Emended diagnosis: Leptictids with large, distinct hypocones on P5 (vestigial or absent in nonleptictines). Anteroposteriorly more elongated P5 and upper molars with lower paracones, protocones, and metacones (taller cusps in nonleptictines). Upper molars with slightly more labially situated conules, absence of a twinned paraconule, and variably developed ectoflexi, but less deeply infolded than in nonleptictines. Upper molar parastylar spurs poorly developed. Elongated p5, with large, projecting paraconid. Skeletons (known in Leptictis dakotensis and Palaeictops matthewi) with distinct distal fusion of tibia and fibula, head of humerus semilunar with sharply defined medial border, head of femur strongly canted to long axis of shaft, and deep trochanteric fossa of femur.

Included taxa: Blacktops Meehan and Martin, 2010; Leptictis Leidy, 1868; Megaleptictis Meehan and Martin, 2012; and Palaeictops Matthew, 1899.

Distribution: Early Eocene (Wasatchian)-Late Oligocene (Whitneyan), North America.

Remarks: Postcranial evidence offers compelling evidence for the monophyly of Leptictinae. Unfortunately, only two species in this group are represented by skeletal material. For the present, the dental features alone must serve to group all species of the subfamily.

Stypolophus Cope, 1880: 746.

Parictops Granger, 1910: 250–251.

Genotypic species: Palaeictops bicuspis Cope, 1880 (described as Diacodon bicuspis by Matthew, 1918).

Referred species: Palaeictops borealis (Russell, 1965), P. bridgeri (Simpson, 1959), P. matthewi (Novacek, 1977), P. multicuspis (Granger, 1910), P. altimontis (new, this paper), and P. robustus (new, this paper).

Diagnosis: Leptictine differing from other members of this subfamily (i.e., Leptictis and Megaleptictis) in having single sagittal crest (known in Palaeictops bicuspis, P. altimontis, and P. robustus); shallow suprameatal fossa (known in P. altimontis and P. bicuspis); more transversely flared basioccipital that overlaps ventrally the promontorium of the petrosal (known in P. altimontis); shallow groove on the paraoccipital process for the digastric muscle (known in P. altimontis and P. robustus); a paraoccipital process that is less extensive, so that the distance is shorter between stylomastoid foramen and posterior margin of basicranium (known in P. altimontis and P. robustus); and a small postglenoid process (known in P. altimontis and P. robustus). Expanded cochlear fossula (in P. altimontis and P. robustus). Lingually swollen protocones on P5 and M1-M3. Similar to Leptictis but different from Prodiacodon in having more bunodont cusps on posterior premolars and molars. Similar to Megaleptictis in having a small suprameatal foramen; lacking the posterior concavity in the coronoid process; and short talonid on p5. Differs from Prodiacodon in having slightly lower trigonids; well-developed hypocone on P5; shallow ectoflexi; elongate p5 with enlarged paraconid; less transverse M2; less developed parastylar spurs; and the presence of only one paraconule on the upper molars. Differs from Myrmecoboides in having well-separated paraconids and metaconids on p5-m3; less elongate, relatively wider talonids; and more closely spaced premolars. Pes in P. matthewi differs from that in all other leptictids where known in having a distinctly pear-shaped sustentacular facet on the astragalus and a very reduced fibular facet on the calcaneum.

Distribution: Wind River, Bridger, Tepee Trail, Wasatch, and Willwood formations, Wyoming; DeBeque and Huerfano formations, Colorado; Uinta Formation, Utah; and Cypress Hills and Swift Current Creek formations, Saskatchewan, Canada. Lower-Middle Eocene (Wasatchian-Duchesnean NALMAs).

Remarks: The above diagnosis documents the presence of cranial and postcranial traits that may exclude Palaeictops from a Leptictis grouping. It is noteworthy, however, that the posteriorly expanded nasals of P. bicuspis are primitive, but not universal for this genus. In P. altimontis the nasals are posteriorly constricted in a manner similar to that in Leptictis. There is, in fact, clear evidence that Palaeictops is closely tied to the history of the Late Eocene-Oligocene taxa. Postcranial features of P. matthewi also support this close phylogenetic relationship (e.g., distal fusion of tibia and fibula, head of femur strongly canted to long axis of shaft, and deep trochanteric fossa of femur).

Although there are distinct differences between Palaeictops and Prodiacodon, some of the criteria demarcating these taxa in Novacek (1977) have been questioned by Bown and Schankler (1982). The latter authors offered the following comparisons:

Stypolophus bicuspis Cope, 1880: 746.

Ictops bicuspis Cope, 1881: 192; Cope, 1885: pl. 29a, figs. 2, 3. Name combination.

Palaeictops bicuspis Matthew, 1899: 31, 35. First use of current name combination.

