« Previous Region | Index | Next Region »

SW Region

Table of Contents:

Southwest Housing:

Iron Gate
Klamath Area
Lost Creek 35JA8 & JA12
Applegate 35JA47
Applegate 35JA42
Salt Caves
Lost & Elk Creek
Elk Creek 35JA59
Beatty Curve 35KL95
The Narrows 35DO53
Klamath Housing
Border Village 35KL16
Horseshoe 35DO400
Collier State Park
Crooked Creek 35Kl209
Soldier TS 35JO22
Fence Line 35KL969
Cole Project 35KL106
Maklaks City 35KL1111
North Bank 35DO61
[back to top]


"Semi-subterranean plank houses typified historic Takelma winter villages. The houses were large rectangular structures built of Sugar Pine planks split with elk-horn wedges and a stone mallet, placed vertically on the walls and horizontally on a slightly gabled roof. The four cornerposts and forked uprights supporting the ridgepole were lashed together with hazel fibers. The door was in the roof at the top of the dirt-bank access ramp or a notched log, and had a sliding door panel. The floor was excavated about two feet below the ground surface. There was a another house-type characteristic of the poorer class and captives/slaves: a circular house of bark slabs supported on poles leaned into a cone shape and banked with earth excavated from the interior. In addition, there were rectangular long houses for community meetings. Construction was similar to the family house. The sweat lodge was also of similar construction, but covered with earth to retain the heat." (Hopkins et al 1976:9-10)

"The winter habitations of the Takelma were described as rectangular in shape and semi-subterranean to a depth of between 18 and 24 inches. The framing consisted of upright posts in the corners, connected to cross beams to which exterior siding was attached. The siding was either split sugar pine planks (reserved for the wealthier class) or with slabs of bark (for the less fortunate). The roofs were of a single gable design covered with planks and sometimes with earth. A fire hearth was located in the center of the structure with a corresponding smoke hole in the center of the roof. Kroeber (1953:905) gives the exterior dimensions as being 12 feet wide and 15 to 20 feet long, with no interior partitions" (Gray 1997:7).

[back to top]


In Cressman's Klamath Prehistory, Klamath houses are described:

"Spier (1930:197-203) has described in detail the permanent or semi-subterranean house. It should be pointed out that only one of many houses we excavated during four seasons' work corresponds very closely to the house pattern informants gave Spier. However, Spier was describing the 'pattern' of a house and we were digging up houses which showed how in this, as every other pattern of behavior, the realization was an approximation of the 'pattern'... Following Spier's description we find the Klamath winter or permanent house consisting of a truncated cone super-structure set into and reaching up out of a circular pit dug to a depth of perhaps 2 feet. The diameter of the pits varies from 10 to 20 and more feet. There are some considerably larger than this. A large house pit on the F.F. McCready Ranch, near Kawumkan Springs, K1-9, excavated in 1951, was 16 m. from rim to rim and 9 m. across the floor. The floor in this house was 2 m. below the surface. Another, K1-12 (only tested), on the same ranch was 19 m. in diameter and 2.5 m. deep. These two houses, however, were unusually large and the growth of trees in the pits indicated a considerable antiquity for them."

"The main support of the house consisted of 4 posts each 12 inches in diameter set in a rectangular pattern and tied together at the top by 4 stringers, 2 of which at each end, the shorter ones, were put in place first and then on top of these 2 longer ones fastened on each of the long sides of the rectangle. The posts have forks about 9 to 10 feet above the ground floor. Four main rafters ran out obliquely from the corners of the rectangle and were fastened in the earth on the outside of the lip of the pit. These poles are split logs. Across each pair of these rafters at the top along the long axis a plank is fastened. Then a series of planks or poles is arranged continuously along each of these horizontal planks with the top resting on the plank and the other end resting on the surface of the ground outside the pit. A continuous series of rafters thus surrounds the house to provide the support for the roof covering. At the ends of the rectangle the poles rest their upper ends directly on the short cross stringers."

"The hatchway is now built by laying a series of poles lengthwise on each side of the long stringers and resting on the short tie pieces, thus narrowing the opening. Since the size of the rectangle formed by the posts will be dependent on the size of the house, it is difficult to give an approximate size for the hatchway. Spier (1930:199) states,:

"After the side rafters are in place the roof is covered with a series of mats laid horizontally in clapboard fashion. Grass is placed over the lower half of the mats and then the excavated dirt is thrown over the grass and mats to form a tight covering. The house thus slopes sharply from the surrounding surface to end in a truncated wedge-shaped top. Access is gained by a kind of combination of ladder-step device made of two small logs with branches projecting from the sides. These are laid close together on the roof so that the projecting branches from each meet the trunk of the other. The space between the cross branches is then packed with dirt to make steps."

"Houses are supposed to have the opening toward the southeast and so the ladder would be on that side. However, the orientation of the house may have depended on its relation to the river. Spier's informants insisted that the house was always oriented toward the southeast, claiming that this was because the prevailing winds came from the west" (Cressman 1956:390-391).

Cressman also published Sterns's notes on Klamath houses:

"The excavation itself is 1 to 4 feet deep... Within, four posts are set in intercardinal position, midway, radially from the center. In the groove in the top of each post a pair of poles will be set. paired poles span adjacent posts on each side; then paired poles resting on top of them bridge the other two sides. This forms a plat upon which rafters are laid, the outer ends resting about a foot beyond the edge of the excavation. Rafters always run outward from the post locations themselves, as well as from other portions of the plate. Horizontal rafters span the space within the plate. on top of this coarse tule mats are laid, with dirt above, to finish the house. A square or rectangular roof entrance-smoke hole is left. A ladder, in form said to resemble our rung-ladder is laid up the eastern side, and a ladder within rests against the eastern edge of the entrance and leads down. The Klamath, by their directional nomenclature, always conceive of the house as facing east... Tom denies that there was a bench or platform running around the inside of the wall. I always heard that the people slept and sat on the ground. The western part of the house was partitioned off by a wall of vertical poles laid against a horizontal beam. This room was used for storage. The eastern portion of the house behind the ladder was devoted to the storage of firewood and other miscellaneous objects... Against the wall, on the northern side, was the place reserved for the owner and his family. The old ladies, e.g. the owner's widowed mother, lived in the northwest corner by the wall of the storage room. Along the south side of the dwelling lived another family, often related" (Cressman 1956:391).

