. The paraconid is transversely compressed and relatively elongate and low. The paraconid portion of the paracristid is oriented obliquely mesiolingually, with the apex of the paraconid mesiolingual to the protoconid apex. The crest slopes up from the carnassial notch, but the lingual half is flat topped due to wear.

On the buccal surface of the paraconid, close to its mesial margin, there is a short, vertical mesiobuccal cingulid, defining an embrasure for the back of the p4 talonid. The cingulid is restricted to this vertical portion and does not continue around the base of the cusp. On the lingual side of the trigonid, the paraconid is separated from the protoconid by a deep groove extending lingually and slightly distally from the carnassial notch.

The protoconid of m1 is larger and substantially taller than the paraconid and is also transversely flattened. Two crests descend its apex. The trenchant protoconid portion of the paracristid is oriented mesially and slightly buccally and descends from the apex to the carnassial notch. The more weakly defined protocristid is directed distally and descends the protoconid vertically towards the talonid.

The talonid of m1 is much shorter than the trigonid and lower than both trigonid cusps. Buccally, there is no hypoflexid separating the trigonid from the talonid. The talonid is narrow and trenchant, with no entoconid or entocristid lingually. The talonid is dominated by the hypoconid, which has a vertical buccal slope and a somewhat gentler lingual slope. The hypoconid is basically flat-topped, but a low, centrally positioned apex is present. Running directly mesial from the hypoconid apex is a short cristid obliqua. This crest ends relatively abruptly at the mesial end of the talonid, where it is separated from the protocristid by a small carnassial notch. Nearly directly distal to the hypoconid, and connected by a weak postcristid, is a small hypoconulid. The hypoconulid is slightly lower than the hypoconid and centrally positioned on the distal margin of the crown. Aside from the mesiobuccal cingulid, there is no development of cingulids on m1.

The m2 of SDSNH 38343 is larger than m1 ( Fig. 2 View Figure 2 ). The trigonid of m2 is equivalent in length to the entirety of m1, and the trigonid cusps on m2 are considerably taller than those on m1. Aside from size, the morphology of the trigonid of m2 is very similar to m1, lacking a metaconid and with a nearly longitudinal paracristid. Additionally, the protoconid of m2 is more distally reclined than that of m1, with a vertical distal carina and more elongate mesial crest, resulting in a relatively longer paracristid than on m1. The distal carina bears serrations, similar in morphology to the canine ( Fig. 4B View Figure 4 ). As with m1, there is a small, vertical mesiobuccal cingulid that helps define an embrasure for the m1 hypoconulid.

The talonid of m2 is relatively smaller and lower than on m1. The basic structure of the talonid is similar, with a hypoconid and hypoconulid, and no entoconid or entocristid.

However, the hypoconulid is more lingually positioned relative to the hypoconid, with the result that the postcristid is oblique rather than longitudinal. The hypoconulid is also slightly larger than on m1, and its apex is better separated from the hypoconid.

Comparisons. The type specimen of Diegoaelurus vanvalkenburghae can be directly compared with all named machaeroidines except Isphanatherium ferganensis , which is known only from an isolated upper molar. Numerous features of the lower dentition and dentary link D. vanvalkenburghae to other machaeroidines, including a broadened symphyseal region of the dentary with a ventral mandibular flange below the canine and mesial premolars, low coronoid process, reduced premolars mesial to p4, loss of m3, and m1-2 with hypercarnivorous features including open trigonids, elongate, sectorial paracristids, and reduced metaconids and talonids.

Compared with the earliest known machaeroidine, Machaeroides simpsoni , D. vanvalkenburghae has a deeper mandibular flange with a more rounded ventral margin, but the anteroposterior length of the flange is similar in the two taxa. The posterior dentary is unknown in M. simpsoni , preventing comparisons, as is the lower canine. Machaeroides simpsoni retains a small, double-rooted p1, which is lost in D. vanvalkenburghae , and a double-rooted p2, which is single-rooted in the San Diego taxon. Known lower molars of M. simpsoni are fragmentary, but the relative molar sizes in the two species do appear to be similar, though m2 does appear to be proportionally larger in D. vanvalkenburghae . Both taxa have elongate, longitudinal paracristids, consistent with hypercarnivory. On the trigonids, M. simpsoni retains small but distinct metaconids, which are lacking in D. vanvalkenburghae . In general, the morphology of Machaeroides simpsoni appears less hypercarnivorously adapted than that of Diegoaelurus vanvalkenburghae .

