Herpetotherium, Cope, 1873 a

Murphey, Paul C., 2019, PE Note: Corrigenda to Murphey et al. 2018, Palaeontologia Electronica (25 A) 21 (2), pp. 1-54 : 13-21

Genus HERPETOTHERIUM Cope, 1873a Type species. Herpetotherium fuzax Cope, 1873a by original designation, emended to Herpetotherium fugax Cope 1873c . Other included species. H. comstocki (Cope, 1884) ; H. valens (Lambe, 1908) ; H. merriami

(Stock and Furlong, 1922); H. marsupium Troxell ,

1923a; H. youngi (McGrew, 1937) ; H. edwardi

(Gazin, 1952); and H. knighti (McGrew, 1959) .

Herpetotherium knighti (McGrew, 1959) in McGrew et al. (1959)

Figures 11.1-9 , 12.1-10 , Table 3

1959 Peratherium knighti ; McGrew, in McGrew et al., p. 147, figure 3.

1962 Peratherium morrisi ; Gazin, p. 21, pl. 1, figure 1.

1970 Peratherium sp. ; McGrew and Sullivan, p.74.

1973 Peratherium knighti ; West and Dawson, p. 35.

1975 Peratherium cf. P. knighti ; Setoguchi, p. 267, figures 3-6.

1976 Peratherium knighti ; Gazin, p. 2, 7.

1976 Peratherium sp. , cf. P. knighti , in part; Lillegraven, p. 86, pl. 1, figures 1a-c, pl. 2, figures 1a-c, 2a-c, 3a-c, pl. 3, figures 1a-c, 2ac, 3a-c, pl. 4, figures 1a-c, 2a-c, 3a-c, 4a-c, pl. 5, figures 1a-c.

1982 Peratherium knighti ; Bown, p. A43.

1983b Peratherium knighti ; Krishtalka and Stucky, p. 235, figure 2.

1984 Peratherium knighti ; Storer, p. 17.

1985 Peratherium sp. , cf. P. knighti ; Eaton, p.347, figure 3a.

1996 Herpetotherium sp. , cf. H. knighti ; Rotheker and Storer, p. 772.

1996 Herpetotherium knighti ; Storer, p. 245, 247.

1996 Herpetotherium knighti ; Stucky et al., p. 44.

1998 Peratherium knighti ; Gunnell, p. 88.

2008 Herpetotherium knighti ; Korth, p. 42.

Referred specimens. From locality SDSNH 5841:

Lm2 or 3, SDSNH 110339. From locality DMNH

4672: LM1, DMNH 75286; Ldp3, DMNH 75324.

From locality SDSNH 5844: LM1, SDSNH 110429;

partial LM1 or 2, SDSNH 110430; partial RM 1 or 2,

SDSNH 110431; partial Rm2 or 3, SDSNH 110427. From locality UCM 92189: LM1, UCM 68793; RM 2s, UCM 70636, 70639, 95783; partial RM 2 or 3, UCM 68622; RM 3s, UCM 68794, 70770; partial RM 3, UCM 69973; RM 4, UCM 95784; partial dentary with Lp2, UCM 95797; Rdp3, UCM 68902; Lp3, UCM 68908; Rp3, 68937; Rm1, UCM 95792; Lm2 or 3s, UCM 70977, 95790; Rm2 or 3s, UCM 78454, 95785; partial Rm2 or 3, UCM 68901; partial dentary with Rm2-3, UCM 68574; partial dentary with Lm3-4, UCM 95780; Rm4s, UCM 67896, 95793.

Description. In the TBM sample, all the upper molars are isolated teeth. In Paleogene and early Neogene herpetotheriids, the M1 can usually be distinguished from M2 by having a larger ap/tr ratio, that is the ap dimension is usually greater than the tr dimension, whereas in M2 the ap/tr ratio is smaller (more transverse) with the ap dimension usually smaller than the tr dimension (Korth, 1994; Kihm and Schumaker, 2015). The M3 is more easily distinguished from the first two upper molars by having a more V-shaped occlusal outline and its ap/tr ratio is usually smaller than that of M2 (Korth, 1994; Kihm and Schumaker, 2015). Thus, the ap/tr ratio progressively decreases from M1 to M3. The above criteria were used in this study to identify the tooth positions of the first three upper molars, but it should be noted that without intact dentitions our identifications for M1 and M2 are provisional. In the TBM sample, two M1s ( Figure 11.1-11.2 ) are identified with ap/tr ratios of 1.03 and 1.11, whereas the four M2s ( Figure 11.3-11.6 ) are identified with a mean ap/tr ratio of 0.92 (range = 0.88-0.96).