"The summer house was usually built with little or no pit and was generally a dome-shaped affair covered with mats. It would leave little archaeological record. Another type was also used. This was built on a rectangular floor plan above ground. Willow poles were set into the ground along the side and sloped at a sharp angle to a light ridge pole. along the sides were placed stringers to tie the sloping willows into a firm structure. Across the top and along the sides, starting from the bottom, were placed specially prepared tule mats as covering. These mats were sewed except at the edges where they were twined. The sewing prevented the rain being caught and soaking through as would have been the case with a twined weft construction. These mats were then placed on in clapboard fashion to facilitate drainage. The houses were sufficiently high to permit an adult to stand upright. Entrance was at one end. This kind of house would leave no direct archaeological evidence but only indirect in the form of tools that may have been abandoned" (Cressman 1956:392).

"Sweat houses were of two kinds, a round, dome-shaped structure used in summer and a more solid gable-shaped type used in winter. The gable-shaped variety is built over a shallow elliptical pit about 5 feet in length. The ridge pole is supported by a pair of struts at each end pushed into the edges of the pit. Sticks and bark are then leaned up against the ridge pole and this 'roof' is covered with dry grass and dirt. The ridge pole is about three and a half feet above ground. The small end entrance is covered with a mat" (Cressman 1956:392).

[back to top]


"Winter houses were rectangular in plan-view, and constructed over a pit up to three feet deep. Houses appear to have measured roughly 20 x 16 feet in size, with a steep roof banked with earth on the side-walls and covered with planks on the ends (Silver 1978:214). The hearth was centrally located on the floor; cooking utensils were stored at one end near the entry where acorn processing was also accomplished. Sleeping areas were located at the opposite end of the house, with work areas located in between. Doorways were usually oriented toward the river (Silver 1978:214; Voeglin 1942)" (Draper 1997:14).

"Assembly structures were also built at larger villages to facilitate meetings, guests, and novice shamans who danced there (Silver 1978:214). These structures measured roughly 20 x 27 ft in size, and were constructed over pits as deep as 6.5 feet. The walls and nearly flat roof were covered with earth, with plank side-walls, a single ridge pole, and a center post near which the hearth was placed (Silver 1978:214-215). A sweathouse for men were also constructed at large villages, while the assembly building served the same function among smaller groups (Holt 1946). The sweat-lodge was usually smaller than the assembly-house, and provided space for 15 to 20 males. Smaller, dome-shaped sweat-lodges, built with poles and skins, also functioned as menstruation huts, childbirth and mourning huts. Separate menstruation huts similar to the winter house were also constructed by Klamath River Shasta groups (Silver 1978:215). Structures at temporary camps in the summer and fall were constructed with pole framework covered with brush and bark " (Draper 1997:14-15).


[back to top]

Iron Gate

While the Iron Gate site is just over the border in California, the data is applicable to southwestern Oregon. The site contained thirteen house pit depressions from 6 to 10 meters in diameter and 20 to 50 cm deep in a row along the river. It was occupied between 1400-1600 A.D.

"The houses at the Iron Gate site differ from those used in the area during the ethnographic period. Shasta houses were basically rectangular plank structures, whereas the Iron Gate houses were circular, bark-covered structures" (Leonhardy 1967:4).

The Iron Gate houses had multiple floors indicating periodic re-use. There were central hearths lined with rock. In house 1, a mano, metate and hopper mortar were found in the NE quadrant of floor three. Floor two had a three mortars associated with the east side of the house. Floor one had a mano and metate associated with the SE quadrant of the house. House 2 contained a circular central fire-pit. A pestle and two hammerstones were found near the hearth. This house had three floors as well. House 3 also had three floors. House 4 had four floors. Floor one had a mano, metate and two pestles in the SE quadrant and a mortar at the edge of the north rim. It also had a cache pit in the NE near the rim and a second cache pit near the SW rim of the pit.

"Reconstructing the archaeological data, the houses at the Iron Gate site are described as follows: A conical framework of poles was erected over a circular pit 5 to 6 meters in diameter and 20 to 30 centimeters deep. large slabs of bark and perhaps planks or large splinters of wood were laid over the framework. Dirt was banked up against the sides of the house and rocks may have been set on the walls to hold the bark in place. Vertical posts may have been used occasionally to support the framework. No information about he entry to the house was recovered. House pits were reused, but when a new house was constructed over an abandoned pit, the debris and accumulated fill from the previous occupation were not removed." (Leonhardy 1967:12-13)

Iron gate house
Leonhardy 1967
[back to top]

Klamath Area: Long-Lalo Ranch

Cressman's work in the Klamath area contains good information on houses. He found that the tended to be circular and that the N-S axis tended to be longer than the E-W axis, but in a few location this trend was reversed. He found two size classes in his data, a large house 6-7 meters in diameter and smaller ones 3-4 meters across. He felt the smaller houses were ancillary to the larger and may have been kitchens, shelters for slaves or dependents (Cressman 1956:435-436).

His measurement for house pits were as follows (Cressman 1956:436):

N-S AxisE-W AxisDepth
7.5 meters7.5 meters1.5 meters
6.9 meters6.8 meters.6 meters
7.5 meters6.5 meters.6 meters
4.2 meters3.0 meters.6 meters
9.6 meters9.0 meters.6 meters
2.4 meters2.6 meters.6 meters
18.0 meters16.0 meters2.0 meters
6.0 meters6.0 meters.4 meters
2.5 meters2.5 meters.4 meters
19.0 meters17.0 meters2.6 meters
Long-Lalo ranch Village
Cressman 1956:434

Cressman (1956:437-441) found that post molds were marked by their absence. One house had a single post mold near the center of the house suggesting a single post support. Two houses had double post molds for possible roof supports. One house had seven post molds, four could have been roof supports and two could have been for the ladder, but he felt they were additional support beams. One house had six posts and came closest to the described ethnographic pattern. The houses tended to have a marked bench between the main pit and the outer house walls. The benches were made by constricting the circumference of the pit and then continuing the excavations for the main saucer-shaped depression. Only one house showed clear re-use. Two houses had a bark floor covering.