Diacodon bicuspis Matthew, 1918: 574–576. Name combination.

Diacodon pineyensis Gazin, 1952: 19. Name combination.

Palaeictops pineyensis Van Valen, 1967: 232. Name combination.

Holotype: AMNH 4802 (fig. 2A-B), consisting of a partial skull that retains C, P4-P5, M1-M3 on the left side; the right side retains two incisors possibly corresponding to I2 and I3 (both are broken), as well as P2 (broken), P4-P5, M1-M2, and M3 (broken). Left ramus with p5-m3 and roots of p1, p2, and p4 (fig. 3). Right ramus with p4-m2 and roots of i?, c, p1, and p2 (fig. 4).

Type locality: Wind River Basin, Wyoming. Wind River Formation, Lower Eocene (Wasatchian NALMA).

Diagnosis (revised from Novacek, 1977: 14): Differs from other species of Palaeictops in having a large, swollen and anteriorly extended paracone of P4 (paracone less anteriorly projecting in P. altimontis and P. matthewi); tricuspid p4 with large anterior cusp on the heel and lack of a cusp anterior to main cusp (p4 in P. multicuspis, P. matthewi, and P. altimontis with cusp anterior to main cusp, but lacking a large anterior cusp on the heel). Differs from P. bridgeri in its smaller size, in lacking a broad prefossid between paraconid and metaconid on p5, and in having a more lingually positioned hypoconulid on m3. P4 and p4 less tall and trenchant than in P. multicuspis (p4) and P. matthewi. P4 more anteriorly extended that in P. altimontis (fig. 5).

Distribution: Wind River Formation (Lost Cabin and Lysite members), Wasatch Formation (Knight Member), Willwood Formation, Wyoming, Lower Eocene (Wasatchian NALMA).

Referred material: AMNH 4255, left ramus with broken m1, m2-m3 and, provisionally, several other specimens from the Willwood Formation, Bighorn Basin, Wyoming, described by Bown and Schankler (1982: 19) (see remarks below). YPM VPPU 13436, jaws with upper and lower cheektooth dentition and several other specimens described by Guthrie (1971: 54–55) from the Lost Cabin Member, Wind River Formation, Wyoming. Provisionally, USNM 19204, left ramus with damaged p5, m1-m3, from the Knight Member, Wasatch Formation, Big Piney La Barge Fauna, Sublette County, Wyoming. Described as the type of Diacodon pineyensis by Gazin, 1952, and referred to Prodiacodon tauricinerei by Novacek (1977: 26). YPM VPPU 13419, upper and lower jaws with cheektooth dentitions and associated distal humerus fragment (described as Palaeictops pineyensis by Guthrie, 1967) from the Lysite Member, Wind River Formation, Wyoming.

Remarks: The most distinctive feature of Palaeictops bicuspis is the enlarged, anteriorly positioned paracone on P4 (fig. 5). This tooth is unknown in P. borealis, P. bridgeri, P. multicuspis, and P. robustus, and the combination of features cited above must be used to separate P. bicuspis from all other members of the genus.

We support Guthrie's (1971) identification of YPM VPPU 13436 from the Lost Cabin Member of the Wind River Formation as Palaeictops bicuspis. The characteristic P4 is clearly present in this specimen.

Bown and Schankler (1982: 16) gave statistical comparisons of tooth measurements for several species of Prodiacodon and Palaeictops. These suggest that the type of Palaeictops (Diacodon) pineyensis (Gazin, 1952) is closer to P. bicuspis than to Prodiacodon tauricinerei, as suggested by Novacek (1977). The type is badly preserved, and assignment is uncertain. At present, we provisionally accept its transfer to P. bicuspis. Also referable to this species are the Palaeictops pineyensis specimens described by Guthrie (1967) from Lysite Member of the Wind River Formation.

Bown and Schankler (1982: 18) identified AMNH 48763, a ramus with p5-m1, as Palaeictops bicuspis. This assignment is contradicted by the morphology of the specimen. The p5 paraconid is lower, less anteriorly positioned, and has a slightly concave posterior face continuous with the lingually opened prefossid. The m1 has a very distinct entoconulid, a feature shared by Prodiacodon and Myrmecoboides. AMNH 48763 is retained within Prodiacodon tauricinerei, as proposed by Novacek (1977: 26).

The nomenclatural history of Palaeictops bicuspis is reviewed in Novacek (1977).

Protictops? borealis Russell, 1965: 7–8, pl. 1, figs. 1, 2.

Palaeictops borealis Storer, 1984: 22–26: figs. 2A-B. First use of current name combination.

Holotype: ROM 1676, right M1 (fig. 6B).

