Belgicasorex ramboeri, Smith & Van Den Hoek Ostende, 2006

Smith, Richard & Van Den Hoek Ostende, Lars W., 2006, Dental and tarsal morphology of the European Paleocene / Eocene " condylarth " mammal Microhyus, Acta Palaeontologica Polonica 51 (2), pp. 381-384 : 381-383

Belgicasorex ramboeri sp. nov.

Holotype: IRSNB−M−1903, fragmentary right maxilla with P4, M1, M2.

Paratype: IRSNB−M−1904, right M3.

Derivation of the name: After Geert Ramboer, who found the holotype and some other specimens from Hoogbutsel.

Type locality and horizon: Hoogbutsel near Boutersem, Flemish Brabant, Belgium ( Glibert and de Heinzelin 1952; Smith 2003), lowermost Oligocene, Boutersem Member, Borgloon Formation, Early Rupelian, reference level MP 21 (see Aguilar et al. 1997).

Other locality: Boutersem TGV (see Smith 2003, 2004).

Differential diagnosis.— Belgicasorex ramboeri gen. nov. sp. nov. differs from all the members of the Heterosoricidae by having a M2 with quadrangular outline, anterior and posterior sides of equal length. Belgicasorex differs from Wilsonosorex Martin, 1978 and Lusorex Storch and Qiu, 2004 by the absence of distinct paraconules and metaconules on M1–M2. It differs from Domnina Cope, 1873 in not having an individualised hypocone and by a much wider posterior basin on the M1 and M2. Belgicasorex differs from Heterosorex Gaillard, 1915 by having a quadrangular outline of P4 rather than triangular, a wider than long M3 with a postprotocrista ending in the trigon basin, and better developed lingual cusps on M1 and M2. The latter character also sets Belgicasorex apart from Quercysorex Engesser, 1975 .

Belgicasorex differs principally from Dinosorex Engesser, 1972 by the absence of reduction in size from M1 to M3 resulting from a square outline of the M2.

Description

The infraorbital canal is well preserved on the maxilla fragment. The canal is short in length but large in cross section. The infraorbital foramen is oval. It lies above the middle of P4 to the middle of M1. The lacrimal foramen is a small circular foramen above the anterior root of M2. Behind the lacrimal foramen a fragment of the jugal is present.

P4 is sub−trapezoidal in occlusal view. The labial side, much longer than the lingual one, is rectilinear except at the base of the paracone which curves convexly. The lingual side is slightly curved convexly. The posterior side is much wider than the anterior one. A well−developed cingulum runs around the tooth. The paracone is high. There is no metacone. The postparacrista is long (about half of the labial side) and curves obliquely from the top of the paracone to the metastyle. The latter is indistinctly separable from the postparacrista. The preparacrista is short and connects the distinct parastyle to the paracone. The parastyle is not really in the anterolabial corner of the teeth but somewhat distant from the labial side. The protocone is low and small. A very small hypocone is situated postero−lingual from the protocone. The posterior basin is pyriform in occlusal view, wide in the lingual part and deep.

M1 has the largest occlusal surface of all the cheek teeth. It is sub−quadrate in occlusal view. The anterior side is slightly curved. The paracone and metacone are high and antero−posteriorly compressed. The sharp ridges of the paracrista, the centrocrista and the metacrista form a wide W−shape. The paracrista is shorter than the metacrista. The paracone is more labially situated than the metacone. The mesostyle is undivided from the centrocrista. The preprotocrista is connected to the base of the paracone and the postprotocrista is connected to the base of the metacone. The lingual cusps are well defined. The hypocone is weaker and somewhat more lingually situated than the protocone. The prehypocrista is connected to the end of the postprotocrista at the base of the metacone, dividing the anterior and posterior basins of the teeth. A posterior ridge runs from the hypocone along a wide and deep basin to the metastyle. The ridge is first convex then somewhat concave. There are neither a paraconule nor a metaconule.

M2 is a little shorter than M1. The outline is more quadrangular. The posterior width is equal to the anterior width. The W−shape is symmetrical because the paracrista is subequal to the metacrista. The paracone and the metacone are equidistant from the labial side. The preprotocrista is shorter than in M1.