Lalo 1 and Lalo 2 Houses
Cressman 1956:438
Long 4 House
Cressman 1956:437

"The houses in the Klamath village of immediate pre-contact time and from earlier periods do not show any pattern of placement with reference to one another. The villages are usually situated on a terrace or bench above the river, possibly to avoid flood damage or perhaps because there the soil was soft and deep enough to permit digging foundations with their limited tools. Frequently the villages (winter) were close to some feature of the river which guaranteed the possibility of fishing during the winter season when there was likelihood of the river freezing. Riffles, springs, or the confluence of a feeder stream provided a good chance that the river would stay open. The pattern of the structure of the houses shows a generally circular floor plan although the north-south axis is usually a little longer than the other. Floors are the shape of a shallow saucer. Most of the houses have a bench or shelf set a shot distance down from the surface which in most cases runs around the houses. Post holes occur only in a few houses. Six of the houses were destroyed by fire, perhaps set accidentally or else burned after the death of an occupant. The evidence for entrances is confusing with some houses suggesting a ladder entrance through the roof hatchway and at least one suggesting a side entrance" (Cressman 1956:445).

Photo Klamath Pit House
Cressman 1981:4
[back to top]

Lost Creek Project 35JA12

In the Lost Creek project Davis (1968:17) noted that sites "JA8 and JA12 lacked evidence of permanent aboveground structures; they may have had temporary structures for summer use. The floor plan is oval and does not resemble the rectangular plan of the planked houses built by the Takelma proper ... The oval floor plan and distribution of lithic materials closely resemble the Irongate houses... but the Lost Creek pits are not as deep and well-defined... The Lost Creek housepits appear to substantiate Leonhardy's (1967:34) guess that the distribution of conical, bark covered house extended north to the Rogue River".

In another report (Davis 1974:43) indicates that ten house depressions were recorded at 35JA23, of which half had been potted. test units were dug into the pot holes, but no features discovered in the disturbed ground.

[back to top]

Elk Creek Project

The 1979 Elk Creek report by Brauner and Honey reported on testing several sites. 35JA11 had been recorded by Cole in 1966 and tested by Davis in 1967, and had visible housepits. By 1979, looters had hit the site so heavily that the depressions were no longer visible. A single test pit was dug between two houses. At 35JA29, they looked for six house depressions but could only find tree throws. A test pit at 35JA100 was interpreted as hitting the edge of a house pit, suggesting buried houses in a long plowed and leveled field. Site 103 had two large muted depressions about 5 meters in diameter and three meters apart. A test pit encountered FCR and a possible house rim confirming a house pit (Brauner & Honey 1979).

Brauner and Lebow summarized the data from Elk creek in 1983. A larger area was excavated at 35JA100 to confirm the house floors. In the process, they recovered crude pottery and figurines. They found two or perhaps three houses superimposed above each other. The houses had been dug about 80 cm below the ground surface. They were six meters in diameter with an unknown superstructure. Some possible posts were found along the west rim. They postulated that a winter village settlement pattern with associated semi-subterranean houses began about 1500-2000 years ago in the Elk Creek drainage (Brauner & Lebow 1983:177).

[back to top]

Applegate Project 35JA47

At the Applegate Lake project, initial surveys recorded a series of sites with houses. 35JA42 was described as containing 5-6 circular house pit depressions in a single row along the river. The depressions ranged between 3-4 meters across and averaged 40 cm in depth (Brauner and Honey 1978:38).

Later at the nearby Applegate project, OSU was excavating houses at 35JA47. An earlier test pit had transected the eastern rim of a house, so excavations were expanded into a 10x10 meter block. Three housepits were encountered and one fully excavated. In another block a possible sweat house was also completely excavated.

House 1
Brauner & MacDonald 1983:87

"House I was circular in plan view with a diameter of 5 meters. The central floor area was excavated 70-80 centimeters below the original ground surface. The floor sloped up toward the rather low angle walls (ca 40° wall slope). The overall appearance would be described as saucer shaped. There was no centrally prepared hearth within the structure. Charcoal staining on the floor was most intense on the central floor area, but fire-cracked rock was abundant. Whether this represented an unlined central hearth or an accumulation of debris on the lowest portion of the floor is uncertain, but the evidence tends to indicate a hearth. There were no post holes around the margins of the structure or penetrating the floor area. Several hopper-mortar bases were situated is a circular pattern on the floor 1 meter from the walls and 1.5 meters from each other. Because the utilized surface of several of these mortar bases was facing downward, these large stones may have functioned as base plates for roof support poles.... A pavement of flat river cobbles along the northeast rim of the house may be related to the entry way. The feature was composed of seven discoidal river cobbles arranged to form an ovoid pavement. The feature was situated on the gently sloping house rim 30 centimeters above the house floor and 30 centimeters below the lip of the housepit" (Brauner & MacDonald 1983:46).

Based on their analysis, the house contained a lithic manufacturing area in the NE, sleeping and storage area in the NW and SW areas, and two food preparation areas, the primary one in the SE and a secondary one in the west.

House 2
Brauner & MacDonald 1983:48
House 2 interiors
Brauner & MacDonald 1983:61

House II was ovoid and 70 cm deep from the original land surface. A 3 meter strip of this house was excavated. The tests opened up the east and NE section of the house, and the artifacts suggested food preparation and lithic manufacturing similar to the other house. Feature o-1 was interpreted to be a sweat lodge. It was full of FCR and was 4 meters in diameter. This possible structure was surrounded by a general work area to the NE, a rock heating area to the SE, a work area to the S, a possible longing area to the SW and a possible entry area to the NW.

