Semi-subterranean House Sites. Sites of ancient semi-subterranean winter houses, modern lodges and what may possibly have been a shell heap were seen and photographed by us in this region. Two of the examples of the remains of semi-subterranean house sites found here, as shown in Fig. 2, Plate IV, had stones on top of the surrounding embankments. Although on the top of the embankments of the remains of similar underground winter houses in the Thompson River region,[124] we saw no stones other than those of the soil. I am informed by Mr. James Teit that such are occasionally to be found there also, but that these stones are generally found only in those places where boulders were removed during the excavation for the houses. He was told that it was the custom to place these boulders around the base of the house. Two semi-subterranean winter house sites, as mentioned on pp. 7 and 15, may be seen on the flat along the north side of the Yakima River about a mile below the mouth of the Naches. One of these may be seen in Fig. 2, Plate III.[125] There are water-worn boulders in and on the embankments surrounding them. These boulders were probably uncovered during the excavation for the house. The holes are situated within twenty-five feet of the river and between it and the Yakima Ridge which rises by perpendicular cliffs, almost immediately behind these winter house sites. In fact, the photograph reproduced in the figure was taken from the hill side north of the pit and just up stream from the cliffs. They are on a little terrace about three feet high which gives them the appearance of having been connected by a ridge. The hole shown in the figure measured from the top of the ridge was nine feet deep. The top of the bank measured at points on the flat between it and the river, up stream from it, and between it and the hill, was four feet, two feet, and two feet, four inches, respectively. Averaging these measurements, the height of the embankment above the level is thirty-three and one third inches. The hole was so near the level of the river, and was so deep that when we visited it on June 18, 1903, which was during high water, the waters of the Yakima had soaked through the terrace and were about two feet deep in the bottom of the hole where it was about eight feet in diameter, measuring north and south. Measuring in the same direction the diameter of the top of the hole from points inside of the surrounding ridge was twenty-two feet, from points on top thirty-three feet, from points outside forty-seven feet, and from points outside of the wash from the ridge fifty-one feet. These measurements give us twelve and a half feet as an approximate width of the ridge or fourteen and a half feet if we measure from the bottom of the wash. The two sites mentioned on pp. 7 and 16 were also examined and photographed by us. One is plainly shown from the north of west in Fig. 2, Plate IV. They are located on a high terrace on the north side of the Naches River about one and a half miles above its mouth. There are angular rocks on each encircling ridge. Some of the large angular rocks found on the embankment of this ridge, may also have been dug out during the excavation for the house if such rocks are found under the surface of the soil in this terrace. Similar rocks are scattered about on the surface so thickly that it must have been necessary to remove a number of them from the site where the house was to stand and possibly others that were scattered about may have been put up around the base of the house in order to clear the immediate vicinity especially since many of them are disagreeably sharp angular fragments.[126]

Measuring the site best shown in the figure, east and west, the level floor inside the extreme wash from the ridge is nine feet in diameter, the rocks fallen from the ridge thirteen feet, the inner edge of the ridge 20 feet, the points on the top of the embankment, twenty-five and a half feet; the outside of the rocks, thirty feet; the extremes of the embankment thirty-five feet. These measurements north and south are respectively, nine feet, thirteen and a half feet, sixteen and a half feet, twenty-one feet, twenty-five and a half feet and thirty-three feet. Judging from these measurements, the original dimensions were probably thirty feet by twenty-five and a half feet over all, twenty-five and a half feet by twenty-one feet for the top of the embankment, twenty by sixteen and a half feet for the inside of the embankment and sixteen and a half feet by fifteen feet for the bottom of the floor. These measurements are also east and west and north and south respectively. The present depth of the hole below the top of the rocks is twenty-nine inches and from the top of the earth embankment is twenty-six and twenty-one inches. The measurements were taken east and west and north and south respectively. The slope of the hill from north to south and its attendant wash, of course, affect the north and south measurements, while the east and west measurements are probably near the original dimensions. Contiguous to this hole on the south, or in the sage brush to the right in the figure, is the other site. It is on the slope of the hill and not so clearly shown in the Plate. This hole measures ten and a half feet by eleven feet across the level floor inside; thirteen by fourteen feet inside of the rocks; nineteen by eighteen feet at the top of the embankment twenty-three by twenty-three feet outside of the rocks; and twenty-seven by twenty-six feet outside of the embankment; fourteen and eighteen inches in depth from the top of the rocks and ten and twelve inches from the top of the earth, the measurements being taken east and west and north and south respectively.

