Pareumys, Peterson, 1919

Kelly, Thomas S. & Murphey, Paul C., 2016, Mammals from the earliest Uintan (middle Eocene) Turtle Bluff Member, Bridger Formation, southwestern Wyoming, USA, Part 1: Primates and Rodentia, Palaeontologia Electronica 7 (8), pp. 1-55 : 31-36

Genus PAREUMYS Peterson, 1919 cf. Pareumys sp.

Figure 10.1-12 , Table 6

Referred specimens. From UCM Locality 92189: P4, UCM 95760; M1, UCM 95758, 92942; M3, UCM 67987, 70905; m1, UCM 69974. From SDSNH Locality 5841: m2, SDSNH 110365, 110366; m3, SDSNH 110367; partial m3, SDSNH 110368. From DMNH Locality 4672: P4, DMNH 75335, 75337; M1 or 2, DMNH 75334. From SDSNH Locality 5843: M1 or 2, SDSNH 110396, 110397. From SDSNH Locality 5844: partial P4, SDSNH 110435; m3, SDSNH 110434.

Description. Four teeth are identified as P4s because they are lacking a hypocone. Their occlusal outline is oval. The protocone is robust and the largest primary cusp. The paracone is slightly larger than the metacone, and both cusps are slightly anteroposteriorly compressed. The protoloph is a complete, relatively tall loph, extending labially from the protocone to connect with the paracone. On one P4, a small swelling occurs on the protoloph (incipient protoconule), near its connection with the paracone. The metaloph is a relatively tall loph that extends lingually from the metacone to terminate in the central valley, near the protocone. The anterior cingulum is moderately strong, extending labially from the anterior base of the protocone to terminate at the anterior base of the paracone. The posterior cingulum is robust and extends in an arc from the protocone to terminate at the posterior base of the metacone.

Confident separation of the first two upper molars in samples of isolated cylindrodontid teeth cannot be made. The protocone is the largest primary cusp. The hypocone is distinct, and separated from the protocone by an indentation along the lingual wall of the tooth. The paracone is robust and compressed anteroposteriorly. The metacone is smaller than the paracone and somewhat compressed anteroposteriorly. The crown is unilaterally hypsodont with the protocone and hypocone significantly taller than the paracone and metacone. The anterior cingulum is robust, extending from the anterior edge of the protocone to terminate at the anterior edge of the paracone. The protoloph is a complete, tall loph connecting the protocone to the paracone. The metaloph is incomplete, extending labially from the metacone to an indistinct metaconule (usually a slight swelling) to terminate either in the central basin (three teeth) or turn posteriorly to connect with the posterior cingulum (two teeth). The posterior cingulum is relatively tall, extending labially from the posterior edge of the hypocone to terminate at the posterior edge of the metacone. The central basin, and the valleys between anterior cingulum and protoloph and between the metaloph and posterior cingulum, are relatively deep.

Two teeth are identified as M3. They have a subcircular occlusal outline and are a little reduced in size relative to M1-2 (mean M3 ap = 95% of mean M1-2 ap). The protocone and paracone are the largest primary cusps. The metacone is distinct and compressed slightly transversely. The hypocone is weak and separated from the protocone by a notch or indentation along the lingual wall of the tooth. The anterior cingulum is strong, extending labially from the anterior edge of the protocone to join the anterior edge of the paracone. The protoloph is complete, relatively tall, extending labially from the protocone to connect with lingual edge of the paracone. The metaloph is tall, and in one M3 it is incomplete, extending lingually from the metacone to terminate in the central basin near the protocone, whereas in the other M3, it is complete, extending lingually from the metacone to join the posterolabial edge of the protocone. The posterior cingulum is robust, extending labially from the hypocone to terminate at the posterolingual edge of the metacone.

Confident separation of the first two lower molars in samples of isolated cylindrodontid teeth is difficult, but the m1 is usually more elongate anteroposteriorly, whereas the m2 has a more squared occlusal outline with the width greater relative to the length. Only one tooth (UCM 69974) meets the above criteria for m1. In length and all other dental characters, it is nearly identical to the teeth identified as m2, but differs by having a small gap at about the middle of the hypolophid. The sections of the hypolophid that are lingual and labial to this small gap are relatively robust and tall, similar in height to that of the m2-3 hypolophid. This gap is regarded as most likely representing individual variation or an aberrant developmental defect. All of the teeth identified as m2 have a tall, complete hypolophid, extending lingually from the ectolophid, just anterior to its origination from the hypoconid, to join the labial edge of the entoconid. The occlusal morphology of the m3 is very similar to that of the m2, but differs by being more elongated anteroposteriorly and by having the hypoconid and posterior cingulid more robust, forming a more convex posterior occlusal margin. The lower molars exhibit a number of shared characters. In early wear, the talonid is nearly as tall as the trigonid. The anterior cingulid is short, extending lingually from the protoconid to join the labial edge of the metaconid. The metalophulid II (posterior arm of the protoconid) extends from the posterolingual corner of the protoconid to terminate near the posterolingual base of the metaconid. The ectolophid is tall and complete, extending from the protoconid to the hypoconid. The central basin and the valley between the hypolophid and posterior cingulid are deep.