"Several similarities are evident in the space-use patterns employed in House I and the sweathouse. First, in each structure the proposed doorway faces a northerly direction, to the northeast in House I and northwest in the sweathouse. Second, the principal manufacturing area in both houses is located adjacent to the doorway, to its left in House I and to the right in the sweathouse...Third, the sleeping area in both houses is situated to the right of the doorway" (Brauner & MacDonald 1983:130).

[back to top]

Applegate 35JA42

Brauner's work at 35JA42, a small village of four houses in a line along the Applegate River with a small structure (menstrual hut?), indicates these houses were occupied one at a time, and for several seasons before being rebuilt at a slightly different location. House 1 was about 4 meters in diameter with a saucer shaped depression. A central hearth was visible as a ring of large discoidal river cobbles placed on end. No post holes were found, but charred Douglas fir wood and bark was recovered. Based on FCR patterns, stone-boiling was done west of the central hearth (1983:23), points were clustered to the west and northeast (1983:29) suggesting weapon manufacturing and repair was done on the northern floor, pestles and metates were found on the west side of the house floor (1983:31). "The primary task area within the House 1 was the northwest quadrant. based on the distribution of fire-cracked rock, bone fragments, pestles, and grinding stones, this was the primary food processing area. Food processing activities spilled into the southwest and northeast floor areas as well. The entire northern half of the house and secondarily the southwest floor area were general work areas. tasks performed included tool manufacture and repair, hide working, and possibly woodworking. The relative absence of debris in the southwest quadrant may indicate a storage, sleeping, longing area which are all non-debris producing activities" (Brauner 1983:34).

Map of Houses
Brauner 1983:16

House 2 was 5.5 meters by 25 cm deep. There was a prepared hearth 40 cm in diameter in the center of the saucer shaped housepit. Again, no post holes were found. "The northwest floor area was definitely the primary work area within House 2. Most of the food-processing activities associated with the house were accomplished here. A secondary floor preparation area was situated northeast of the hearth. Other activities (such as hide working, wood working, lithic tool manufacture and repair) were accomplished across the entire northern floor area.... The floor area south of the hearth was almost devoid of cultural material... As with House 1, this may have been a storage, lounging, sleeping area" (Brauner 1983:52).

House 3 was just over 4 meters, but without a defined central fire hearth. This house had been burned and the lower 10-20 cm of wall planks preserved by charring and because earth had been banked against them. "In plan view, House 3 was probably square. The north and south walls were composed of parallel planks set vertically. The walls appeared to be straight. The east and west walls were not as intact as the north and south walls, but the construction detail (i.e., vertical planks) was the same. The east and west walls were also parallel. The wall intersections approached 90°. The floor area, as described by the remnant walls, was slightly under 4 x 4 meters. The hearth was centrally positioned on the floor."

"The wall planks were set up to 30 cm in width and averaged about 6 cm in thickness. Most of the planks were incense cedar with bark remaining on one side. Two planks were Douglas fir with the bark remaining on one side. The bark side of the planks faced outside. Two small vertical posts were preserved. One post was situated near the southern end of the east wall. The other post was centrally positioned along the west wall. Both posts were Douglas fir. Ponderosa pine bark fragments were also associated with the house, but how, or if, they relate to the superstructure is unknown" (Brauner 1983: 56-59).

"The primary food processing area was confined to the western floor area. Weapon storage may have been confined to the eastern half of the house. The entire floor area could be termed general work space. No apparent sleeping, storage, or lounging area was noted, although a relatively low frequency cultural debris area was noted south and west of the hearth" (Brauner 1983:70).

House 5 was a steep-walled depression, and the steep walls were from looting. Excavation was stopped when the extent of looting was defined.

House 1
Brauner 1983:22

House 6 was a 2x3 meter depression about 20 cm deep. It turned out to be a small oval area dug into alluvial gravel about 30 cm. The floor space was 4.3 by 2.8 meters and it contained no hearth. "Activities and tools usually associated with a male presence in the structure were absent. The total cultural assemblage from House 6 indicates short-term usage of the structure by women, The content, size, and isolated position of the structure leads this researcher to propose that House 6 functioned as a menstrual hut" (Brauner 1983:82).

"Although four houses were present in the village complex, they were not occupied at the same time. The backdirt from the construction of House 1 was thrown into the House 2 depression. House 2 was therefore abandoned when House 1 was built and occupied. Cultural debris within the House 3 depression was smeared and blended more than houses 1 and 2, and some elements (i.e., hearthstones) were missing. Although tenuous, the conclusion might be drawn that House 3 laid open, exposed to scavenging by subsequent occupants of the site, longer than houses 1 or 2" (Brauner 1983:86).

House 3 Wood Elements
Brauner 1983:57

Brauner postulated that the houses represented four successive winters by the same group building in a new location each winter. The biggest surprise was the square house built over a circular saucer depression. If it had not been for the burning of the one house, Brauner would have assumed a quite different house superstructure. Based on the evidence, he feels all of the houses were square and placed over a round shallow depression. The material culture supports a proto-historic date for occupation (Brauner 1983:87).

[back to top]

Salt Caves

The Salt Caves project was surveyed by Elliott Gehr, and initial results described in 1984. Site 35KL16 had been recorded earlier and contained 23 housepits in three clusters aligned along the river. The 84 survey identified five additional housepits. 35KL17 contained two probable housepits and a third possible one. 35KL18 (Baby Rock Village) contained 41 housepits, 5-6 meters in diameter in rows along the river bank. 35KL20 (Klamath Shoals Village) had 11 housepits. 35KL22 (West Bank Pine Village) included seven depressions recorded in 1961. 35KL23 (Crayfish Creek Village) had three depressions 4-5 meters in diameter and a small depressions. 35KL25 (B-6 Village) with 11 depressions also contained five cache pits. 35KL26 (Powerhouse Overlook Village) had nine depressions from 3 to 5 meters across and about 50 cm deep. Additional sites included Flume View (4 depressions), Butler's Lookout Point (1 depression), Salt Cave Lookout (1 depression), Feather Flats (1 depression), and Ponderosa Estates (10 depressions in two clusters).