Mr. G. R. Shafer informed me that there were holes, the remains of old houses on the flat in the Naches Valley, twelve miles above the Nelson Bridge which crosses the river a short distance above the mouth of Cowiche Creek. At Fort Simcoe, immediately south of the Indian agency, on the north edge of "scab land" overlooking a small ravine as mentioned on p. 8, is a large pit surrounded by an embankment of earth, the remains of a winter house site. This hole is so deep and the embankment is so high that both Mrs. Lynch and the Indians call it a fort. About fifteen miles above Kennewick on the eastern side of the Columbia River, according to Mr. D. W. Owen, there were the remains of hundreds of semi-underground winter houses and we saw several large and deep sites immediately below Mr. Craig's house above Priest Rapids as mentioned on page 20.

A semi-subterranean winter house, with an entrance through the roof, seen by Lewis and Clark[127] on the north side of the Columbia near the mouth of White Salmon River, was uninhabited at that time (1805). As described, it does not differ from the winter house of the Thompson Indians. The Chinook, so far as we know, never erected such houses. The pit of an underground house, according to Clark[128] was found among the Nez Perce. Gibbs[129] mentions what were probably similar pits on the Lower Yakima. Kane[130] describes a somewhat similar house used by the Walla Walla but much ruder. Such houses were used by the Klamath.[131]

Not far from the ranch of Mr. Frank Turner on Rock Creek about six miles below Rock Lake on Section 6, Town 18 north, Range 40 east in the country locally known as "The Rocks," there are two pits that are supposed to be the remains of houses which with other remains (pp. 29, 82, 140) have been in their present condition since about 1874 when they were first seen by the whites. Both the pioneers and the old Indians are said to know nothing about them. Mr. Turner's place is best reached from Sprague on the Northern Pacific Railroad, although his Post Office is Winona. My information regarding these two pits is from Mr. J. S. Cotton, then in charge of cooperative range work in Washington.

It is quite possible as pointed out by Lewis[132] that the introduction of the buffalo skin covered lodge which probably came after the advent of the horse into this region, had something to do with the apparent scarcity of the semi-subterranean winter house in the Yakima region in historic times, the buffalo skin lodge possibly having taken the place of the earth-covered dwellings.

The so-called cremation circles near Cherry Creek and near the mouth of the Naches which were mentioned on pp. 12 and 15 and described on pp. 163 and 157, may be the remains of small houses of the type of semi-subterranean winter house sites that were made especially as grave houses. As before mentioned, this type of semi-subterranean circular lodge is found as far north as the Thompson River country, and I have seen one site on the prairie near Rochester, Thurston Co., probably of this type. In the Nez Perce region to the east, remains that appear like those of semi-subterranean houses consisting of ridges of earth about a foot above the general level of the ground, surrounding a circular pit, from three to five feet deep, measuring from the top of the ridge; and about seventy feet in diameter, are found near the mouth of Tammany Creek on the east bank of Snake River, a few miles above Lewiston. The site may be identified with Hasutin.[133] The place is known to have been used as a camp until about 1878, especially during the season of lamprey eel fishing. These house rings are in several groups. A little charcoal, some unio shell, flint chips, a digging stick with a bone handle, glass beads and other objects are reported to have been found in them. Somewhat similar house rings about twenty-five feet in diameter were found on the south bank of the Middle fork of Clearwater River, near the town of Kooskia. Spinden[134] refers to Lewis and Clark[135] for evidence of considerable antiquity for the circular house rings in this Nez Perce region. They mention one as being about thirty feet in diameter with a rim over three feet high and the floor sunken four feet below the surface of the ground or seven feet below the top of the rim. The Mountain Snakes, according to Ross[136] never used underground houses.

