Wittry, then a student at the University of Wisconsin, reported on June 22 and worked the remainder of the time spent in the area. Landowners in the survey area and various residents of Glen Ullin uniformly cooperated and provided assistance of various kinds. Hewes provided us with information regarding his investigations and made helpful suggestions based on his experience in the Koehler site.
Riley, Department of Geology, University of Nebraska, assisted with the identification of the stone material, and Dr. Members of the laboratory staff at the River Basin Surveys headquarters in Lincoln were instrumental in the preparation of this report in several ways.
INTRODUCTION
The region's climate is characterized by long, cold winters, hot summers, relatively little rainfall, and an average growing season of approximately 120 days. There are favorable factors in the region's climate that tend to counterbalance the shortness of the growing season and the decrease in rainfall. These are, on the one hand, the long hours of daylight during the summer months and, on the other hand, a relatively low rate of evaporation and the fact that a large part of the annual precipitation occurs during the growing season (approximately 70 percent during the months April to September).
West of the Missouri River in areas immediately adjacent to North Dakota, two sites showing relationships to Missouri River complexes have been investigated and reported. The first number indicates the country and represents the country's numerical position in the alphabetical list of the United States; the two letters following the number are the symbols for the district; and the final number indicates the specific location in the district.
THE KOEHLER SITE (32GT1)
EXCAVATIONS
TERRACE
One baseline extended from this point to the northeast end of the terrace in a direction 30 degrees west of magnetic north, and the other was perpendicular to the same point. In the following description of the excavations, references to directions are to these baselines and not to actual compass directions. Because the time and manpower available for site investigation were strictly limited, only relatively small-scale excavations could be undertaken.
They consisted of trenches, usually 2.5 feet wide, through various parts of the occupation area. A generally similar stratigraphy was found to prevail over the overall parts of the investigated site, except in Excavation Unit 6, which was at a higher elevation, and an area that had been recently disturbed. It was exceptionally well represented on the part of the N45 line shown on Plate 3, h. Zone B was a layer of light-tan material, predominantly a very fine one. sandy over most of the site, but with a clayey component in excavation1, near the foot of the adjacent slopes.
The thickness of the layer varied from 0.25 to 0.8 feet and averaged about 0.5 feet in most of the site. It was not detectable in limited areas of the site where it may have been destroyed by the activities of the indigenous inhabitants. The top of the upper culture layer was 1.6 feet below the surface, the base of the lower was 2.8 feet.
Feature 19 was a dense deposit of bones, mostly bison, found at the western end of Excavation Unit 1 (Fig. 1) and extending north and south of the ditch boundaries. The depths at which heat affected the color of the underlying soil also varied considerably, indicating differences in the duration of the fires that burned on the hearths. In the main part of the site, dominated by the Classic profile, hearths were found from the top to the bottom of Zone D, although the majority (8 out of a total of 12) were in the lower half of this deposit, with 4 lying at or near its base.
ARTIFACTS
A large proportion of ordinary sherds are probably from the neck area or from the rim, although very small. The outer surface of the edge was thickened by adding a 9mm spacer. The latter example is also unusual in the series in terms of its thickness and the height of the decorated rim.
Thinning of the lips in the last category was usually achieved by chamfering the inner rim edge (pi. Of the 5 sherds (3 vessels) from unmodified, extended rims, 1 is certain and 2 others (1 vessel) are probably decorated only on the lip, which in each instance is thickened to the outside, is flat and slopes outwards. 2 sherds (1 vessel) are decorated on the lip and both faces of the rim with closely spaced, unusually fine cord-impressed lines.
The absence from the sparse Koehlersite collection of the more elaborate forms and of features such as checkered surfaces, handles, and cord-wrapped stick decorations is of little significance as these are reportedly rare in the major collections of the Missouri River sites. The largest difference between our sample from the DoubleDitch site and the material from the Koehler site is in the thickness of the sherds; the mean for the latter is considerably smaller than that for the samples collected from the surface of the former. The points shown by the smaller fragments probably did not greatly exceed the range of the complete specimens; the largest (pi. 8.5) is 18.5 mm.
One specimen has a grave-like point at one end of the beveled edge of the scratch (p. 8, IJj). Most of the original surfaces have been reduced, but a small remnant near a rounded corner of one of the pieces bears a short segment of an incised line (pp. 8, 26). As with the pottery, it appears that the stone artifacts from the Koehlersite fall within the range of materials from the Double Trench site.
WORK IN BONE AND ANTLER
Although so fragmentary that no definite statement is possible, these objects seem to resemble incised rectangular and tabular stones reported by Will and Spinden from the Double Ditch site (1906, p. Of similar material is an asymmetrical but basically disc-shaped bead with a conical central bead Perforation (pi. 8, 30) Finally, as regards tools or other stone objects, there are two unshaped cobbles of igneous rock, each with one end somewhat bruised, suggesting that they used as a hammer.