Besides the M2s being more transverse relative to the M1s, their occlusal morphologies are very similar. They have a dilambdodont centrocrista and a very shallow ectoflexus. The paracone is smaller and lower in height than the metacone. The protocone is robust and anteriorly positioned. The preprotocrista and postprotocrista extend labially from the protocone to terminate at a well-developed protoconule and metaconule, respectively. All stylar cusps are positioned along the labial edge of the stylar shelf. Stylar cusp A is absent in five specimens, only represented by an anterolabial spur that is connected to the labial terminus of the anterior cingulum, and on one specimen it is vestigial (a very small, low cusp on the spur). Stylar cusp B is robust, connected lingually to the preparacrista and positioned opposite of the paracone apex. Stylar cusp C is moderately developed and positioned just posterior of the ectoflexus. Stylar cusp D is moderately developed, slightly smaller than C in

five specimens and about equal in size to C in one specimen and positioned very close to C in four specimens and joined to C in two specimens. A stylar cusp E is lacking on all specimens.

The occlusal morphology of the M3 ( Figure 11.7-11.8 ) is quite similar to that of the M1-2 ( Figure 11.1-11.6 ), but minor differences can be discerned. As noted above, the M3 has a more Vshaped occlusal outline and is slightly more transversely broad than M1-2, which is due to a relatively shorter metastylar wing on the stylar shelf. The ectoflexus of M3 is slightly deeper than that of M1-2, but still quite shallow. The stylar cusp C is positioned slightly more anteriorly, close to the center of the ectoflexus. Stylar cusps C and D are slightly better developed and slightly more separated.

The M4 ( Figure 11.9 ) has a very transversely elongated occlusal outline. The paracone is about equal in size to the metacone. The protocone is robust with its apex nearly in line with that of the paracone. The paraconule and metaconule are small and poorly developed. A metastylar shelf is lacking posteriorly, whereas the stylar shelf extends anterolabially from near the labial base of the metacone to join the anterolabial terminus of the elongated preprotocrista. The anterior cingulum is distinct, extending from the protoconule to the anterolabial projection of the stylar shelf. A single stylar cusp is present that is positioned just slightly posterior of the paracone apex as a flattened, oval bulge on labial edge of the stylar shelf, and probably represents cusp C. The ectoflexus is nearly straight.

The p2 and p3 have a simple morphology, typical of those of Herpetotherium (Lillegraven, 1976; Korth, 1994), with a relatively tall protoconid and weakly-developed anterior and posterior cuspulids that are nearly in line with the protoconid apex. The p2 differs from the p3 by being smaller, narrower, and more elongated anteroposteriorly with the protoconid apex positioned more anteriorly, resulting in a longer, more steeply inclined central ridge from the protoconid apex to the posterior cuspulid and by having a slightly better developed posterior cuspulid.

Two lower teeth agree well in proportions and overall structure to those previously identified as a deciduous p3 of Herpetotherium (Lillegraven, 1976; Rothecker and Storer, 1996; Kihm and Schumaker, 2015). They exhibit an occlusal morphology that is similar to that of the lower molars, including a well-developed entoconid that is taller, larger, and separated from the shelf-like, posteriorly projecting hypoconulid by a distinct notch ( Figure 12.1 ). It differs from the lower molars by the following: 1) smaller size; 2) a lower, relatively smaller and slightly more anteriorly projecting paraconid, resulting in the trigonid being longer than wide; 3) the protoconid larger than the metaconid with their apices positioned relatively closer to each other; and 4) the trigonid and talonid narrower relative to the length.

One tooth is identified as an m1 ( Figure 12.3 ). It differs from m2-3 ( Figure 12.2, 12.4-12.8 ) by being slightly smaller in size and by having the paraconid projecting slightly more anteriorly and positioned slightly more labially. Otherwise, the m1-3 are very similar in occlusal morphology. The protoconid is the tallest primary cusp and is about equal in size to the metaconid. The paraconid is anteroposteriorly compressed, projects anterolabially, and is lower in height than the protoconid and metaconid. The width of the talonid is usually equal to or slightly wider than the trigonid. The hypoconid is the largest and tallest talonid cusp. The entoconid is robust, conical, and taller than the hypoconulid, and separated from it by a distinct notch. The hypoconulid is shelf-like, projecting well posterior of the entoconid, and is connected to the hypoconid by a distinct, relatively tall postcristid. The cristid obliqua extends anterolabially from the hypoconid apex to terminate on the posterior wall of the talonid, about halfway between the protolophid notch and protoconid apex. The anterior and posterior cingulids are moderately robust.

