Boreodon matutinus

DESCRIPTION: although many isolated lower teeth and mandibles of Boreodon have been recovered in the Oldman Formation (Russell, 1952), the upper dentition has yet to be described. At least four upper teeth are here assigned to the genus on the basis of size, morphology, and a similarity to those of Didelihodon. The only complete upper tooth is an M 1 and possibly a DP 3. Other teeth referable are the labial and lingual halves of either M 2 or M 3. DP 3, AMNH 77386, is a large, worn, but almost complete DP 3 of Boreodon (fig. 14 C). The width of the crown is 3.6 mm. and this is approximately equal to the estimated length of the tooth. The stylar shelf is broken posteriorly and the most prominent cusp is stylar cusp B. The parastylar area is expanded. The paracone is higher than the closely situated metacone and is linked to stylar cusp B by a ridge. Both conules are present. The protocone is anteroposteriorly compressed. Ml; a worn, but complete M', AMNH 77385 (fig. 14 A), has been recovered. Its stylar shelf is incipiently bilobate with a long posterior wing. Stylar cusp A is small and connected by a ridge to cusp B, which is the largest stylar cusp. Two ridges originate from the apex of cusp B, one connecting it to the paracone and the other to cusp C. Cusp C is situated at the stylar cleft or slightly anterior to it. Three small cuspules lie on the wide stylar shelf labial to the paracone and metacone and between stylar cusps C and D. Cusp D is elongated anteroposteriorly, and a small cuspule is present slightly anterior to it. The shelf is not expanded labial to cusp D as it is in the posterior molars. Cusp E (metastyle) is connected to the metacone. The conules are well developed, the paracone is distinctly larger and is connected to stylar cusp A by a cingulum. The protocone is a low cusp. Although roots are not present it is possible to make out an interradicular crest linking the lingual sides of the two labial roots. M 2 or M 3; a tooth fragment, AMNH 79383, is probably the labial part of an M 2, but it could be part of an M 3 (fig. 14 B). Its stylar shelf is asymmetrically bilobate, and stylar cusp A is small and well separated from cusp B. Stylar cusp B is not connected to the paracone by a ridge and is situated farther from that cusp than in M 1. Stylar cusp C is situated slightly anterior to the stylar cleft, and it is smaller than stylar cusp D. As in M 1, three cuspules are present on the broad stylar shelf lingual to cusps C and D. Stylar cusp D is relatively well developed and the shelf area labial to it is expanded. A similar condition is found on molars of Didelphodon. Another tooth fragment, AMNH 77381, is the lingual portion of either an M 2 or an M 3. Both its conules are present, but the paraconule is the better developed. Pre-paraconular and post-paraconular wings diverge at a greater angle than the pre-metaconular and post-metaconular wings. The protocone is a low cusp. P 2; the partial tooth, AMNH 77387, is probably a P 2 (fig. 14 G, H). It is large, robust, and double rooted (length 4.0 mm., width 2.4 mm.) with two cusps. The anterior cusp is massive and its apex is worn. The posterior accessory cusp is not so sharply differentiated from the anterior cusp as in P 3 but is connected to it by a ridge. I consider the type specimen of Boreodon matutinus Lambe (1902) to be P 2 of the species generally known as Eodelphis cutleri Simpson (1928) and not a P 3 as was suggested by Russell (1952). P 3; this premolar is larger than P 2 and M 1. The anterior cusp of AMNH 77388 is so worn that it is not possible to make out many details (fig. 14 I). The large cusp is narrow transversely with a worn ridge leading from its apex to its valley separating the posterior accessory cusp. The latter is a small, pointed cusp with two ridges arising from its apex, one extending toward the anterior cusp, the other extending labially. P 3 iS two rooted, and the posterior root is larger than the anterior. M 2; although AMNH 77393 (length 4.4 mm., fig. 14 J, K) is nearly the same length as M 2 in the mandible, BM Ml 1532, described by Woodward (1916) as the type specimen of Cimolestes cutleri, the tooth is narrower (2.8 mm.). Width of M 2, BM M 11532, is 3.5 mm. The trigonid is anteroposteriorly compressed, the paraconid is a large, lingual cusp. The metaconid is slenderer and smaller than the paraconid in AMNH 77393. The protoconid, although worn, is the largest cusp of the trigonid and is situated slightly anterior to the metaconid. The rugose anterobasal cingulum is well developed but does not extend to the labial side of the protoconid. Width of the trigonid is equal to that of the talonid. The hypoconid is the largest cusp in the talonid and well separated from the hypoconulid and entoconid. The hypoconulid is broken, but a cingulum encircles its labial side. The crista obliqua intersects the posterior wall of the trigonid slightly labial to its center. The posterior root is larger than the anterior. M 3; a mandible, AMNH 77394, found during the first year of prospecting (1964) fortunately has an unquestionable contact with an M 3 found from the same locality. This specimen, AMNH 77394, from the Judith River Formation, consists of the posterior part of the ramus with the associated M 3 and the alveolus for M 4. M 3 (fig. 14 D-F) is larger (length 5.5 mm.) than AMNH 77393, which is discussed above. But the trigonid is similarly anteroposteriorly compressed, and the paraconid is slightly taller than the slender metaconid. The talonid cusps are similar in the two teeth, except that the hypoconulid is directed posteriorly. The anterobasal cingulum extends to the labial side and is not so restricted as in AMNH 77393. All articular surfaces are broken away from the mandible (fig. 14 D-F). The masseteric fossa is deep and well defined. The greatest transverse width of the dentary is at M 4, 6.8 mm., but in BM M 11532 itis 8.0 mm.