Although the descriptions of, for example, projectile points from the latter location are too general to allow a satisfactory comparison, there seems no doubt that this is at least the case. I don't recall seeing this kind of material, at least in quantity, on other cultural affiliate sites and the like. Three are made from the lateral surface of a large rib, one is probably from the thickened edge of a large scapula.
These specimens appear to be similar to the bone objects from Double Ditch illustrated by Will and Spinden (1906, fig. 35, n, o) and, in general form and nature of the working ends, to objects from the vicinity of Mobridge, S .Dak., described by Wedel (1955, pp. 123-126). One is an otherwise unaltered point that was apparently broken rather than cut from the antler (Fig. 11, 11), and the other appears to have been made from the thickened rim of a large spatula. Cut marks are visible at both ends and there are several short transverse cut lines unevenly distributed at various points on the specimen, which is 14 mm long and 6 mm in diameter (Fig. 11, ^).
Its function is uncertain, but it is suggestive of specimens from the Double Ditch site (WillandSpinden, 1906, p. 172 and pi. 36,x). The remaining specimen in the collection of the Eiver Basin Surveys is the distal end of a bison metapodia, which has been cut from the shaft and is presumably a by-product of preparing the shaft for the manufacture of an artifact. In each case a few narrow notches are cut into the rim, and the wear produced by use is on the cut side of the rim.
CONCLUSION
5, h) that it seems certain that they were deliberately ground, probably for cooking for the manufacture of bone fat (Leechman, 1951). It is not possible to make a good estimate of the number of individuals represented by the bison and the dog bones, but based on the number of the most common elements present (divided by 2) it can be said that there are at least eleven bison and at least are three dogs. Morrison, they include six species of freshwater mussels, most of which are represented by only a few specimens.
No evidence of structures of any kind was discovered, nor was there any evidence of burial pits, which facts lead to the conclusion that. If the bones found in this area can be considered conclusive, bison were exploited almost to the exclusion of other game animals; nodeer orelkbonesh are identified in collections. The extremely fragmentary condition of the bones, perhaps indicative of the extraction of bone fat, seems to imply that the site.
The artifact inventory, while leaving much to be desired in terms of quantity and range of forms, seems unmistakably to point to a close relationship with certain sites that have been investigated on the Missouri River near the mouth of the Heart River. This negative evidence, together with the positive evidence of a close cultural relationship with the double ditch site, points to occupation in the 18th century and perhaps into it. It seems likely that most of the other sites recorded in the reservoir area (see Appendix) can be similarly identified, although the collections are too few to allow more tentative identifications.
In any case, there seems no doubt that the western tributaries of the Missouri River constitute a fertile field for future archaeological research.
LITERATURE CITED
APPENDIX
SITES IN THE HEART BUTTE RESERVOIR AREA
The shelter faces northwest, in the direction of the prevailing wind, and would be uninhabitable on windy days; the sand blown by such a wind made work almost impossible during the half day the Kiver Basin Surveys group spent here. The pottery, which is apparently in the same tradition as that at the Koehler site, has been described by Hewes (1949b, p. 61). A point in the collection on loan from the State Historical Society of North Dakota (fig. 12, c,8) is small and triangular and has a slightly concave base and a few side notches.
Four small pottery shards, collected from the terrace slope below the site, are decorated with lines impressed with string and again. PAP Ncf'/j^''' HEART BUTTE RESERVOIR AREA — COOPER 39 Site 32GT9 is a circle with some boulders said to be in the SE14 SWi^ sec. The surface cultural evidence consisted of a single-cord-impressed rim sherd, similar to that found in the excavations at the Koehler site, two re-.
Site32GT11, indieNWi/4 SWy4 sec.3 T136N R90W, is an occupation site buried under slope wash on a terrace cut by the Hart River. Two pottery sherds, one plain and one plain stamped (with parallel grooves and ridges), a triangular kert end scraper, a broken blade of "Knife River flint" and cores, splines and flakes, mostly of "Knife River. Site 32GT12 is high land on the north side of the Hart River in the SEi/4 SWi^sec.
3 T136N R90W, is a campground located on a terrace about 15 feet above the Heart River and on the north side of that stream. Four small single-stamped sherds and fragments of "Knife River flint" were found on the surface of the terrace, which is well below the reservoir level. Site 32GTlli.is a camp site on a fairly low terrace on the north side of Heart River, in NEi^ SW14 sec.
Site 32GT15 occupies the summit of a node on the north side of the HeartRiverintheNW14 NIV14 sec.34T137N R90W and within the floodplain. 169 Site S2GT16 is a small area on the north side of the Heart River which has yielded a number of flakes and cores of "Ejiife River feldspar".