Type locality: Swift Current, Saskatchewan. Swift Current Creek Formation, Middle Eocene (Uintan NALMA).

Diagnosis: Palaeictops borealis can be easily distinguished from other species of the genus by its size (larger than P. altimontis, P. bicuspis, P. bridgeri, P. multicuspis, and P. matthewi, but smaller than P. robustus). Like P. matthewi in having P5 with a more extensive precingulum in lingual region (less extensive precingulum present in P. altimontis and P. bicuspis). Differs from P. altimontis, P. bicuspis, and P. matthewi in having the upper molars with a more extensive precingulum in lingual region. Like P. altimontis and P. bridgeri in having the lower molars with lower trigonids (higher trigonids present in P. bicuspis, P. matthewi, and P. multicuspis).

Distribution: Cypress Hills and Swift Current Creek Formations, Saskatchewan, Canada. Middle Eocene (Uintan-Duchesnean NALMAs).

Referred material: RSM P1654.220, right DP5; RSM P1654.221-222, right P5; RSM P1654.226, right m1; ROM 23595, left m1; RSM P1654.225, right m2; ROM 1685, right m2; RSM P1654.223-224, left m1 or m2; Swift Current Creek Formation, Saskatchewan, Middle Eocene. RSM P1899.1472, right DP4; RSM P1899.1486, left M1 or M2; and P1899.1450, left m1 or m2; Cypress Hills Formation, Saskatchewan, Middle Eocene.

Remarks: Russell (1965) could not assign with certainty the holotype of Palaeictops borealis to a specific locus (P5 or M1), but subsequently Storer (1984) considered the holotype to be an M1, an assignment with which we agree. Likewise, other teeth were not assigned to a specific locus by Russell (1965) and Storer (1984); ROM 1685, ROM 23595, and RSM P1654.223-226 could not be assigned as either an m1 or m2. Based on our revision of all the Palaeictops material we were able to assign four of the six teeth to either m1 or m2 (see Referred Material), but we were not able to do this for RSM P1654.223 and RSM P1654.224 because of their poor preservation. We were not able to review the material from the Cypress Hill Formation (RSM P1899.1472, right DP4; RSM P1899.1486, left M1 or M2; and P1899.1450, left m1 or m2), but from the illustration it is clear that at least one specimen (RSM P1899.1450) belongs to P. borealis (Storer, 1995: fig. 1a).

Diacodon bridgeri Simpson, 1959: 1–4, fig. 1.

Palaeictops bridgeri Van Valen, 1967: 232. First use of current name combination.

Holotype: AMNH 56032, consisting of a right ramus that retains p5-m3 (fig. 7).

Type locality: Locality 6 (Misery Quarry) of McGrew (1959) in the vicinity of Tabernacle Butte, Wyoming. Upper Bridger Formation, Middle Eocene (Bridgerian NALMA).

Diagnosis: Like Palaeictops altimontis (see diagnosis below) and unlike other species of the genus in having a p5 with a low paraconid bordered posteriorly by a distinctly opened, lingual prefossid, a shortened talonid, and a deeply excavated talonid basin opened lingually and bordered posteriorly by steep faces of the entoconid and hypoconid (fig. 7). Both the types of P. bridgeri and P. altimontis have slightly worn, but low trigonids on m1-m3. Differs from P. altimontis in larger size (length of m1-m3 in P. bridgeri equals 11.23 mm, length in P. altimontis equals 9.18 mm.), and in having a p5 metaconid positioned more posteriorly relative to the protoconid and a more anteroposteriorly “opened” trigonid due to a more salient paraconid on m2. M2 possibly referable to P. bridgeri is more transverse with a shallower ectoflexus than M2 of P. altimontis.

Distribution: Bridger Formation, Bridger Basin, western Wyoming, Middle Eocene (Bridgerian NALMA).

Referred material: AMNH 9873, left M2, Main locality, Tabernacle Butte, Sublette Co., Wyoming. Upper Bridger Formation.

Remarks: Palaeictops bridgeri was clearly described and illustrated by Simpson (1959) who first noted the distinctive characters of its lower, last premolar. The new Palaeictops altimontis species from the Tepee Trail locality (see below) is most similar to P. bridgeri, but the two species differ in size and in details of the lower dentition. Moreover, the possibility that a leptictid M2 (AMNH 98731) from Tabernacle Butte is referable to P. bridgeri provides further evidence for the distinction of this species from P. altimontis.

Holotype: FMNH P26904 (fig. 8A-B), consisting of a damaged skull that retains P2-P5 (broken), M1-M2 (roots), and M3 (broken) on the left side; the right side retains roots of C-P1, P2 (broken), P4-P5, and M1-M3. Right ramus with c-p1 (broken), p2-p5, m1-m2 (broken), and root of i2 (fig. 9). Left ramus with c (broken), p1, p2 (broken), p4-p5, m1-m3 (broken), and root of i2 (fig. 10). Partial skeleton including lumbar and caudal vertebrae, pelvis, femur, tibia, carpals, tarsals, and phalanges.