Figure 14 A-K

REVISED DIAGNOSIS: A large didelphoid with robust upper teeth which have wide, thick, prominent stylar shelves, becoming progressively more asymmetrically bilobate from M 1 to M 3, and stylar cusp B D A = C E.

DISCUSSION: Boreodon matutinus was a species proposed by Lambe (1902) on the basis of a large lower premolar (NMC 1887) from the Oldman Formation, near Steveville, where most of the Campanian mammals from Alberta were collected (Russell, 1952). Although Lambe (1902) placed the specimen in the Multituberculata, Matthew (1916) pointed out its therian affinities. Simpson (1929) was the first to comment on its didelphid relationships. Its didelphid relationships were supported by Russell (1952), who suggested that the holotypic premolar of Boreodon matutinus was probably a P 3 of " Eodelphis " cutleri. Clemens (1966) in his study of the Lance marsupials also considered this likely. No author, however, formally put Eodelphis in synonymy with Boreodon. The chief reasons for not synonymizing the two was because the isolated premolar, NMC 1887, which is the holotype of Boreodon matutinus, was worn and fragmentary, and also because there was lack of information on the kind and size of other marsupials that may have been present in the Judith River and Oldman formations. Also, little notice was taken of an edentulous jaw referred to B. matutinus by Stanton and Hatcher (1905). This lower jaw is similar to the many mandibles assigned to " Eodelphis " cutleri (Russell, 1952). The holotype of Boreodon matutinus, NMC 1887, is a P 2 not a P 3 as was suggested by Russell (1952) and accepted by Clemens (1966). The reasons for considering NMC 1887 as P 2 rather than P 3 are the following. Dimensionally, the tooth is nearly identical with P 2 (AMNH 77387, with a length of 4.4 mm.) rather than P 3 (AMNH 77388, with a length of 5.6 mm.). Morphologically, the posterior median ridge on NMC 1887 arises from the apex of the anterior cusp and extends to the small posterior accessory cusp as in P 2 (AMNH 77387). The anterior cusp is sharply separated from the distinct posterior accessory cusp in AMNH 77383, and the posterior median ridge descends only to the deep valley separating the two cusps. Differences that exist between specimens NMC 1887 and AMNH 77387 are mainly produced by wear. The question of synonymy of Cimolestes cutleri Woodward (1916), based on a right mandible with P 3, M 2, and M 3, BM Ml 1532, and Eodelphis browni Matthew (1916) based on a left mandibular ramus with a complete but worn dental series, AMNH 14169, has been repeatedly discussed in the literature (Simpson, 1928, 1929, 1930; Russell, 1952; and Clemens, 1966). The specimens on which these genera are based were collected close to each other, and from the same stratigraphic horizon, near Little Sandhill Creek, near Steveville, Alberta. The description of Cimolestes cutleri Woodward (1916) was published on May 30 of that year, whereas that of Eodelphis browni Matthew (1916) was published on July 24. Woodward's (1916) reference of specimen Ml 1532 to Cimolestes was based on its resemblance to C. curtus. Later the specimen was used by Simpson (1 927 b) as the holotype of Diaphoradon, and still later it was put in synonymy with Didelphodon vorax by Clemens (1966). Simpson (1928, 1929) in his review of the Mesozoic mammals, commented on the similarity of the two forms and finally (Simpson, 1929) put E. browni Matthew (1916) in synonymy with Cimolestes cutleri Woodward. Russell (1952) agreed with Simpson's point of view on the basis of the many mandibles at his disposal. More