Remarks. The taxonomy of Bridgerian cylindrodontids has a complicated history. Leidy (1871) named Mysops minimus , the type species, based on a partial right dentary with moderately well-worn m2-3 from the lower Bridger Formation at Grizzly Buttes. The diagnostic characters of the species are as follows ( Leidy, 1871; Matthew, 1910; Troxell, 1923b; Wilson, 1938b): 1) an incomplete m2 hypolophid; 2) a quadrate m2 occlusal outline (tra and trp subequal); 3) a complete m3 hypolophid; and 4) a narrow valley between the m2 entoconid and posterior cingulid. Wilson (1938b) referred one additional specimen to the species, which also exhibits an incomplete m2 hypolophid. Wilson (1938b) stated that the referred specimen is "presumably from the Bridger Formation, but its exact locality is unknown."

In 1872, Marsh named three new taxa from the Bridger Formation; Tillomys parvus based on a partial dentary with a well-worn m1, Tillomys senex based on a partial dentary with m2, and Taxymys lucaris based on a partial maxilla with P3-4. Leidy (1873) named Mysops fraternus based on a partial right dentary with m1-3 from an unknown locality within the Bridger Formation. Matthew (1910) first suggested that the holotype of T. parvus may actually be cogeneric with Mysops minimus . However, Troxell (1923b) followed Marsh (1872) and retained T. parvus . Troxell (1923b) questionably considered Mysops fraternus as a junior synonym of T. parvus . Although Troxell (1923b) recognized Mysops minimus as valid, he considered Taxymys lucaris to be cogeneric with Tillomys , which made Taxymys a junior synonym of Tillomys . Troxell (1923b) further recognized two subspecies of Tillomys parvus ; T. parvus parvus , represented by the holotype of T. parvus , and Tillomys parvus plicatus , based on a partial dentary with p4-m3, both of which are from Grizzly Buttes ( Troxell, 1923b; Wilson, 1938c). Troxell (1923b) distinguished T. p. plicatus from T. p. parvus by it smaller size and a more complicated occlusal pattern. In a series of papers revising certain Bridgerian rodents, Wilson (1938a, 1938b, 1938c) provided convincing evidence that Taxymys and Tillomys are representatives of Sciuravidae , whereas T. parvus is actually cogeneric with Mysops minimus as Mysops parvus , and both of the latter taxa belong in Cylindrodontidae . Wilson (1938b) listed the following characters in the diagnosis of M. parvus : "hypolophid of m2 generally strongly developed; [lower molar] entoconids usually not especially robust, more or less crested; posterior valleys of [lower] molars well-developed; m2 more elongate than in M. minimus ." Wilson (1938b) further noted that the occlusal pattern of the type specimen of M. p. plicatus is no more complicated than additional specimens that he referred to M. p. parvus . Wilson (1938b) stated that because the holotype of M. p. parvus is so incomplete, with only m1 present, the characters used by Troxell (1923b) to distinguish it from M. p. plicatus may very well represent individual variation instead and not rise to a subspecific level. Wilson (1938b) noted the following additional differences between the holotype of M. p. parvus and that of M. p. plicatus : 1) the anterior termination of the masseteric fossa slightly farther forward; and 2) the mental foramen relatively more superior and anteriorly positioned. Wilson (1938b) also regarded the differences between M. p. plicatus and M. fraternus as possibly just examples of individual variation, but noted that "because of the uncertain relation of M. p. parvus to M. p. plicatus and M. fraternus ," he tentatively retained all three taxa. Wilson (1938b) acknowledged that the then known sample of Mysops from the Bridger Formation is highly variable in occlusal morphology and stated "there are probably at least two valid species present, one form represented by the holotype of M. minimus and the other best represented by the holotypes of both M. p. plicatus and M. fraternus ." However, M. p. parvus has priority over M. parvus plicatus and M. fraternus , but unfortunately it is represented by the poorest holotype ( Wilson, 1938b). In summary, Wilson (1938b) recognized M. mimimus as represented by the holotype and one referred specimen, M. fraternus as represented by the holotype only, and the subspecies M. p. plicatus by the holotype only, whereas all other specimens were assigned to M. p. parvus . Wilson (1938b) discussed in detail the wide degree of variation seen in the specimens he referred to M. p. parvus , but he was unable to correlate these differences into consistent groupings. His analysis was also hampered by a general lack of stratigraphic information for many of the specimens.