The m4 ( Figure 11.9-11.10 ) only differs from the m1-3 by having a considerably narrower talonid that results in it appearing slightly more anteroposteriorly elongated and a slightly more labially positioned hypoconulid.

Remarks. The familial and subfamilial status of Herpetotherium and other closely related North American and Eurasian genera vary among investigators, with some allocating them to the subfamily Herpetotheriinae Trouessart, 1879 , within the Didelphidae Gray, 1921 (e.g., Korth, 1994, 2008; Hayes, 2005), whereas others have elevated the subfamily to familial rank as Herpetotheriidae (e.g., Case et al., 2005; Kelly, 2010; Ladevèze et al., 2012; Williamson et al., 2012). Regardless of which rank is preferred, most investigators agree on their generic composition. Here, we follow Willamson et al. (2012) and allocate Herpetotherium to Herpetotheriidae .

The Paleogene and early Neogene herpetotheriid marsupials of North America have a rather complicated taxonomic history. Krishtalka and Stucky (1983b) and Korth (1994, 2008) documented detailed historic accounts of their taxonomy, from which a brief updated summary is provided here.

Cope (1873a) described a new genus and species, Herpetotherium fuzax , from what is now known as the late Eocene and early Oligocene White River Formation of Colorado, which he regarded as an insectivore. Shortly afterwards, Cope (1873c) corrected the specific name for the genotype to Herpetotherium fugax and described five additional species of Herpetotherium , which he also regarded as insectivores. Eleven years later, Cope (1884) recognized that Herpetotherium fugax actually represented a marsupial and transferred all six of his species to the European genus Peratherium Aymard, 1850 . Most subsequent investigators followed this generic allocation and referred many Paleogene and early Neogene North America marsupials to Peratherium (e.g., Matthew, 1903; Stock and Furlong, 1922; McGrew, 1937; Stock, 1936; Gazin, 1952; Galbreath, 1953; Guthrie, 1971; West and Dawson, 1974; Setoguchi, 1974; Lillegraven, 1976; Krishtalka and Stucky, 1983a, 1983b, 1984; Russell, 1984; Storer, 1984). However, of Cope's (1873a, 1873c) original six species, one ( Herpetotherium marginale ) was later transferred to the geolabidid eulipotyphlan genus Centetodon Marsh, 1872a (McKenna, 1960; Lillegraven et al., 1981), and two ( Herpetotherium hunti and Herpetotherium stevesonii ) were transferred to other marsupial genera (Crochet, 1980; Scott, 1941; Krishtalka and Stucky, 1983a, 1983b; Korth, 1994).

Lavocat (1951) and Hough (1961) first proposed that the type species, Herpetotherium fugax , should be regarded as generically distinct from Peratherium based on the incorrect assumption that an inflected mandibular angle was lacking in the former, but continued to include most of the remaining North American species in Peratherium . Crochet (1977) went further with the resurrection of Herpetotherium , referring many of the North American species to the genus based on certain differences in the cheek teeth, but his proposal was not widely accepted. Stock (1936) described Peratherium californicum from the Uintan of California and Gazin (1952) described Peratherium chesteri from the late Wasatchian of Wyoming. Setoguchi (1973) proposed that Peradectes protinnominatus Simpson, 1928 , is a junior synonym of Peratherium chesteri , which he considered as probably assignable to Peradectes within a peradectid lineage that included Peradectes Matthew and Granger, 1921 ,

