In the third subsection some important trends in the cultural record are documented in terms of the distribution and changes in artifact and pit house types. Here the emphasis is upon defining continuity within the sequence of phases in the Vantage locale.
THE SITE AND ITS STRATIGRAPHY
The Site. The Sunset Creek Site, 45KT28, is situated at the north end of Quilomene Bar where Sunset Creek empties into the Columbia River. It extends southward from this creek's mouth in a long, narrow ribbon which flanks the Columbia (Figs. 1, 2, 3 & 4). Directly across the river, to the east, a huge scabland channel flood bar marks the entrance to the Pot Holes, a scabland channel remnant dating from late Wisconsin times. A short distance to the south this steeply faced bar terminates against rugged cliffs which plunge vertically into the Columbia River. To the north it gradually loses its steep scarp, intersecting and covering a series of low basalt benches and projecting westward in a gently sloping tongue which forms a shallow bend in the Columbia River and marks the white water known as Lodged Pole Rapids. Directly across from this tongue of land a sheer cliff 600 feet in height towers above the river, sealing off the north end of Quilomene Bar. One-half mile to the south this cliff is broken by a deep, V-Shaped canyon carved into the basalt bedrock by Sunset Creek whose alluvial fan spills around the end of the cliff and forms the northernmost portion of Quilomene Bar.
The site is located along the southwestern margin of this fan and on the adjacent river flood plain. To the south, Quilomene Bar extends for three miles, terminating in a high, steeply faced scabland channel floodbar whose surface is deeply pocked with giant kettles. Quilomene Rapids is adjacent to the end of this ancient bar, and Quilomene Creek, the largest stream in the immediate area, cuts through the aeolian and floodplain deposits at its base.
 In addition to the Sunset Creek Site, two other pit house sites and at least three separate burial sites are located on Quilomene Bar. In the high, terraced basalt cliffs which rim the bar there are also a series of storage shelters. The survey of these shelters is reported in Appendix C.
Aggregated together, these sites constitute the Quilomene Bar Site Complex as identified in Figure 3 on page 265.
Site Stratigraphy. Interpretation of the stratigraphic sections at 45KT28 has been done with the aid of Mr. Roald Fryxell, who has visited the site and prepared a geological map of the northern part of Quilomene Bar. This information will be published at a later date.
This report was never forthcoming. It is being looked for amongst Roald's papers. If anything is found it will be presented on this website.
The geological section at the site reveals interbedded floodplain and alluvial deposits which are frequently altered by aeolian processes. These deposits were tested to a depth of nineteen feet in two separate locations and to depths of between six and eleven feet at six other stations within the confines of the site (see Fig. 4). On the basis of these excavations six major strata have been recognized. The earliest of these are apparently Altithermal in age and consist of alluvial gravel associated with the Sunset Creek alluvial fan (Stratum 1), a brown-yellow beach sand (Stratum 2), and a more finely textured, grey-brown sand (Stratum 3). The stratigraphic relationships between these strata may be seen in Figures 5, 6, and 10.
Originally, lengthly stratigraphic descriptions accompanied the various profiles. These were simplified and a composite description in stratigraphic order was produced, and is presented below. It, too, was removed in the editorial process.SEDIMENT DESIGNATIONS IN THE HP15 ESCAVATION
1. Top soil, containing rootlets. This layer, always very thin, corresponds to Stratum VD.
2. Light yellow-brown sand, deposited in historic times by the Columbia River. This layer, designated Stratum VC, probably corresponds to the flood of 1948.
3. Dark , sandy soil, redeposited from the crown, or ridge, forming the western border of the site. This layer, designated Stratum VB , probably corresponds to the flood of 1898.
4. Light, yellow-brown sand of the same composition as that described under Soil Designation 2 (see above). This layer, designated Stratum VA, represents an historic flood, probably earlier than 1898.
5. Historic intrusions.
6. Sandy soil, gray-black in color. This soil, which comprises the bulk of Stratum IV, is a combination of cultural refuse and sands similar to those described under Soil Designations 2 and 4 (see above).
7. Dark, charcoal stained, sandy s6ir, varying in color from intense gray-black to black, and usually representing remains of scattered earth ovens or hearths. Its occurrence was common in Stratum IV; it was also encountered in Strata IIIE and IIIL.
8. Light yellow-brown sand of the same composition as that described under Soil Designations 2 and 4. This material was, diversely, deposited by flood waters of the Columbia or carted up from the beach in front of the site for use in cooking and fire dousing.
9. Dark brown-gray soil, with a heavy, humus-like quality; probably decayed organic matter; late in Stratum IV.
10. Dark gray-black, humus-like soil, containing large quantities of fish and mammal bone , and occasional patches of organic material . This material: forms a part of a house floor in House Pit 5.
11. Yellow-white, clay hardpan.
12. Light gray, sandy soil, composed of cultural debris and sand similar to that described under Soil Designations 2 and 4. This material is essentially the same as that described under Soil Designation 6, only lighter. It is characteristic of some of the earlier portions of Stratum IV.
13. Light yellow-grey sandy soil . This material appears to be a mixture of light sand, such as that described under Soil designations 2 and 4, and the dark sandy soil which is denoted by Soil Designation 6.
14. Fine, white ash . 15. Fine, orange-brown ash. This material was encountered in Strata IIIM and IV (Cultural Component VIIA).
16. Fine, light grey ash VIIA.
17 Fine, pink ash (VIIA).
18. Dark red-brown ash (VIIA).
19. Fine, dark brown-black ash (VIIA).
20. Fine, white, burned sand (VIIA).
21. fine brown ash (VIIA).22. Grey-black sand, somewhat darker than the sandy soil described under Soil Designation 6. This material comes forms Strata IIIF and IIIQ.