Wood (1973) named a new species of Mysops , M. boskeyi , based on isolated teeth from the early Uintan Whistler Squat Quarry in the lower member of the Devil's Graveyard Formation, Texas, which Campisano et al. (2014) have recently re-dated as between 45.04 - 44.88 Ma. Wood (1973) also provided a more complete summary of the dental measurements of the sample of M. parvus from the Bridger Formation and noted that the high coefficients of variation suggest that more than one species may be represented in the sample. Campisano et al. (2014) state " Wood (1973) described M. boskeyi as a high crowned species of Mysops with clear difference in height between the trigonid and talonid that is not characteristic of later cylindrodontids." Korth (1994) suggested that M. boskeyi should be referred to Pareumys (also see Black and Sutton, 1984), which was followed by Walsh and Storer (2008). However, although Campisano et al. (2014) acknowledged Korth's (1994) proposal they did not mention Walsh and Storer's (2008) referral of the species to Pareumys and chose to retain it in Mysops in the absence of a more formal argument for transferral. Although a reanalysis of the hypodym of M. boskeyi is beyond the scope of this study, judging from the illustrations and descriptions of M. boskeyi ( Wood, 1973) , the molars are distinctly higher crowned than the previously recognized species of Mysops from the Bridger Formation. The ratio of the m1-2 unworn protoconid height to m1-2 ap of Mysops specimens from the Bridger B- D (Br2-3) ranges from 0.63 to 0.65. In the holotype M1 of M. boskeyi , the protocone length (measured from the apex of the protocone to the lingual base of the enamel in anterior view) relative to its greatest transverse width (tra) equals 0.87, whereas that of M1-2 of Mysops from the Bridger B-D ranges from 0.65 to 0.70. Another character seen in M1-2s of M. boskeyi that is similar to those of Pareumys is that the metaloph is commonly directed toward and connected to the posterior cingulum (= posteroloph), whereas in the Mysops specimens from the Bridger B-D, the metaloph is directed towards the protocone and does not connect with the posterior cingulum. Mysops is very similar to Pareumys and most investigators believe that Pareumys was derived from Mysops (e.g., Wilson, 1938b; Wood, 1973), so where one draws the line between the two genera becomes somewhat arbitrary. However, we believe that Korth's (1994) and Walsh and Storer's (2008) assignment of the species to Pareumys is reasonable based on its greater hypsodonty, increased size, and the common occurrence of a connection between the M1-2 metaloph and posterior cingulum.

To summarize, the three species of the cylindrodont Mysops are currently recognized from the Bridger B-D ( M. minimus , M. parvus and M. fraternus ). Mysops parvus is further provisionally divided into two subspecies (M. p. parvus and M. p. plicatus ). The validity of these species and the subspecies is unclear because the holotypes of the type species M. minimus and M. parvus are poorly preserved so adequate comparisons to the other species or subspecies is difficult. Moreover, many specimens of Mysops collected during early explorations of the Bridger Formation lack stratigraphic control, which also makes comparisons difficult. However, the degree of occlusal variation and wide observed ranges in the dental measurements ( Wilson, 1938b; Wood, 1973), strongly suggests that when additional stratigraphically controlled samples become available, it may be possible to better define the actual species diversity of Mysops from the Bridger Formation. Until then, we follow Wilson's (1938b) taxonomic hierarchy. Contrary to Campisano et al. (2014) and following Korth (1994) and Walsh and Storer (2008), we consider the Texas cylindrodont from the early Uintan (Ui1b) Whistler Squat Quarry to represent a primitive member of Pareumys , P. boskeyi .