Joanne Mack summarized the excavation data form the 1960's University of Oregon work at Salt Caves. At Border Village (35KL16) one houses was excavated and two were tested of 19 depressions in two irregular rows along the Klamath River. Depression 1 was eight meters across and 1.5 meters deep. The crew found four floors. The maximum floor-4 diameter was 6 meters with a depth ranging from 7 to 27 cm. There was a central firepits about a meter in diameter. The house had a clay rim, and charred beam fragments were recovered. The presence of a cremation burial suggested that the house had been burned at the death of an occupant. Floor-3 was 6.5 meters in diameter and 10 to 40 cm deep. It also had a central firepit. The house probably had a bench marked by a clay rim, and it had burned. Floor-2 was 6.5 meters wide and 23-53 cm deep. It had the same central firepit, similar bench and rim, and had burned. Floor-1 was 6.5 meters by 15-67 cm deep. "The arrangement of the wood indicated that the house, at the time this floor was in use, was built with supporting posts near the edges and in the center, with slabs or beams laid from the house edge to the center. When the house collapsed most of the wood fell inward. The vertical wood near the perimeter may indicate the interior walls were also lined with wooden slabs or planks" (Mack 1983: 33-38).

35Kl16 House 1 Floor 1
mack 1983:40

Depression 2 had three superimposed floors. The houses ranged from 5.5 to 7 meters in diameter and from 10 to 85 cm deep. "The placement of the posts and fallen wood indicated a house with support posts near its edges and beams or planks laid from the house rim to the center of the house" (Mack 1983: 41).

Depression 16 had two floors, 5.5 and 6 meters across and 10-95 cm deep. The placement of two post holes suggested the possibility of both a central group of posts as well as an outer ring of posts.

"Several similarities were observed between the three excavated house pits of Border Village. All of the house pits were: 1) multifloored; 2) oval in plain (SIC) view; 3) saucer-shaped in profile; and 4) burned at least once. Each house and usually each floor had: 1) a central firepit; 2) milling stone and mullers on the floor; 3) a superstructure of wood; 4) rocks and cobbles concentrated adjacent to the rim; 5) a clay cap or overlay over part of the rim; and 6) post holes and posts around the floor edge. A few floors had one post or post hole near the center of the floor" (Mack 1983:44).

Big Boulder Village (35KL18) contained 14 houses in three irregular rows along the river. Three houses were excavated. House-3 was a 5 meter by 50 cm depression. It contained a single house floor 5.2 meters ranging from 40-78 cm deep. It contained a central fire pit and a secondary one in the NE section. There was one central post and four around the perimeter. House-11 depression was 5 meters across and 55 cm deep and ended up 4.1 across and 40-50 cm deep. A large central fire pit was recorded. Mack noted that the post hole patters suggested a central post and an outer ring of posts. There was a possible bench to the SE. House-13 depression was 5.5 meters and 45 cm deep. The excavation suggested the house was 4.5 meters across at floor level and 20-75 cm deep. It had been burned and :Most of the timbers laid with their long axis toward the center of the house. Some timbers were at right angles to others and may have been tie beams" (Mack 1983:47-56).

35Kl18 House 13
Mack 1983:53

"The three excavated house pits of Big Boulder Village have several characteristics in common: 1) a single floor or occupation level; 2) a central firepit; 3) cache pit adjacent to the outer rim; 4) a ring of cobbles or stones around the rim; 5) a superstructure of wood, which was usually incense cedar, although some pine was used; 6) possible entrances on the east side; 7) partial benches or shelves between the floor and the outer rim; 8) hopper mortar bases or grinding slabs embedded in the floors; and 9) an oval plan view and a saucer-shaped profile" (Mack 1983:56).

Mack noted that depressions ranged from 3.5 meters at 35KL20 to 8 meters at Border Village. Mack speculated that Big Boulder was Modoc while Border Village was Shasta or Takelma (Mack 1983: 217-219). She felt the area was a fringe area in historic times between the traditional territories of groups to the NE in Oregon and SW towards California.

In a report on survey work at Salt Caves in 1984, Elliott Gehr updated site forms for a series of sites containing possible house depressions with a shamanistic interpretation of function by Anaraiko, a Shasta medicine woman as well as other Shasta informants. Site 35KL16 (Border Village) contained 23 depressions in three clusters. Cluster one contained 3 depressions that Anaraiko felt was for salmon-callers to invite fish upstream. Cluster two with 18 depressions, was interpreted as a village. Cluster 3 with 2 depressions was for the salmon-callers to send the fish to spawn. The depressions averaged 5 meters in diameter with several larger ones 7 and 8 meters and from 25 to 100 centimeters deep. Site 35KL17 had three depressions 4-5 meters across and very shallow.

Site 35KL18 (Big Boulder Village) had 41 depressions (also known as Baby Rock Village by the shaman because a "rain calling" rock was found there). The depressions were 4-6 meters across in rows along the river. The map of the site shows irregular clusters of pits, some of which form rough irregular rows. Site 35KL20 (Klamath Shoals Village) contained 13 depressions, eight of which were 4-5 meters across and 50 cm deep, the other pits were smaller and of unknown function. The depressions were in an arc along the edge of the terrace. Site 35KL22 (West Bank Pine Village) contained 7 depressions from 4-6 meters across and 50 cm deep arranged in a row.

Site 35KL23 (was given the name Crayfish Creek Portal) contained 4 depressions that the shaman interpreted to be house for ritual entry and exit for visitors to the canyon. The pits ranged from 3-5 meters across. Site 35KL25 (B-6 Village) had 10 depressions 3-8 meters across and up to 75 cm deep.

Site 35KL26 (given the name Men's Hunting Sweat) contained 8 depressions 4-5 meters across and 30-50 cm deep. The shaman said the village was used for ritual sweats for hunting. Site 35KL550 contained 3 possible depressions from 3-5 meters across and 30 cm deep. Ms Vanderploeg interpreted 35KL551 (named Butler's Lookout) to be a shamans medicine area and the 3 meter and 50 cm deep depressions as a house where medicine men went to isolate themselves and meditate. Site 35KL552 contained 3 depressions 3-6 meters across and 30 cm deep.