M3 is much smaller than M1 and M2. The labial side is strongly oblique. The paracone is well developed. The metacone is reduced. The W pattern is incomplete due to the absence of a postmetacrista. The protocone is small and placed close to the anterior side. The preprotocrista is short and ends at the base of the paracone, separated from it by a notch. The postprotocrista is not clearly defined and ends in the trigon basin. This basin is lingually wide open. The lingual side is semi−circular. There is no hypocone.

Discussion

The presence of Belgicasorex ramboeri in the Belgian localities Hoogbutsel and Boutersem TGV attests that the family Heterosoricidae was already present in Europe at the earliest Oligocene (MP 21). At that time, endemic insectivores were in competition with new immigrants ( Smith 2004). In all probability, Belgicasorex ramboeri must be considered with the erinaceid Tetracus nanus and the nyctitheriid Oligonyctia hoffmani as new immigrants to Europe.

The oldest representatives of the Heterosoricidae are known from North America, where the genus Domnina Cope, 1873 has been found in the late Eocene of Wyoming (Krishtalka and Setogushi 1977). In Asia, the oldest published find was Gobisorex kingae Sulimski, 1970 from the late Early Oligocene of the Hsanda Gol Formation in the Gobi Basin ( Reumer 1998). Unfortunately, it is known from the lower dentition only, so that no comparison with Belgicasorex can be made. In the meantime, Gobisorex has also been found in lowermost Oligocene strata from the Hsanda Gol Formation (Gerhard Storch, personal communication 2005). Belgicasorex in Europe and the early presence of Gobisorex in Asia show that by the Early Oligocene the Heterosoricidae had already undergone a diversification and had a wide geographic range. In contrast, the oldest known Soricidae comes from the Lower Oligocene Ergilin−Dzo Svita ( Mongolia) ( Yanovskaya et al. 1977). Therefore, the stratigraphic distribution, in combination with the anatomical differences underlined by Reumer (1987), seems to indicate that the Heterosoricidae evolved independently from the Soricidae , although both could be derived from the Nyctitheriidae .

The description of a new Oligocene heterosoricid genus in Europe also provides fresh insights into the early history of the family. The previously European Oligocene finds of Heterosoricidae were placed either in Quercysorex or in Dinosorex . Van den Hoek Ostende (1995) noted the peculiar stratigraphic distribution of Dinosorex , which was present in the Oligocene of Europe, but absent during most of the Early Miocene of that continent. During that period it was found in Anatolia. This picture has changed somewhat, as Ziegler (1998) transferred the only species of Dinosorex described from the Oligocene, D. huerzeleri , to Quercysorex . Still, this left two records of Dinosorex in the European Oligocene, Dinosorex sp. from Cournon−Les Souméroux (MP 28) ( Brunet et al. 1981), and Dinosorex sp. from Boudry−Trois−Rods ( Mojon et al. 1985). The M2 found at Cournon−Les Souméroux ( Brunet et al. 1981, fig. 11) is only marginally narrower at the posterior side than at the anterior side. In this respect it resembles Belgicasorex more than Dinosorex . Since we have not seen the original material, we tentatively assign the assemblage from the French locality to Belgicasorex , and consider it best classified as cf. Belgicasorex sp. Comparison with Dinosorex sp. from Boudry−Trois−Rods ( Switzerland, Oligocene) is not possible, since the species is represented by a fragmentary mandible only. However, this fragment shows no diagnostic features that allow identification even at the genus level.

Because there is no evidence for the presence of Dinosorex in Europe during the Oligocene, the pattern in the distribution of Dinosorex observed by Van den Hoek Ostende (1995) may well be artificial. It would be more logic to assume the genus migrated into Europe from Asia at the end of the Early Miocene, Dinosorex anatolicus from the Lower Miocene of Turkey being its oldest representative. This scenario is corroborated by the morphology of the P4. Although both Belgicasorex and the younger species of Dinosorex posses a trapezoid to square P4, this element has a triangular outline in the most primitive Dinosorex , D. anatolicus . Thus, the square outline of the P4 developed separately in Belgicasorex and Dinosorex . Assuming this picture is correct, it is noteworthy that both in the Oligocene and Miocene the heterosoricids are represented by a genus with well developed lingual cusps ( Belgicasorex and Dinosorex , respectively) and one in which these cusps are much less developed ( Quercysorex and Heterosorex ). Presumably, this represents some trophic adaptation, related to different niches.