Type locality: Exposure near top of Buzzard Pass, Mesa Co., Colorado. De Beque Formation, Rifle Member, Lower Eocene (Wasatchian NALMA).

Diagnosis (modified from Novacek, 1977): Similar to Palaeictops multicuspis in having tall, trenchant p4 but differs from P. multicuspis in slightly smaller size and in having a more slender ramus; p2 with only one main, trenchant cusp and a basal posterior cuspule; p4 with anterior main cusp larger than posterior main cusp and a short transverse ridge on the posterior heel separated from the rest of the crown by a transverse trough (figs. 9, 10). P4 trenchant tooth, taller than P5 with a very prominent paracone, smaller metacone and low protocone (fig. 8B). Species has postcranial features used to diagnose Palaeictops (see above diagnosis).

Distribution: De Beque and Huerfano formations, Colorado, Lower Eocene (Wasatchian NALMA).

Referred specimens: AMNH 17555, skull fragment, partial lower jaws and partial skeleton. Garcia Canon region, Colorado. Lower beds of Huerfano Formation.

Remarks: Bown and Schankler (1982: 11–12) stated that Palaeictops matthewi is probably conspecific with P. multicuspis, because differences in size and p2 morphology are trivial. Synonymy is, however, unwarranted. The jaw proportions and p2 and p4 in these species are clearly distinct as described in the diagnosis of each species (this paper). Bown and Schankler (1982) maintained that such characters might be insignificant and subject to intraspecific variation. There is, however, no sample evidence supporting this assessment and, in absence of data to the contrary, P. matthewi is here recognized as diagnosed above.

Parictops multicuspis Granger, 1910: 250–251.

Palaeictops multicuspis Van Valen, 1967: 232. First use of current name combination.

Holotype and only specimen: AMNH 14741, a left ramus with p2-m3 and a right ramus with p4-m3 and alveoli for a double-rooted p2, single rooted canine, and three incisors (figs. 11, 12).

Type locality: Alkali Creek (Buck Springs) of Wind River Basin, Wyoming. Lost Cabin Member, Wind River Formation, Lower Eocene (Wasatchian NALMA).

Diagnosis: Differs from other species of the genus (except P. robustus) in its large size and deeper ramus. Also differs from other species (including P. robustus) in having a trenchant p2 with four anteroposteriorly aligned cusps; p4 with very small anterior accessory cusps, two main cusps and two low cusps on the shortened heel; p5 with small, but distinct, cuspule at labial base of the paraconid. Similar to Palaeictops matthewi in having a trenchant p4 taller than p5 but differs from P. matthewi in having greater number of cusps on p2 and p4. Similar to Prodiacodon and Myrmecoboides in having lower molar entoconulids, but differs from these taxa in showing characters that diagnose Palaeictops (e.g., large, swollen p5 paraconid).

Remarks: See Novacek (1977).

Palaeictops sp. Novacek, 1977: 21.

Holotype: AMNH 96250 (figs. 5B, 13B), consisting of a nearly complete skull that retains P1-P2, DP4-DP5, M1-M3 on the left side; the right side retains P5-M3. Left ramus with p2, and p4-m3 (fig. 15A-B). Right ramus with p4, p5 (damaged), and m1-3 (fig. 15B).

Referred specimens: AMNH 88400, right ramus with p4-5; AMNH 101955, left ramus with P2, P5; AMNH 105032, left ramus with p4, m3; AMNH 99301, left maxilla with P4-M3; AMNH 113880, left ramus with p4-m3. All from the type locality.

Type horizon and locality: Unit 24 (bone bed A) East Fork Basin, northeast of Dubois, Freemont Co., Wyoming, about 500 feet above the local base of the Tepee Trail Formation (see McKenna, 1980: 337), Middle Eocene (Uintan NALMA).

Etymology: From the Latin: altus, “high,” and mons, “mountain.” Refers to the dramatic montane settings of the type locality (see Love, 1939; McKenna, 1980).