23. This material forms the matrix of Stratum III, and varies in content and color from place to place It varies in color from light yellow-grey to light yellow-brown.
24. Compacted clay, stained dark gray-black by cultural debris. This material characterizes Stratum IIIM.
25. Fine, light yellow-brown sand. This material is slightly coarser than the matrix send of Stratum III. It characterizes Strata IIIH, III-I, IIIJ , IIIK, IIIL, IIIN, IIIO, and IIIP. 26. Hard, light yell
ow-brown clay. This material is characteristic of Stratum IIIE. 27. Fine, light gray sand, evidently stained by cultural debris. This material is characteristic of Stratum IIID.
28. Fine , light yellow-gray sand, noticeably coarser than the matrix sand of Stratum III. This material has been designated Stratum IIIC.
29 . Coarse, light gray sand. This material is characteristic of Stratum IIIB.
30 . Coarse, gray sand mixed with fine pea gravel . This material is characteristic of Stratum IIIB.
31-35 in and below the distributary channel on which Cultural Component I rests. 31. Dark brown clay.
32. Light, hard, yellow-tan clay.
33. Unsorted, angular, basalt gravel.
34. Coarse, angular gravel, somewhat larger than pea gravel.
35. Fine, light gray sand.
In anticipation of Roald's visit to Quilomene Bar, we excavated a deep sounding along the margin of House Pit 28 (Fig. 4), which is on the slightly elevated floodplain 50 feet in back of House Pit 15, locus of the main excavation. It was excavated to a depth of 11 feet below the surface of the floodplain. It was not included in the publication because it was supposed to become a part of Roald's report on the geology of the site and of Quilomene Bar.
The scabland flood bar, which is the foundation of Quilomene Bar, is a long ramp that begins at river level in the north and rises gradually until it terminates in a downstream slipface at its southern end. The northern toe of the bar is covered with horizontally bedded fluvial deposits. It is these deposits that are revealed in the House Pit 28 sounding.
The Columbia River cut a deep channel into the face of these deposits. There then followed a long period of aggradation during which the six strata were deposited which contain the archaeological remains at the Sunset Creek Site.
The tests in these deep strata, which ranged from twelve to nineteen feet below the present ground surface, were not extensive. Nevertheless, they revealed several cultural assemblages, one of which was sufficiently large to be assigned to the earliest phase currently defined in the Vantage locale.
Mantling these early deposits is a thick layer of sand and floodplain loess roughly equivalent to the first half of the Medithermal. Designated Stratum 4, this deposit varies from seven to ten feet in thickness and contains many culture-bearing lenses. The earliest of these represent the Frenchman Springs Phase and the more recent the Quilomene Bar Phase.
Stratum 5 is a deposit of midden, house fill, and floodplain loess which varies from four to six feet in thickness and is coextensive with Cultural Component VII. The cultural assemblage from it represents the Cayuse Phase at the site.
Stratum 6 is a thin mantle of fully historic deposits including the flood silts of the 1896 and 1948 floods. It contains a variety of historic artifacts as well as a few secondarily deposited aboriginal items. 
The Role of Sand and Loess
This section was omitted from the original publication because it might prompt collectors to dig in places they had never before considered interesting.
Above, the Sunset Creek alluvial fan and the adjacent site of 45KT28. Below, an enlargement annotated for the purposes of the discussion that follows.
Prior to around 2,000 BP, sites along the banks of the Columbia and Snake rivers are characterized by the occupation of sandy beaches during seasonal periods of low water. Between ca. 8,000 and 2,000 BP, the floodplain is aggrading so that beach occupations are covered with sterile layers of sand. Sometimes, there is local seasonal erosion of occupation surfaces, but this is often minimal and many times there is little disturbance at all. Hence, long sequences of occupations are found at sites such as 45KT29, 45CO1, and the Dalles. Such sequences require (1) that sand be deposited consistently in the same places over long periods of time, and (2) that people preferentially make their camps at these locations.
Such deposits are made possible by topographic structures that create conditions in which sand is preferentially deposited. At 45KT28, the Sunset Creek alluvial fan diverts the flow of the Columbia River so that an eddy forms in the lee of the fan ensuring a forever sandy beach. Bedrock obstructions and the downstream slip faces of scabland channel bars also commonly ensure sandy beaches and the preservation of early archaeological materials during periods of floodplain aggredation.
Why did people of yore camp on sandy beaches? Sand is comfortable to live on with very little preparation, is well-drained in case of rain or melting snow, easy on the feet, provides a good surface for beaching canoes, and is relatively free of vermin such as poisonous snakes, spiders and scorpions. However, it will not support seasonably occupied structures.
As aggradation slowed, beach deposits become finer and are eventually dominated by very fine sand and silt. Around 2,000 years ago, the floodplain at 45KT28, which was previously very narrow, widens so that is now being deposited on top of the earlier beach deposits, protecting them. It now consists almost entirely of silt, which is commonly referred to as floodplain loess. About six feet of this loess is deposited over the next ca. 1950 years, an average rate of just over 300 years per foot.
The loess floodplain is quickly occupied and used to construct pit houses, creating sites that are the central places in the winter village pattern described elsewhere in this report. This change coincides with increased site size, the density of cultural remains, technological styles, and the diversity of functional categories that make up tool assemblages.
Sometime relatively recently, the Snake and Columbia Rivers begin to erode floodplain margins, including middens and the remains of pit houses. This may be in response to industrial-scale water management in the Columbia Basin.
LAST REVISED: 31 MAR 2016