At the site near Kooskia there is another type of house site such as I have not seen in the Yakima, Thompson or Coast regions. Spinden describes them as long and narrow, about sixty to eighty-five feet long by eighteen feet wide. The interior is sunken from one to three feet and surrounded by well marked elevated rims. As a rule, these pits are not so deep or clearly marked as those of the circular type. The axis of the house is parallel with the river. He states that these house sites have not been used for a long time and that trees, some of which are eighteen inches in diameter grow directly out of them. Excavation revealed a number of fireplaces about twelve feet apart along the axis of these houses suggesting that they were communal lodges.[137] We discovered no indications of communal dwellings in the Yakima region.Circles of Stones (Summer House Sites). A circle of stones which marked a small lodge site was examined and photographed. The stones were no doubt cleared from the interior and all or part of these possibly with others, were no doubt used to hold down the lodge covers. Although I saw no such circle of stones in the Thompson River region I am informed by Mr. Teit that they are occasionally to be seen there and that they represent old lodge sites. The circle of stones above-mentioned as described on p. 15 was found on a terrace somewhat lower than the one on which were situated the remains of the two semi-subterranean houses described on p. 52 This terrace is a few yards down stream from the one on which they stand, and is separated from it by a small ravine. The site is a little further down the stream and towards the southeast. It is shown in Fig. 1, Plate IV,[138] from the point on the hillside a few feet above it to the north, shown on the lower end of the slope in Fig. 2, Plate IV and in negative nos. 44491, 5-4, and 44492, 5-5. This circle of stones on the level ground was made up of angular rocks such as are scattered on the immediate surface. It measures ten by eleven feet in diameter inside; fifteen by seventeen feet from the top of the circle; and twenty-two by twenty-three feet over all. The top of the highest stones was from fourteen to twelve inches above the middle of the space enclosed which as before stated, seemed to be on a level with the outside, all measuring being east to west and north to south respectively. Among the rocks was found a chipped piece of jasper or brown chalcedony.

No saucer-shaped depressions were seen in the Yakima region, although it is quite probable that they formerly existed and have been obliterated by weathering. It will be remembered that such saucer-shaped depressions are often made by sweeping out the summer lodges in the Thompson River region[139] and that they marked the sites of such houses.

Two summer lodges photographed[140] by us near Ellensburg which were mentioned on page 12 and the one seen below Union Gap down stream from Old Yakima, resemble those of the Thompson River region to the north. It will be remembered that mat covered tipis are found in the Nez Perce region to the east.[141] Lewis and Clark[142] mention but one buffalo skin lodge among the Nez Perce in 1806 and that was apparently reserved for special occasions, but a few years later this type of lodge had practically supplanted the mat lodge among that tribe and was in common use among all the interior Salish and Sahaptin tribes. The mat houses of the Yakima are mentioned by Gibbs in the Pacific Railroad Reports.[143]

A pile of stones shown in Fig. 2, Plate V[144] and mentioned on p. 20 as uncovered by the wash of the flood waters of the Columbia, was seen on the bottom-lands on the western side of the Columbia, south of Sentinal Bluffs and within a hundred feet north of the house of Mr. Britain Everette Craig. It is possible that this may have been a house hearth or ancient cooking place, although the presence of human bones among these stones, suggests that it was a grave covered with flat oval river pebbles. Near by, uncovered by the same wash, was a small patch of fresh water unio shells shown from the west of south in Fig. 1, Plate V.[145] This was probably kitchen refuse. The little pits, each encircled with a slight embankment made up of the soil thrown out in making it, p. 15, are probably the remains of food caches near the houses.