Diagnosis: Like Palaeictops bridgeri in having p5 with open prefossid and shortened talonid, but differs in smaller size, in lacking an anterior accessory cuspule at labial base of paraconid; p5 metaconid aligned less obliquely relatively to protoconid; more anteroposteriorly compressed trigonid on m2; and less transverse M2 with a deeper ectoflexus. Differs from P. bicuspis, P. multicuspis, and P. matthewi in having smaller, more anteriorly separated paraconid on p5 and lower trigonids on m1-3. Skull like P. bicuspis but unlike Leptictis in having single sagittal crest (fig. 13A). Skull like Leptictis but unlike P. bicuspis in having posteriorly narrow nasal elements. Basicranium differs from Leptictis (cf. fig. 14 and Novacek, 1986: fig. 22) in having: (1) a postglenoid foramen positioned more laterally and farther from the anteroexternal edge of the tympanic cavity; (2) a more transversely flared basioccipital that overlaps ventrally the promontorium of the petrosal; (3) shallower grooves on the promontorium for the promontory and stapedial branches of the internal carotid artery; (4) a posterior lacerate foramen only slightly larger than the stapedius fossa (much larger and more oval than stapedius fossa in Leptictis); (5) a cochlear fossula with a more expanded, dorsal rim; (6) a narrower bridge of the mastoid tubercle of the petrosal extending from ventral rim of the cochlear fossula between the stylomastoid foramen and stapedius fossa; (7) a paraoccipital process that is less extensive, so that the distance is shorter between stylomastoid foramen and posterior margin of basicranium; and (8) a shallower groove on paraoccipital process for the digastric muscle.

Description and Comparisons: The front of the skull of Palaeictops altimontis (AMNH 96250) is badly damaged (fig. 13A-B), and the premaxillae are not preserved, except for a small nasal process on the right side of the skull. Most of the paired nasal elements can be seen in dorsal view (fig. 13A). Posteriorly the nasals are narrow, as in Leptictis. However, unlike the latter, the contact of the nasal and the frontal is along a more obliquely oriented suture (fig. 13A). It is difficult to ascribe much taxonomic significance to this difference, as the form of this suture varies to some degree in Leptictis. On the right side of the skull, the maxilla has been displaced laterally, and the contact of this element with the nasal is marked by a faint ridge that divides the dorsal (horizontal) process of the nasal from a more vertically oriented process (fig. 13A). From this, it is clear that the maxilla overlaps the nasal for a considerable extent on the skull roof and nasalfacial exposure (fig. 13A). Differential growth of the maxilla and nasal may thus account for the relatively narrowed nasal exposure in Palaeictops and Leptictis.

The maxilla has a broad contact with the frontal along an oblique suture (fig. 13A). The condition resembles that in Leptictis. In Palaeictops bicuspis this contact is narrower, because the posterior nasals are much broader in their contact with the frontals (fig. 2A). The antorbital fossa is distinct, though shallower than in Leptictis. The ventral border of the fossa is marked by a pronounced ridge. The anterior foramen of the infraorbital canal (preserved only on the right side of the skull) is a circular, but somewhat smaller, opening than in the Leptictis (fig. 13A-B). As in the latter, the infraorbital canal is relatively short in length; it opens above M1 (fig. 13A-B). Unfortunately, the important relationships of the maxilla with other elements of the orbital region cannot be seen due to poor preservation.

The palatine has a basically similar construction as that in Leptictis. Its posterior margin (coincident with the posterior margin of the palate) lies between the last molars (fig. 13B). The margin is biconcave with a distinct, rounded postpalatine torus (fig. 13A-B). The minor palatine foramen is a large opening in the pars perpendicularis that meets the pterygoid (fig. 13A-B). The dorsally trending route of the minor palatine foramen cannot be traced as orbital foramina are obscured by damage. It appears, however, that the dorsal exit of the minor palatine foramen and the sphenopalatine foramen were in closed proximity within a shallow depression, a condition also seen in Leptictis.

The lacrimal is a small triangular element vaguely demarcated on both sides of the skull (fig. 13A-B). As in Leptictis, the lacrimal foramen is confined fully within the orbit. It faces posteriorly and lies directly below the dorsal ridge of the lacrimal (fig. 13A-B). A lacrimal tubercle is present, though it appears much weaker than this structure in Leptictis.

The jugal is not preserved, but a broken, rugose surface on the anterior root of the zygoma suggests that this element contacted the lacrimal (a common and probably primitive eutherian trait seen in other leptictids).

The frontal is poorly demarcated in the orbit, so it is impossible to determine whether it was isolated from the orbital process of the maxilla by the intrusive palatine. Its separation from the maxilla, however, seems likely, because the palatine-maxillary boundary lies directly above the pars perpendicularis of the palatine. This relationship is like that in Leptictis. Accordingly, it's probable that Palaeictops shared with the latter the same orbital mosaic. A small ethmoidal foramen is preserved near the frontal-orbitosphenoid suture in the right side of the skull (fig. 13A-B).