The younger type species of Machaeroides , M. eothen , is better known than M. simpsoni . As with the latter species, M. eothen has a less developed, more triangular mandibular flange that is similar in length to D. vanvalkenburghae . More posteriorly, the coronoid process is somewhat taller than in D. vanvalkenburghae . The lower canine is incompletely preserved in M. eothen , inhibiting comparisons to D. vanvalkenburghae . Serrations are absent from the distal carina of the canine of USNM 17059, but this surface is worn from contact with the upper canine ( Gazin, 1946, p. 343). Again resembling M. simpsoni , Machaeroides eothen has a double-rooted p1 and p2, differing from D. vanvalkenburghae , which lacks p1 and has a single rooted p2. The crown of p3 is preserved in M. eothen . It differs from D. vanvalkenburghae in lacking a paraconid and in being larger relative to the remainder of the dentition. As with M. simpsoni , m1-2 of M. eothen are more similar in size than in D. vanvalkenburghae , in which m2 is distinctly larger than m1. The trigonids of M. simpsoni are taller and less elongate than in the San Diego form. Again as in M. simpsoni , M. eothen retains metaconids on m1-2, unlike D. vanvalkenburghae . Talonid morphology is documented in the Bridgerian taxon, which retains low entocristids, contrasting with D. vanvalkenburghae , which has fully trenchant talonids. Talonid size is similar. As was the case with M. simpsoni , the morphology of Machaeroides eothen is similar to but less hypercarnivorously adapted than that of Diegoaelurus vanvalkenburghae .

Comparisons with the Chinese machaeroidine Apataelurus pishigouensis are limited due to the fragmentary nature of the type and only specimen of this taxon, a dentary fragment with p4-m1. Most of the dentary is not preserved, and the shape of the mandibular flange cannot be evaluated. However, the flange appears to have been larger than in D. vanvalkenburghae , extending to a point beneath the mesial margin of p4. The m1s of both species are similar in having a relatively low, elongate trigonid. Both taxa also lack metaconids. Unlike D. vanvalkenburghae , the talonid of A. pishigouensis is very short, comprising approximately a quarter of the length of the tooth. Based on the comparisons that can be made, both D. vanvalkenburghae and A. pishigouensis are relatively hypercarnivorous machaeroidines. However, the apparent difference in the size of the dentary flange suggests differences in saber morphologies and, potentially, in feeding ecology.

The type species of Apataelurus , A. kayi , is better known than A. pishigouensis , permitting more extensive comparisons with Diegoaelurus vanvalkenburghae . The mandibular flange of A. kayi is incompletely preserved, but it extends to the level of p4, indicating a larger flange than in D. vanvalkenburghae . The coronoid process of A. kayi is much lower than in D. vanvalkenburghae , lower than the trigonid of m2. Asingle-rooted p1 is retained in A. kayi , contrasting with the loss of this tooth in D. vanvalkenburghae . Similarly, A. kayi , has a double-rooted p2, contrasting with the single-rooted tooth of D. vanvalkenburghae . The premolars of A. kayi are closely spaced and separated from the canine by a large diastema. In contrast, p2 is separated from p3 by a distinct diastema in D. vanvalkenburghae . As a result, there is less of a postcanine diastema in the San Diego form, despite the loss of p1, and it is possible shortening of this portion of the dentary helps to explain the loss of p1 in D. vanvalkenburghae . More posteriorly, A. kayi and D. vanvalkenburghae share a reduced p3, although A. kayi lacks the paraconid present in D. vanvalkenburghae .

The m2 of Apataelurus kayi is much larger relative to m1 than in D. vanvalkenburghae .