recently, however, Clemens (1966, p. 59) has shown that Eodelphis browni Matthew (1916) and Cimolestes cutleri Woodward (1916) are not conspecific, as there are differences in incisors and premolars. The present material from the Judith River Formation includes an upper molar of Alphadon cf. A. rhaister, AMNH 77372, and a lower molar, AMNH 77371, comprising a large species consistent in morphology with the holotype of Eodelphis. There is no really satisfactory reason for Eodelphis browni and " Cimolestes " cutleri being considered to be congeneric except for their similar size. This implies that " Cimolestes " cutleri does not belong to any described genus except for the long abandoned Boreodon Lambe (1902), with which Simpson (1929), Russell (1952), and Clemens (1966) have suggested a close relationship. None of these authors, however, put the material referred to " Eodelphis " cutleri in synonymy with Boreodon matutinus because of the eroded and worn nature of the type specimen of B. matutinus, now established as P 2. Evidence provided by the premolars (AMNH 77387 and 77388) from the Judith River Formation indicate that Boreodon matutinus Lambe (1902) is conspecific with " Eodelphis " cutleri (Woodward, 1916). Consequently, material previously referred to Eodelphis cutleri is put in synonymy with Boreodon matutinus Lambe (1902). Alternatives to this solution would be referral of " Cimolestes " cutleri to Didelphodon, or the erection of a new generic name to substitute for Boreodon. Neither of these alternatives seems useful. A few upper teeth have been discovered that may belong to Boreodon. Description of a DP 3 and of upper molars is the first attempt to establish the morphology of the upper teeth of this genus. Upper teeth have not been found in association with lowers, and the referral of the teeth is based on their large size and similarity to those of Didelphodon vorax. Boreodon matutinus of the Campanian apparently occupies the same ecolological niche that D. vorax does in the Maestrichtian. Both are large omnivorous marsupials with carnivorous tendencies. Clemens (1966) has discussed the possible ancestry of the stagodonts Boreodon and Didelphodon. He rightly pointed out that the wide stylar shelf with many cusps, especially the prominent stylar cusp B, are primitive features and indicative of a lineage that must have originated sometime before the Campanian and after the advent of the Forestburg therians. The pediomyids differ from the stagodontids in the same way as they differ from the various species of Alphadon, namely in the narrowing of the stylar shelf labial to the paracone, the absence of the feeble development of cusp B, and of the ridge linking that cusp with the paracone. The lower molars of Didelphodon and Boreodon, however, differ from those of the other Cretaceous marsupials by a more prominent anterobasal cingulum, and by more anteroposterior compressed trigonids, on which the metaconid was the shortest and most slender cusp. The paraconid and protoconid are of the same height and form an effective shearing blade. The talonid is relatively wider than in Alphadon and the orientation of the crista obliqua is more labial.

DISTRIBUTION: Judith River Formation, Montana; Oldman Formation, Alberta.

TYPE LOCALITY: Near Steveville, Red Deer River Valley, Alberta.

Boreodon matutinus LAMBE, 1902, p. 79. Boreodon matutinus: STANTON AND HATCHER, 1905, p. 128.

Cimolestes cutleri WOODWARD, 1916 a (May), p. 30; 1916 b (Sept.), pp. 525 - 528.

Eodelphis cutleri: SIMPSON, 1928, p. 148; 1929, pp. 128, 129; 1930, pp. 29 - 32.