The White River Group of the Great Plains includes the Chadron and Brule Formations of late Eocene to Miocene age. It reaches its maximum thickness in northwestern Nebraska, decreases in thickness northward, and disappears in northern North Dakota (Larson and Evanoff 1998). Pioneer Trails Regional Museum (PTRM) (Bowman, North Dakota) locality V89002 is located in the Medicine Pole Hills of southwestern North Dakota (Figure 1; South Rhame quadrangle). The lower part (Chalky Buttes Member) of the Chadron Formation, remnant of a former sedimentary "blanket" (Trimble 1980), caps those hills in many places (Murphy et al. 1993; Pearson 1993), and V89002 is located on one of the hilltops. Fossils were first reported from the area some 80 years ago (Leonard 1922), and since 1989 the Pioneer Trails Regional Museum has conducted excavations at V89002.

The sediment is a brownish, poorly sorted, poorly consolidated, generally medium- to fine-grained sandstone (Pearson 1993). A 2.5 m local section (Figure 2) shows two trough cross-bedded units (units 2 and 6); unit 5 is composed of sediment indistinguishable from that of unit 2. Along the bedding planes are strewn gray mudballs, often comprising 15% of total rock volume. The troughs of units 2 and 6 have trends at high angle (~80°) to one another. Unit 3 is a fine-grained, parallel-bedded sandstone. Unit 4 displays planar cross-bedding.

Rapidly shifting channels that deposit planar and trough cross-bedded sandstone units (lithofacies Sp and St of Miall 1996) are characteristic of the distal reaches of some modern braided rivers (Einsele 1992), and I postulate that this section was deposited as part of a braidbelt environment. Parallel-bedded units, like unit 3, are also deposited in such environments during waning flood stages (Einsele 1992: 45). The mudballs probably derive from the underlying Paleocene strata (Kihm personal commun., 2004). Many fossils are quite worn, presumably by stream abrasion. The similar color (creamy to tannish yellow) and preservation of nearly all specimens argue for a single provenance.

The Medicine Pole Hills l.f. is taphonomically similar in many ways to the Calf Creek l.f. of Saskatchewan (Holman 1972, 1976). The Cypress Hills Formation, which yields the Calf Creek l.f., is also interpreted as a braidbelt deposit (Leckie and Cheel 1989). As at V89002, the bone in the Cypress Hills Formation is set in a poorly sorted matrix that also contains mudballs torn from interfluves (Leckie and Cheel, op. cit.), although I interpret the Medicine Pole Hills fossil horizons as fluvial in origin rather than as debris flows. All fossils were collected from units 1 and 2 by dry-screening quarried sediment; fossils of the Calf Creek l.f. were also collected by screening (Holman 1972). Neither deposit appears to exclude particular size classes, both preserving everything from large brontotheriids to small marsupials (Storer 1996; Kihm et al. 2001; Pearson, personal commun., 2001).

The Chadron Formation in the Medicine Pole Hills and the Cypress Hills Formation in the Calf Creek area are both Chadronian in age, although more carefully considered, they may differ by 1 m.y. or so. The leptomerycid artiodactyls present in the Medicine Pole Hills l.f., including a probable early form of Leptomeryx yoderi, suggest an early Chadronian age for these deposits (Heaton and Emry 1996; Hoganson et al. 1998). In contrast, the Calf Creek l.f. contains L. speciosus and L. mammifer (Storer 1996), which are indicative of a medial Chadronian age (Emry et al. 1987).