Both taxa share low, open trigonids on both molars, and both lack metaconids and entocristids. However, A. kayi has much more reduced talonids, with a short m1 talonid and a rudimentary, unicuspid m2 talonid, contrasting with the large, bicuspid talonid present in D. vanvalkenburghae . Overall, while Diegoaelurus vanvalkenburghae is advanced relative to Apataelurus kayi in its degree of premolar reduction, the remainder of the morphology of the Utah taxon appears more specialized. As with A. pishigouensis , the apparently larger mandibular flange of A. kayi suggests further differences in saber morphology that are presently undocumented.

Phylogenetic Results. Analysis of the character-taxon matrix produced a single most parsimonious tree (L: 30; CI: 0.83; RI: 0.61) ( Fig. 6 View Figure 6 ). The analysis recovers a monophyletic Machaeroides as the sister taxon to a clade comprising Diegoaelurus and a monophyletic Apataelurus . Of the three ingroup nodes, only the node subtending middle Eocene machaeroidines ( Apataelurus plus Diegoaelurus ) has Bremer support greater than one, with two additional steps required to recover trees that do not include this grouping.

Monophyly of Machaeroides is supported by a single unambiguous synapomorphy, presence of a double-rooted p1 (character 8, state 1). In contrast, Apataelurus kayi and both outgroups have a single-rooted p1 (state 1), while D. vanvalkenburghae lacks p1 (state 2). Monophyly of middle Eocene machaeroidines is supported by two unambiguous synapomorphies, loss of the metaconid on m1-2 (character 17, state 1) and absence of an entocristid on m1 (character 18, state 1). Monophyly of Apataelurus is supported by two unambiguous synapomorphies, reduction of the length of the m1 talonid to approximately 1the length of the crown (character 19, state 1) and posterior expansion of the mandibular 4 flange to a point below p3 (character 21, state 1).

Unambiguous autapomorphies are only identified in two of the five ingroup taxa. Machaeroides eothen has seven autapomorphies. However, all but one of these is a feature of the upper dentition, which cannot be scored for Diegoaelurus or either species of Apataelurus . The remaining feature reconstructed as an autapomorphy of M. eothen is the presence of a symmetrical p4 protoconid, which contrasts with M. simpsoni , both Apataelurus species, and the oxyaenine outgroup. Three autapomorphies are reconstructed for D. vanvalkenburghae . Two of these describe reduction of the anterior premolars: loss of p1 (character 8, state 2) and presence of a single-rooted p2 (character 9, state 1). The third autapomorphy of D. vanvalkenburghae is the presence of a paraconid on p3 (character 11, state 1). Neither species of Apataelurus is reconstructed as having any autapomorphies. This is initially surprising given the distinctive morphology of A. kayi . However, the fragmentary nature of A. pishigouensis renders optimization of most of these features ambiguous. Thus, features such as an enlarged m2 (character 16, state 1), reduced m2 talonid (character 17, state 1), and low coronoid process (character 23, state 1) are equally parsimonious as Apataelurus synapomorphies or A. kayi autapomorphies. Presence of a tall p4 paraconid (character 13, state 1) and broad p4 talonid (character 14, state 1) is shared by both Apataelurus species but cannot be evaluated in Diegoaelurus , while a mandibular condyle positioned below the toothrow (character 24, state 1) cannot be evaluated in either D. vanvalkenburghae or A. pishigouensis . One distinctive feature of Diegoaelurus , presence of a rounded border to the mandibular flange (character 22, state 1) cannot be evaluated in either Apataelurus species.

Table 2 Measurements (mm) of the holotype of Diegoaelurus vanvalkenburghae.

SDSNH 38343
Length of dentary 71.5
Length of dentition (i2 alveolus-m2) 51.4
Length of cheek dentition (p2 alveolus-m2) 35.2
Depth at flange 23.9
Depth of flange below line of jaw 5.3
Depth below m1 11.6
m2 length 10.4
width 3.8
height 8.0
m1 length 7.9
width 3.0
height 6.7
p3 length 4.5
width 2.1
p2 length (alveolus) 2.5
width (alveolus) 1.7
c1 length 4.7
width 3.5
height 13.4
i2 length 3.3
width 1.8
height 5.1














GBIF Dataset (for parent article) Darwin Core Archive (for parent article) View in SIBiLS Plain XML RDF