Site 35KL553 was a single depression (renamed Salt Cave Lookout site) interpreted as a hunting lookout. Site 35KL558 (Feather Flats) was a single depression 6.5 meters and 100 cm deep, interpreted as a medicine area. Site 35KL576 (renamed Women's Ceremonial Area) consisted of two clusters of depressions, six to the north and two to the south. It was seen as a female sweat area for ceremonies (Gehr ND).

[back to top]

Lost & Elk Creeks

Davis (1983), in his final reports on the work at Lost and Elk Creeks, described the testing results from 35JA27 (Johan II). Three depressions had been recorded as possible housepits. The inner rim of depressions 2 was 3.6 by 3 meters. The crew digging this depressions reported finding a house floor, but it did not correspond to any of their filed notes on soils or features. Davis felt the evidence was inconclusive. Depression 1 was felt not to be a housepit, but rather a mix of looter pits and erosional features. Depression 3 was 3.8 by 4 meters. After testing, it was determined that it was a looter pit (Davis 1983:99-101)

[back to top]

Elk Creek 35JA59

During excavations of 35JA59 at Elk Creek, Mountain Research dug a semi-subterranean housepit dating between 740-920 years old and noted another possible house as feature 4. Feature 1 was a housepit a maximum of 1 meter below the original surface and about 4.6 meters across. It appeared to be oval with its long axis oriented northeast-southwest. A hearth was found offset from the center to the northeast. The house was exposed in profile, and no post holes were seen within the very limited exposure (Budy, Drews & Elston 1986).

[back to top]

Beatty Curve 35KL95

During testing at site 35KL95 (Beatty Curve), Musil (1987:24) suggested that a house pit or other large feature was seen in the profile of the excavations in test pit 1. The feature had a layer of shell above a layer of rock, both suggesting house floors. The pit feature was about 60 cm deep and at least two meters across (1x2 meter profile).

[back to top]

The Narrows 35DO153

O'Neill (1987: 7) reported on a test pit that hit a possible housepit in 35DO153 (The Narrows), where a C14 date of 1020±60 BP was associated with the house floor. The profile suggests a 60 cm deep house with a 50 cm wide bench about 20 cm deep at its margin.

[back to top]

Lebow Summary

In 1988, Clay Lebow gave a paper summarizing the house features in southwestern Oregon at the 41st NW Anthropology Conference. "In profile, all the house depressions excavated by INFOTEC were bowl-shaped, with steep walls, and sloping, but relatively flat floors... The floors were defined by lenses, no more than 2 cm thick, of soils more compact than the covering fill, and with relatively higher densities of charcoal, faunal remains, and fire-cracked rock. The house walls had the same characteristics as house floors, and were defined as that portion of the feature surrounding the floor in which the rise was greater than the run. Depths from the outside surface to the lowest point of the floor ranged from 38 to 56 cm. In plan view, four of the houses were roughly circular, while the lower of the two super-imposed houses at 35JA100 was roughly rectangular; it had two parallel sides, and the ends roughly convex. The circular houses ranged from 3.8 to 4.8 m in diameter, while the rectangular house measured 6.2 by 4.3 m. The floors of the houses at 35JA100 appear to have been artificially smoothed during house construction by packing fine-grained sediments over the freshly excavated course-grained in-situ sediments; one of the floors appears to have been created by packing and then firing clay on the freshly excavated surface" (Lebow 1988: 1-2).

"All the houses at both sites had centrally located hearths, which averaged 108 cm in diameter and 45 cm deep; all of these hearths were pits excavated into the house floor. Only one hearth also included a ring of rocks. All of these hearths contained large quantities of fire-cracked rock, and the sides and bottoms of the pits were baked. The rectangular house had a second, smaller, circular pit located 30 cm from the hearth, which was thought to be a second, smaller hearth, as it also contained fire-cracked rock and baked earth. Non-hearth pits were located in the floor of two of the houses; in fact, one house at 35JA100 had seven such pits.... Another feature on a house floor was a raised earthen platform in the northwest corner on the floor of the rectangular house; this bench was approximately two meters long and one meter wide, and was elevated 10 cm above the floor" (Lebow 1988: 2-3).

"Among the features recovered within the house depressions were charred posts, post molds, and charred bark slabs lying on the floors, all of which provide insights into the superstructures over the house depressions. The remains of two posts were recovered, with the ends butted up against or set slightly into the wall of the house; alluvial cobbles were piled around the base of the post, and remains of the posts indicated that the poles pointed towards the center of the house. The two poles were identified as Douglas-fir and maple, two species of tree readily available in the are today. Five small, shallow, circular depressions (four of which were in the walls of the rectangular house) were examined, and were inferred to be post molds. Like the posts, all were located in the house walls, and all encompassed by large alluvial cobbles. Numerous fragments of burned bark were found on the house floors. It appears, therefore, that these prehistoric houses were pole structures covered with bark. There was some evidence that at least the lower portions of the outer bark covering was blanketed with dirt -- probably with earth removed during excavation of the house depressions. We found no evidence of doors or entry-ways" (Lebow 1988: 4-5).

"One of the interesting differences between the sites with house features in the upper Rogue River and those in the upper Klamath River is the number of known houses per site. Based on the Irongate Site and sites in the Salt Caves area, village sites on the upper Klamath River contain from 2 to 19 houses, with an average of 11 houses per sites (Leonhardy 1967; Mack 1983). On the other hand, the maximum number of houses in a single site in the upper Rogue River is 8, with most sites containing far fewer" (Lebow 1988: 6).

House 1 at 35Ja100
Pettigrew & Lebow 1987:26
[back to top]


In their 1987 report, Pettigrew & Lebow describe House 1 at 35JA100 as 4.6 by 4 meters. The profile of the house is shown above. The relatively steep side walls, yet shallow and relatively flat bowl shape is apparent in the profile. Features 29, 32 and 35 are the hearths and feature 37 is the packed and fired section of floor. It dates from AD 1300-1500.