The most notable feature of the skull roof is the single sagittal crest, which is best exposed in the posterior parietal near the junction of the sagittal and lambdoidal crests (fig. 13A). More anteriorly, bone is missing and the brain endocast underneath shows the medial longitudinal fissure (fig. 13A). In addition to having only a single median sagittal crest, this skull differs from Leptictis in showing much less sculpturing for attachment of the temporalis muscle on the parietal and squamosal. This, in combination with the comparatively smaller postglenoid process, suggests a weaker development of the temporalis complex of the jaw-closing apparatus. As noted in Novacek (1986), Leptictis shows a strong emphasis of orthal shear.

Another interesting feature of the skull roof in Leptictis is the extension of the parietal around the lambdoidal crest and its exposure as a small triangular process on the occiput. On the left side of the skull of Palaeictops altimontis, there is a break that may indicate the boundary between the parietal and interparietal, suggesting that the parietal in this form also developed an occipital process. Unfortunately, because of poor preservation, this rather anomalous mammalian condition is not clearly identified in Palaeictops.

Features of the orbitosphenoid and alisphenoid in the orbit are not clearly preserved (fig. 13B). The presphenoid appears to have a ventral median keel as in Leptictis. On the left side of the skull, there appears to be a short, but badly damaged alisphenoid canal just anterior to the foramen ovale.

The pterygoid shows prominent, vertical, entopterygoid crests that form the medial walls of the ectopterygoid fossae (fig. 13A-B). Although this region is damaged, there is evidence for the presence of lateral ectopterygoid crests of the alisphenoid, indicated by the extensive development of the fossae and the flaring of a ridge on the right side of the skull that lies lateral to the entopterygoid crests (fig. 13A-B). As noted in Novacek (1986: 45) it is probable that the manner in which the internal pterygoid muscles originated from the skull is quite similar in Palaeictops and Leptictis.

The basisphenoid is closely fused with adjacent elements. In ventral view it forms a trapezoidal platform with weak sculpturing for pharyngeal grooves and rectus capitis muscles (fig. 13A-B).

The squamosal is strongly distorted by the dorsoventral flattening of the skull (fig. 13A). However, a few features of interest are recognizable. As noted above, there is no marked rugosity or ornamentation of the lateral moiety of the squamosal in the temporal region. Unlike Leptictis the suprameatal fossa behind the zygomatic process of the squamosal is quite shallow. A small opening lies directly above this fossa on the left side of the skull. This is likely the suprameatal foramen. In Leptictis, the suprameatal foramen is much larger and is situated more ventrally, well within the deep suprameatal fossa. The glenoid fossa for articulation with the lower jaw is a shallow, broad surface, somewhat more extensive and less concave than in Leptictis (fig. 14). Although the skull on both sides is damaged in the region of the postglenoid process, it is clear that this feature is smaller than in Leptictis. There is only a very narrow bridge of bone representing the root of the postglenoid process. Furthermore, the postglenoid foramen lies posterior and slightly lateral to this area (fig. 14). In Leptictis the foramen has shifted to a more medial position, presumably repositioned by the marked expansion of the swollen postglenoid process.

The tympanic region of AMNH 96250 is remarkably well preserved and, thus, shows much more detail than other regions of the skull. There are some obvious departures from the condition in Leptictis. The promontorium of the petrosal is more smoothly rounded and does not taper as distinctively in its anterior region (cf. fig. 14 and Novacek, 1986: fig. 22). The regularity in surface features of the promontorium is the result of two other characteristics. The sulci for the promontory (= internal carotid) and stapedial arteries are very narrow and faint, whereas in Leptictis they form distinctly deep troughs (cf. fig. 14 and Novacek, 1986: fig. 22). In addition, the petrosal crest that arcs on the medial surface of the promontory is only a weak flange, whereas in Leptictis it develops as a distinct and prominent ridge with a rugose contact surface with the entotympanic bulla. Despite the weaker petrosal crest in Palaeictops altimontis, it is likely that this form also had an entotympanic bulla. Although this structure is not present in the specimen, it is also commonly not preserved in skulls of Leptictis, so it seems to be a feature readily lost during burial. Another feature of the promontorium peculiar to P. altimontis is the very broad development of the dorsal lip of the cochlear fossula (fig. 14), which overhangs the fenestra rotunda. In Leptictis this feature is also distinct but is less prominent.

There is, as in Leptictis, a distinct glaserian fissure in the lateral tympanic roof of AMNH 96250. Lateral to the region of the excavated facial canal, there appears to be an epitympanic recess, although its size is uncertain, because the meatal bridge of the squamosal has collapsed (fig. 14). This bridge, the suprameatal surface of the squamosal, is much narrower than in Leptictis (cf. fig. 14 and Novacek, 1986: figs. 20, 21). Moreover, the postympanic process, which forms the posterior buttress for the roof of the meatus, is weaker in Palaeictops. Behind this buttress, the paraoccipital process of the petrosal is somewhat broader in exposure, and it lacks the distinctive groove for the digastric muscle seen in Leptictis.