and Nanodelphys McGrew, 1937 . Setoguchi (1973) also suggested that Nanodelphys may be a junior synonym of Peradectes , whereas Lillegraven (1976) retained Nanodelphys as a valid genus and transferred Peradectes californicum to Nanodelphys as N. californicus . Bown (1979) followed Setoguchi's (1973) proposal and formally synonymized Peratherium protinnominatus with Peratherium chesteri , and assigned Peratherium chesteri to Peradectes . Crochet (1978, 1979, 1980) provided a different taxonomic scenario as follows: 1) erected the subgenus Peradectes for the type species, Peradectes elegans Matthew and Granger, 1921 ; 2) reduced Nanodelphys to a subgenus of Peradectes , which resulted in Peradectes californicus along with Nanodelphys minutus McGrew, 1937 , and Peradectes protinnominatus being assigned to Peradectes (Nanodelphys) and 3) transferred Peradectes chesteri to Herpetotherium (a mistaken assignment that Krishtalka and Stucky [1983b] later corrected). Subsequently, Krishtalka and Stucky (1983b) demonstrated that Peradectes protinnominatus is specifically distinct from Peradectes chesteri and, following Setoguchi (1973), regarded Nanodelphys as a synonym of Peradectes . However, Korth (1994) provided convincing evidence that Nanodelphys minutus is a junior synonym of Herpetotherium hunti Cope, 1873c , then recombined the nomenclature for the genotype to Nanodelphys hunti and demonstrated that it is generically distinct from Peradectes , making the genus monotypic.

Fox (1983) recognized certain differences in the antemolar dentition between Herpetotherium fugax and European species of Peratherium , supporting Crochet's (1977, 1980) original proposal. Korth (1994) provided further evidence that most Chadronian through Hemingfordian herpetotheriid species previously referred to Peratherium should be allocated to Herpetotherium , but retained all Duchesnean and older species in Peratherium . Based on size and certain dental differences, Korth (1994) also erected a new genus, Copedelphys , wherein he transferred two Chadronian species previously assigned to Peratherium ( Peratherium titanelix Matthew, 1903 , and Peratherium stevensonii Cope, 1873c ) to his new genus. Rothecker and Storer (1996) concluded that Peratherium innominatum Simpson, 1928 , actually represents a primitive species of Copedelphys and reassigned it to C. innominata . Rothecker and Storer (1996) also recognized additional differences between the cheek teeth of Herpetotherium and Peratherium , and referred almost all Wasatchian through Hem-

ingfordian North American herpetotheriids to Herpetotherium , which was also followed by McKenna and Bell (1997) and Korth (2008). However, Beard and Dawson (2001) referred material from the early Eocene (Wasatchian) Tuscahoma Formation of Mississippi to the European species, Peratherium constans . Thus, except for Peradectes constans , Peradectes californicus , Peradectes chesteri , Nanodelphys hunti , the three species now assigned Copedelphys ( C. titanelix , C. stevensonii , C. innominata ) and the one species transferred to Centetodon ( C. marginale ), the current consensus is that all the other Paleogene and early Neogene North American herpetotheriids originally assigned or transferred to Peratherium are now allocated to Herpetotherium . Two additional herpetotheriid genera have also been recognized from the Paleogene of North America, Swaindelphys Johanson, 1996 , from the middle Paleocene of Montana and Estelestes Novacek et al., 1991, from the middle Eocene of Baja California, Mexico. Horovitz et al. (2009) referred two partial skulls to Herpetotherium sp. , cf. H. fugax that exhibit procumbent lower incisors and a large stylar cusp C on M3, characters typical of the genus (Fox, 1983; Korth, 1994).

Copedelphys innominata , Peradectes chesteri , and two species of Herpetotherium , H. knighti (McGrew, 1959) , and H. marsupium (Troxell, 1923a) , are currently recognized from the Bridger Formation (Gazin, 1976; West and Hutchinson, 1981; Krishtalka and Stucky, 1983b; Gunnell, 1998; Rothecker and Storer, 1996; Korth, 2008). The TBM specimens can be confidently assigned to Herpetotherium because they exhibit the following diagnostic characters ( Setoguchi, 1975; Krishtalka and Stucky, 1983a, 1983b; Rothecker and Storer, 1996; Korth, 2008): 1) a dilambdodont centrocrista on M1-2; 2) a reduced paracone on M1-2 that is smaller and lower in height than the metacone; 3) a large, conical, relatively tall entoconid on m1-3; and 4) a posteriorly projecting hypoconulid on m1- 3 that is shelf-like, smaller, and notably lower in height than the entoconid and separated from it by a distinct notch.