House 2 35JA100
Pettigrew & Lebow 1987:37

The above profile shows the two house floors in House 2. The house was between 4.5 and 4 meters in diameter. Again the house profile is bowl-shaped with relatively steep walls on one side and relatively flat floor. Feature 2 and 22 are central hearths. The two house floors are shown in the profile drawing. It dates between AD 1200-1700.

Photo of House 4
Pettigrew & Lebow:1987 55

House 4 was 4.8 by 5.2 meters in diameter. It also had relatively steep side walls and a bowl-shaped floor with a central fire pit as well as a number of storage pits excavated into the floor that show up as the dark areas in the photo. It dates between AD 1200-1500.

"The prehistoric houses excavated at JA100 and JA59 were all semi-subterranean, with the depressions circular to rectangular in plan view. The circular depressions ranged from 3.8 to 4.8 m in diameter, while the single rectangular depression was 6.2 x 4.3 m; all were excavated from 38 to 56 cm below the original ground surface (the surface present when the house was constructed), with an average depth of 49 cm. In profile, the house depressions appeared bowl-shaped, with relatively steep walls, and a floor gently sloping towards the center. During house construction, the floors appear to have been smoothed by packing fine-grained sediments (clayey soil) over the freshly excavated surface, and, in at least one instance, firing the clay. All the houses had hearths located near the center; these hearths, only one of which was rock lined, were pits excavated into the floor. Three of the houses had small pits (of unknown function) in the floor or walls, including one house (House Site 4 at JA100) with seven small pits. The rectangular house (the upper house at House Site 2 at JA100), had a 1 m wide and 2 m long bench along the western wall. All of the houses had large flat alluvial cobbles setting on the floors; these may have served as furniture, implements for household chores, or as platters (Holt [1944:309] notes that the Shasta placed deer heads and other bones with little meat adhering on flat stones near the fire, where anyone could help themselves)" (Pettigrew & Lebow 1987 11.54-55).

"The superstructure over the depression appears to have been constructed with poles covered by bark. The poles were butted up against or set slightly into the wall or the base of the wall around the periphery of the house, and rocks were piled around the base of the poles for support; the poles pointed towards the center of the house, indicating a conical structure (although this is conjectural, as the remains of only two poles were observed). There is no evidence of a central pole or poles. The poles, which included Douglas-fir and maple, were covered with pine bark, and earth (possibly what was removed during excavation of the depression) was piled around at least the base of the structure; it is possible that earth was used to cover the entire structure" (Pettigrew & Lebow 1987: 11.55).

[back to top]

Border Village 35KL16

In 1987, additional testing was done at 35KL16 (Border Village) for the Salt caves Project by Jensen & Associates. The test of house 14 indicated that "no discrete housepit or occupational 'floors' were discernible, and it seems reasonable to assume that the deposit represents a more or less continual succession of occupation which was accompanied by the gradual accumulation of cultural material within the confines of the original structure's perimeter. There was no evidence for any catastrophic events (fire) or major episodes of 'rebuilding'.... Features such as hearths, rock, bone and artifact concentrations were observed at various depths and locations throughout this deposit, indicating that things had indeed been 'moved around' inside of the structure over time" (Jensen 1987:3.21). No post holes or superstructure was encountered. While four strata were identified, and two major episodes of occupation, the report disclaims these were occupational floors in the house. The house was excavated 10-20 cm into a yellow clay (the layers totaled about 50 cm of depth. The house was circular and basin shaped (6.5 meters in diameter).

House 19 was also tested. It was a 5 meter depression 27 cm deep. After the first layer was dug, the excavations were terminated by Jensen & Associates as it was felt that the data was redundant. The exposed floor was refilled with 3-5 cm of soil (Jensen 1987 3.77). The depression at 35KL551 was tested and found not to be a structure.

[back to top]

Horseshoe Site 35DO400

In a testing report for 35DO400 (Horseshoe Site), Spencer (1987) speculated that a two rock features may be on the edge of a possible house depression: "It is felt that this rock cluster potentially represents the western boundary of a culturally produced depression and that this depression may have been a house" (Spencer 1987:23).

[back to top]

Collier State Park

Dave Cole wrote up some field notes in 1965 on excavation of a house in Collier State Park including letters describing what he found. This material is stored as SHPO report 9661. He and Harvey Rice surveyed the Park in 1965 and recorded two large depressions measuring 10 and 12 meters in diameter and 80 cm deep. Another six houses ranging from 6 to 11 meters were recorded in a different location. two of the depressions were deep (1.2 and 1.5 meters) and the others shallow. A number of shallow depressions were noted nearby (no description or numbers) as well as two shallow and one deep depression 10 meters across and 1.2 meters deep. This house was dug and two occupations found at 35 and 80 cm from the surface (Cole 1965).

In an attached letter dating 1969, Cole described the house (a drawing is missing): "House Pit #1 was the first occupied house, which, we suspect was covered with timbers radiating out from the center, like spokes on a wheel. The Fire Place in the center is associated with the floor of House Pit #1. Also note that the east wall of the central depression of House Pit #1 was mostly eliminated with the excavation of House Pit #2. The truncated remnant of the east rim of House Pit #1 is seen approximately 20 cm west of the later House Pit #2 east rim... House Pit #2 was a later house which apparently had the east and south walls outside wall in common with House Pit #1. The west outside wall however, was nearly two meters east or inside of the existing wall of House Pit #1. The central depression of House Pit #2 likewise was further inside the present rim on the west side. The floor of the remnant of House Pit #2 was 3-6 cm higher than the floor of House Pit #1, aligned with the floor of the remnant of House Pit #2 on the east side of the existing pit... The structural remains of House Pit #2 were found in excavation. This house was covered with large logs laying parallel, straight across the house from rim to rim, and oriented E-W. Bark had covered the logs then soil had been placed upon the bark. We were unable to determine the location of the entrance of House Pit #2, but assume that it was central.... To reconstruct the house I would suggest constructing a frame with four upright logs as center posts with supporting beams in a square to support ceiling beams extending in from the sides. As far as I could determine, there were twelve such beams. Seven of the beams would have been used on a little more than the northern half of the pit. That is the number I would recommend. I would suggest that the center posts be about five feet apart with a five foot square frame on top. Logs twenty feet long and at least eight inches in diameter, including bark, at the smallest end, should be used as roof beams. They should be situated, small end in, over the places on the rim where the yellowish or gray pumice extends fartherest towards the center of the pit.... In the center of the pit let the ceiling beams extend over the supporting beams far enough to from a three foot diameter opening ... Over the ceiling beams I would place branches, then bark, then cover with soil" (Cole 1965: letter).