Posterior to the facial canal is a very well-defined stylomastoid foramen. This opening seems even somewhat larger than in Leptictis, though the difference here is of dubious significance. In the left tympanic region, there is a small process that arises from the mastoid just medial to the postympanic process and extends below the facial canal just short of contact with the promontorium in the region of the fenestra vestibuli. This process is likely that of the mastoid tubercle, which may also represent the fusion of the tympanohyal with the petromastoid (fig. 14). The cup-shaped ventral depression on this tubercle characteristic of Leptictis is not present, although the tubercle is so badly damaged that presence of this fossa in Palaeictops cannot be ruled out. The posterior lacerate foramen is, as in Leptictis, merged with the jugular foramen, so there is only one exit for the internal jugular vein and cranial nerves IX, X, and XI. Because the ventral lip of the cochlear fossula is so expanded, there is no narrow trough between the stapedius fossa and the posterior lacerate foramen (fig. 14). This latter feature is distinctly present in Leptictis.

The occiput in Palaeictops altimontis is badly compressed and damaged, but it is apparent that there was a prominent mastoid exposure in this region. The anterior edge of the ventral occipital condyles has the sigmoid curvature seen in Leptictis and lipotyphlans.

The mandible is deepest below m1 (fig. 15). There is a small mental foramen located below p2 and below the posterior root of p4. At the back of the jaw the masseteric fossa is well excavated. The coronoid process shows some tapering dorsally (fig. 15), but does not show the extreme posterior concavity seen in Leptictis (figs. 15 and Novacek, 1986: fig. 1). Because the extremity of the coronoid process is missing, it is uncertain whether the process had a hooklike outline as in Prodiacodon tauricinerei or a more bluntly round curved process as in Palaeictops robustus (fig. 17). The angular process is damaged in both right and left mandibles. The articular surface of the jaw condyle is slightly broader medially than laterally. The surface features of the condyle are poorly preserved.

There are no upper incisors or canines preserved in AMNH 96250 (fig. 13). The canine alveolus is present on the right side of the skull. P1 is bicuspid, single rooted, and triangularly shaped in lateral outline. P2 is tricuspid, double rooted, and triangularly shaped in lateral outline. The central cusp of P2 is dominant while the most anterior cusp is very minute (fig. 5B).

AMNH 96250 shows a remarkable condition, wherein the right DP4-5 are present and only moderately worn, and the left P5 is already erupted (left P4 is missing) (figs. 5B, 13B). DP4 is roughly triangular in occlusal view with well-developed, somewhat inflated paracone, metacone, and protocone (fig. 5B). There is also a crenulated postcingulum. DP5 is molariform with well-developed paraconule, metaconule, and postcingulum in addition to three inflated main cusps (fig. 5B). There is a pronounced parastylar spur that is overlapped ventrally by the metastylar spur of DP4 (fig. 5B).

A P4 is preserved in AMNH 99301, but it is badly worn. The tooth is roughly triangular in occlusal view, although its parastylar spur is very strong, as is typical of leptictids. P5 is fully molariform with distinct conules. The hypocone is also well developed to nearly a third the height of the protocone. P5 has a strong parastylar spur, but the labial margin of the tooth shows no appreciable invagination.

The upper molars are essentially like P5 although they have broader crowns, more inflated cusps and conules, and (except for M3) relatively larger hypocones and talon basins (fig. 5B). The general construction of the dentition is very like that in other species of Palaeictops and in Leptictis. The labial margins of the upper molars have shallow inflexions in contrast to the condition in Prodiacodon. There is no evidence of a doubling of the paraconule as in the upper molars of Prodiacodon (fig. 5B; cf. Novacek, 1986: fig. 4B).

Lower incisors, canine, and p1 are missing from all specimens (fig. 15). The p2 is elongate, trenchant, and two rooted, with a small anterior cuspule, a large central cusp, and a low heel (fig. 15). The p4 resembles p2 but has four cusps: a small anterior cuspule, a large “central” cusp, a small cuspule on the posterior ridge of the central cusp, and a low cuspid heel (fig. 15). The p5 is molariform with a moderately developed paraconid well separated from the metaconid and protoconid and has an elongated talonid basin bordered by three cusps. The crista oblique contacts the posterior wall of the trigonid nearly directly below the posterior trigonid notch. As noted above the p5 resembles Palaeictops bicuspis in having a very open trigonid due to the salient paraconid. The metaconid is less oblique in its alignment with the protoconid than in P. bicuspis. The lower molars are very characteristic of Palaeictops and Leptictis as a whole (see Novacek, 1986: 17), although the trigonid of m2 is notably compressed due to the crestiform paraconid.