The cylindrodont teeth from the TBM (Ui1a) are interesting because they appear to be intermediate in morphology between Mysops species from lower in the Bridger Formation (Br2-3) and P. boskeyi . They exhibit significantly higher molar crown height than Bridgerian specimens of Mysops , wherein the observed range of their unworn M1-2 protocone height/tra equals 0.83-0.88, similar to that of the holotype M1 of P. boskeyi . As noted above, a Rm1 (UCM 69974) from the lowest locality in the TBM (UCM Locality 92189) has the hypolophid almost complete, but with a small gap present at about the middle of the hypolophid. Considering the degree of variation of the hypolophid seen in the larger samples of Bridgerian (Br2- 3) Mysops , the significance of this character is debatable, and it likely represents either individual variation or a developmental defect. Even though the protoconid is moderately worn on UCM 69974, it has a protoconid height/ap ratio of 0.64, which is similar to those of unworn m1-2s of Mysops specimens from lower in the section. In unworn m1-2s of Bridgerian Mysops , the protoconid is equal to or only slightly lower than the metaconid height (unworn height of metaconid/protoconid height = 0.98-1.00), but by moderate wear its height is significantly reduced relative to that of the metaconid, so presumably, this ratio would be greater if UCM 69974 was unworn. The m3 (SDSNH 110434) from SDSNH Locality 5843 is abraded, but in early wear, and has a protoconid crown height/ap ratio of 0.83, which is significantly higher than those of unworn m3s of Bridgerian specimens of Mysops (observed range of unworn m1-3 protoconid height/ap = 0.63- 0.65). Of the five M1 or 2s from the TBM, two have the metaloph directed towards and connecting with the posterior cingulum, whereas in the other three, the metaloph is directed towards the hypocone, but well separated from it. The TBM sample of teeth are smaller than those of P. boskeyi , but within the observed size ranges of Mysops species from lower in the Bridger Formation.

Previously, two upper molars (SDSNH 110396 and 110397) from SDSNH Locality 5844 were referred in the SDSNH catalog to Pareumys , whereas all of the isolated teeth from SDSNH Locality 5841, which is 104 m lower in the TBM near its base, were referred to Mysops . In the UCM catalog, a number of isolated lower molars from UCM Locality 92189, which is also near the base of the TBM at 103 m below SDSNH Locality 5844, were referred to Mysops . The occurrences of both genera in the TBM were also listed in Gunnell et al. (2009) and Murphey et al. (2011), but no explanations for their generic assignments were included. These investigators then included the last occurrence of Mysops and the first occurrence of Pareumys as part of their characterization of the earliest Uintan biochron Ui1a. In addition to the above three samples, three more cylindrodontid teeth from SDSNH Locality 6242 are now known. The fact that the small sample of cylindrodontid teeth from the TBM are significantly higher crowned than Mysops specimens from lower in the Bridger Formation (Br 2-3) and they exhibit a tendency of the M1-2 metaloph to connect to the posterior cingulum (40% of the teeth), indicates that they exhibit slightly more similarity to P. boskeyi than to Bridgerian Mysops . To reflect this putative relationship, the TBM specimens are only compared to the genus as cf. Pareumys sp. This phylogenetic scenario implies that the TBM specimens may represent a transitional form between Bridgerian Mysops and early Uintan P. boskeyi , which is not unreasonable, considering that the TBM is older (~2.0 Ma) than the Whistler Squat Quarry.

Figures

FIGURE 10. cf. Pareumys sp. from TBM: 1, RP4, UCM 95760; 2-3, RM1 or 2, SDSNH 110396; 4, RM1 or 2, SDSNH 110397; 5, RM1 or 2, DMNH 75334; 6, LM1 or 2, UCM 95758; 7, LM3, UCM 67987; 8, Rm1, UCM 69974; 9, Lm2, SDSNH 110366; 10, Rm3, SDSNH 110367; 11-12, Rm3, SDSNH 110434. 1-2, 4-11 occlusal views. 3, anterior view. 12, labial view. Scale bar equals 1 mm.

FIGURE 11. Metanoiamys sp. from TBM: 1, LdP4, DMNH 74135; 2, LP4, DMNH 75331; 3, LM1 or 2, DMNH 74136; 4, LM1 or 2, DMNH 75251; 5, RM1 or 2, SDSNH 110442; 6, LM3, UCM 98813; 7, Rdp4, SDSNH 110369; 8, Lp4, DMNH 74143; 9, Lm1 or 2, DMNH 75252; 10, Lm1 or 2, DMNH 75250; 11, Lm1 or 2, DMNH 75254; 12, Lm3, DMNH 74144. All occlusal views. Scale bar equals 1 mm.

FIGURE 12. Elymys? emryi new species from TBM, macrophotographs with corresponding line drawings of specimens: 1 and 9, LM1, SDSNH 110463; 2 and 10, RM1, SDSNH 110465; 3 and 11, RM1 (holotype), SDSNH 110466; 4 and 12, RM2, SDSNH 110444; 5 and 13, LM3, SDSNH 110448; 6 and 14, Rm1, SDSNH 110456; 7 and 15, Lm2, SDSNH 110457; 8 and 16, Lm3, SDSNH 110461. All occlusal views. Scale bar equals 1 mm.