[back to top]

Crooked Creek 35KL209

Elizabeth Budy (1990) re-evaluated site 35KL209 during the Crooked Creek project. The site had been recorded in 1979 as a village with 34 depressions. Budy had the depressions measured. They ranged from 1.6 to 7.2 meters in diameter with an average (excluding the single large pit) of 2.9 meters. She felt the pits were too small to be winter housepit depressions, but could be summer lodges or sweat-lodges. No FCR or prehistoric artifacts were directly associated with the pits. Instead, a thin scatter of historic artifacts was found. She speculated the site may have been a post-contact Klamath summer camp (Budy 1990:12) associated with the nearby Fort Klamath.

[back to top]

Soldier Timber Sale

Boyd recorded site 35JO27 with 7 depressions during the Soldier Timber Sale. The depressions were 5-6' in diameter and 6-12 inches deep. Only one flake was recorded with the depressions. Based on the site form, measurements they appear to be tree-throws ... but were in an Alder grove (Boyd 1990).

In 1991, Budy recorded another possible housepit site. The depression was 7.2 meters across and 58 cm deep. It had a steep-sided bowl shape. A few lithics and some FCR were recorded nearby.

[back to top]

Fenceline Survey 35KL969

In a 1991 range fenceline survey, Sobel recorded a site with three rock rings. It was recorded as site 35KL969. Two of the features had been vandalized and were at the base of rimrock, and the third (also vandalized) was out in the open. "Features #1 and #2, stone enclosures, have been highly disturbed. It is now impossible to determine original form of features. They appear to have been enclosed of cleared spaces, along rimrock base, with high concentrations of artifacts. Rimrock encloses each space on the east side. On other sides, enclosures are defined by piles/walls of naturally occurring stones and stones placed there by vandals" (Sobel 1991: site form). Feature #1 was 40 by 12 feet, feature #2 was 13 by 11 foot and feature #3 was not measured on the site form, but the sketch map suggests about 18 feet in diameter.

[back to top]

Cole Project 35KL1062

George Knight (1992) for the Cole project on the Chiloquin Ranger District, recorded site 35KL1062, a village with between 25 and 48 depressions (numbers vary with resurvey). No measurments were taken nor a map prepared of their locations within the site. This may be the village named Gpgá si noted by Spier in 1930. Another site, 35KL1066 was single depression with a berm of backdirt. Not measured or drawn. Site 35KL1067 is listed with 5 depressions ranging from 5-6 feet in diameter, circular with slight berms.

[back to top]

Maklaks City 35KL1111

David Ball (1993) recorded site 35KL1111 (Maklaks City) during the Odell Creek Fish Enhancement Project. The site had 29 depressions, no measurements given on the site form, but a map to scale suggests they ranged from 2 to 8 meters across. Fifteen of the depressions clustered near the largest house in the NE, smaller groupings of 2-3 houses were strung out to the SW.

[back to top]


Budy, McNair & Gentry (1993) recorded a site with a single depression 6 meters across and 70 cm deep with a nearby possible cache feature 2 meters long and 60 cm deep for site 35KL1125. In another project, Budy recorded site 35KL1122 with a two depressions. "one large pit (12 meters in diameter) and 1 average housepit-sized feature (6 meters in diameter) are situated on the top of the river bank in a manner similar to 'housepit-like' features recorded elsewhere along this stretch of the river" (Budy 1993: site form). The large one was a meter deep and the other 40 cm deep.

Anan Raymond (1996) recorded site 35KL1820 during an inventory for fire lines on the Klamath Forest National Wildlife Refuge. His crew observed 14 depressions and they measured some. They were circular and about 1 meter deep. House 1 was 6 x 5 meters, house 2 11 x 9 meters, house 3 was 10 x 8 meters, house 4 was 7 x 8 meters, house 5 was 8 x 8 meters, house 6 was 7 x 7 meters, house 7 was 10 x 5, and house 8 was 11 x 8. Site 35KL1821 contained two depressions: 4 x 5 and 5 x 7 meters. Site 35KL1078 contained over 80 depressions when looked at during this project. No measurements were taken.

[back to top]

North Bank 35DO61

Gray (2000) directed volunteers working at the North Bank Site (35DO61) on Roseburg BLM land along the North Umpqua River. They dug a trench through a house (Feature 7) and its associated hearth (Feature 2). The hearth was dated to 90±40 and 210±30. They were able to put together composite samples from the house floor: 330±40 and 200±40. The recovery of trade beads and metal fragments suggested a date range of 1800-1810 AD. The house was distinguished by a distinct soil change, its hearth, and its concave shape in the trench profiles. The E-W house width was estimated at 4 meters (13 feet) and the N-S as 6.5 meters (21 feet). The house was saucer shaped and about 35 cm (14 inches) deep. A wood fragment on the house rim was dated 280±60. During the testing, 294 prehistoric and 17 historic artifacts were recovered: 88 projectile points, 40 point fragments, 4 preforms, 5 drills, 2 perforators, 6 end scrapers, 4 unifaces, 6 modified flakes, 63 bifaces, 44 cores, 2 abraders, 5 net weights, 12 cobble tools, 4 pestles, 5 anvil stones, 2 choppers, and 2 bone awl tips.

The report, expanding on other work, suggested that shoulder-width, not neck-width determines if a projectile point was a dart or arrow. Above 20 mm they are dart points and below 20 mm they are arrowheads. Based on this, Coquille-series broad-necked points are arrowheads, not dart points, so their appearance in late sites would be expected (Gray 2000: 94-97) and were simple special function robust arrowheads.

« Previous Region | Index | Next Region »