As with other Paleogene and early Neogene marsupials, species of Herpetotherium are differentiated primarily by size, dental proportions, and differences in the occurrence and positions of the upper molar stylar cusps (Korth, 1994, 2008). Although many early studies dealt with very small samples, more recent investigations utilizing larger sample sizes for North American species of Herpetotherium have demonstrated that a fair degree of individual variation in size and in the occurrence and positions of the upper molar stylar cusps can be discerned (e.g., Eberle and Storer, 1995; Rothecker and Storer, 1996; Hayes, 2005; Kihm and Schumaker, 2015; Korth, 2015). Ladevèze et al. (2005) reported similar examples of individual variation in the herpetotheriid marsupials of Europe. These studies indicate that the frequency and morphological trends for these character states should be utilized to differentiate species rather than definitive diagnostic statements regarding these character states.

Based on a partial dentary with p3-m2 (YPM 13518) from the Bridger Formation of the Bridger Basin, Troxell (1923a) described Herpetotherium marsupium and differentiated it from Entomacodon minutus Marsh, 1872a (= Herpetotherium knighti see McKenna, 1960 and Robinson, 1968; see also Krishtalka and Stucky, 1983b for nomenclatural priority) by having a presumably more diagonally transverse trigonid wall on the lower molars. The holotype of Herpetotherium knighti , a partial maxilla with M1-3 (AMNH 55684), came from the Bridger Formation at Tabernacle Butte and was initially characterized by McGrew (1959) by having the following: 1) length of M1-3 equals 5.9 mm; 2) three principal stylar cusps on M1-3 that are low and positioned on extreme labial border of the stylar shelf; and 3) a well-developed paraconule and metaconule on M2-3. Lillegraven (1976) assigned specimens from the Bridger Formation and Teepee Trail Formation of Wyoming to Herpetotherium knighti and specimens from the Uintan Friars and Mission Valley formations of California to Herpetotherium sp. , cf. H. knighti . However, Krishtalka and Stucky (1983b) considered Lillegraven's sample from California to represent a mixed sample of the larger Herpetotherium knighti and the smaller Copedelphys . innominata . Subsequently, samples ranging from the Wasatchian through the Duchesnean have been referred to Herpetotherium knighti and Herpetotherium marsupium , with H. knighti being distinguished from the latter by having the following ( Setoguchi, 1975; West and Dawson, 1975; Bown, 1982; Krishtalka and Stucky, 1983a, 1983b, 1984; Russell, 1984; Storer, 1984; Rothecker and Storer, 1996): 1) smaller size; 2) significantly shallower M2-3 ectoflexi, especially on M3; 3) M1-3 stylar cusps positioned along extreme labial border of the stylar shelf; 4) stylar shelf usually lacking crenulation along its labial border; 5) less tendency toward reduction of cusp C on M1-3 and less separation of cusps C and D on M1-2; 6) lack of a tendency towards twinning of stylar cusp

C on M1-3; and 7) m1-3 cristid obliqua with ten- dency to terminate slightly more labially on the pos- terior wall of trigonid. The TBM specimens are intermediate in size between typotypic samples of Herpetotherium knighti and Herpetotherium marsu- pium, but within the observed ranges of the larger specimens of Lillegraven's (1976) sample of Her- petotherium sp., cf. H. knighti from the Uintan of California, which Krishtalka and Stucky (1983b) referred to H. knighti . In all other dental characters, especially the shallow ectoflexi on M1-3, labially positioned stylar cusps and the close positioning of cusps C and D on M1-2 with the tendency to share a common base, the TBM specimens are indistin- guishable from those of Herpetotherium knighti . Considering the individual variation in size noted above for large samples Herpetotherium , we refer the TBM specimens to Herpetotherium knighti .

Figures

FIGURE 11. Upper molars of Herpetotherium knighti: 1, LM1, SDSNH 110429; 2, RM1, DMNH 75286; 3, LM2, UCM 68793; 4, LM2, UCM 70637; 5, RM2, UCM 95783; 6, RM2, UCM 70639; 7, RM3, UCM 68794; 8, RM3, UCM 73770; 9, RM4, UCM 95784. All occlusal views. Scale bar = 1 mm.

FIGURE 12. Lower cheek teeth of Herpetotherium knighti from TBM: 1, Rdp3?, UCM 68902; 2, Rm2-3, UCM 68574; 3, Rm1, UCM 95792; 4, Lm3-4, UCM 95780; 5, Lm2 or 3, SDSNH 110339; 6, partial Rm2 or 3, SDSNH 110427; 7, Lm2 or 3, UCM 70977; 8, Rm2 or 3, UCM 95785; 9, Rm4, UCM 67896; 10, Rm4, UCM 95793. All occlusal views. Scale bar = 1 mm.