Remarks: The upper cheek teeth of Palaeictops altimontis resemble more closely the corresponding teeth of Leptictis than any other Paleocene or Eocene leptictid. Palaeictops altimontis lacks, however, the distinctive trenchant p4 of Leptictis, a condition more closely approached in P. matthewi. In addition, P. altimontis retains the single sagittal crest and several basicranial characters that are clearly modified in Leptictis. Aspects of molar morphology are specializations that separate all known species of Palaeictops from Leptictis.

Palaeictops bridgeri and P. altimontis are thus far known only from Middle Eocene assemblages. They depart from Early Eocene Palaeictops species in the structure of the p5 paraconid.

Holotype: CM 11954, consisting of a damaged skull that retains P1-P2, P4-M3 on the left side; the right side retains one incisor possibly corresponding to I2, as well as P4-M1 (fig. 16A-B). Left ramus with i1-3, c, p1, p2, and edentulous right ramus (figs. 17, 18).

Type horizon and locality: Leland Bench Wash, 4 mi. west of Ouray, Utah; Wagonhound Member, Uinta B horizon, Middle Eocene (Uintan NALMA).

Etymology: From Latin robustus, “hard” or “strong.” Refers to the notably large size of the skull.

Diagnosis: Teeth (badly worn) less transverse than in Prodiacodon or Myrmecoboides. Single sagittal crest unlike Leptictis but like Palaeictops altimontis and P. bicuspis (figs. 2A, 13A, 16A; Novacek, 1986: fig. 2). Significantly larger than all other North American leptictid species (see table 1; cf. Novacek, 1977: table 2).

Description and Comparisons: The skull of Palaeictops robustus, despite its poor condition, reveals several features of taxonomic interest. It possesses traits that are shared with Palaeictops altimontis but differ in Leptictis. There is only a single sagittal crest (fig. 16A). The postglenoid process (though damaged) is weak and confined to the lateral glenoid region. The postglenoid foramen is posterior (rather than posteromedial) to the postglenoid process (fig. 16A). The promontorium is broad and rounded, with a very weak median petrosal crest. There is a broadly expanded dorsal lip of the cochlear fossula.

The right side of the skull shows the base of a large canine. Anterior to this tooth, there is a damaged alveolus and, more anteriorly, the base of an incisor. Beyond the incisor, the premaxilla protrudes to form the floor to the external nares. There is no clear evidence of additional incisor alveoli, so it is likely that the tooth in question is I2, and as in other leptictids, there were only two upper incisors. The dorsal rim of the coronoid process of the mandible is rounded, rather than the hooklike dorsal coronoid process of Leptictis and Prodiacodon tauricinerei (figs. 17, 18).

Remarks: The teeth in this specimen are so badly worn that reference to Leptictidae is open to question. However, the last upper premolar is of dimensions that suggest its molariform construction (fig. 16A-B). Otherwise, Palaeictops robustus might just as well be a pantolestid. This species marks the youngest occurrence of Palaeictops.

Remarks: Two designated species from Middle Eocene beds are clearly leptictids, but their validity as separate taxa is dubious. The species are Hypictops syntaphus described by Gazin (1949) from a maxilla with a right P1-M3 (USNM 13445) found north of Lone Tree in the Bridger Basin of Wyoming, and Viverravus? nitidus Marsh (1872), known only from a left dp5 (YPM VP 11888) from Henry's Fork, also in the Bridger Basin.

The teeth of Hypictops syntaphus are badly worn (Gazin, 1949: fig. 1), but they are comparable in size to upper teeth that would occlude with the lower dentition of the holotype of Palaeictops bridgeri (contra Simpson, 1959). The M2 of “Hypictops” is also similar to the M2 from Tabernacle Butte referred here to P. bridgeri.

Matthew (1909: 342) remarked that YPM VP 11888, the type of Viverravus? nitidus Marsh, 1872, was probably a deciduous last lower premolar of a leptictid. McKenna et al. (1962) suggested that this tooth could be a permanent p4 (equals p5 of this paper) of Hypictops syntaphus. Either assessment might be correct. The posterior position of the metaconid in YPM VP 11888 is very similar to the condition of the p5 in Palaeictops bridgeri. Some differences are apparent, however, and perhaps this is related to differences found between deciduous and permanent teeth of the same taxon.

Thus, there is a strong possibility that Viverravus? nitidus, Hypictops syntaphus, and Palaeictops bridgeri are one and the same taxon. Synonymy would unfortunately establish YPM VP 11888, a single enigmatic tooth, as the type for Palaeictops nitidus, and the only specimen referable to Hypictops is too badly worn to allow positive identification. It seems best to resist the action of synonymy in these cases.