Site History: Cliff Palace, Chapin Mesa
by Kathleen Fiero, NPS

Cliff Palace is the iconic cliff dwelling every visitor to Mesa Verde National Park has heard about and wants to see. The size and spectacular setting of the village does not disappoint. It contains the greatest number of rooms (rooms, kivas, open areas) of any cliff dwelling. It is this dense clustering of rooms within a large alcove, at various levels on the sloping floor, that appeals to all--undoubtedly to the original builders as much as to modern visitors. Like Mug House, this is not a preplanned village. Rooms were added, abandoned, modified throughout its occupation. The village as it appears today dates to the late 13th century. There are depressingly few tree-ring dates considering the size of the site but the bark dates from the wood that does exist are in the A.D. 1260s and 1270s. Cliff Palace is located on a finger of Chapin Mesa overlooking Cliff Canyon. This canyon drains into Soda Canyon, one of the major drainages of Mesa Verde. The Cliff Palace alcove is 80 m long and 20 m deep. The main portion of the site is on the sloping floor of the alcove. Because of the slope, the rooms and kivas are on various levels within the alcove and this creates much of the aesthetic appeal of the site. There are also ten rooms on a natural ledge above the very back portion of the main village and a kiva within a tower outside the alcove, approximately 20m west of the north end of the site at the base of the cliff face. Cliff Palace is at an elevation of 6790' (2270 m) with pinyon and juniper the dominant overstory plants.

Cliff Palace is the largest village in a cluster of cliff dwelling villages. Sunset House and Swallows' Nest are down canyon. Balcony House is across the mesa to the east, and Oak Tree House, New Fire House and Fire Temple are across the canyon to the west. Besides these sites, there are numerous smaller cliff dwellings between Cliff Palace and the major villages. Also right across Cliff Canyon, on the mesa top, is Sun Temple--a late 13th century ceremonial site. The question of whether there was enough water for all the people who lived in the area is an important one. The occupants of Cliff Palace needed it for daily consumption and construction. In his 1911 report, Fewkes states that "there was no water at Cliff Palace" and that they developed a good supply in the canyon below the ruin "where there is every reason to believe the former inhabitants had their well" (1911:12). There is a very dependable spring across Cliff Canyon from Cliff Palace at the mouth of Fewkes Canyon, and seeps at the top of the talus slope at the contact of sandstone/shale lenses both north and south of Cliff Palace.

Nordenskiold visited Cliff Palace (Nordenskiold's Ruin 2) in 1891. He created a plan view map of the unexcavated site identifying over a hundred rooms which includes 17 kivas. He took numerous photographs of the site. He noted two estufas (kivas) that differed in construction from the ordinary form being square with rounded corners. Although not mentioned by Nordenskiold, they also do not have pilasters and in fact there is no evidence that they were ever roofed. None of the artifacts illustrated in Nordenskiold's publication are from Cliff Palace so quite possibly all the museum quality artifacts had been removed before his visit. The absence of structural wood throughout this site was so dramatic that it was noted by Nordenskiold. He speculated that it might have something to do with the "extinction or migration of the former inhabitants" (1893:66).

Cliff Palace was the second cliff dwelling excavated and stabilized after the creation of the park. Jesse Walter Fewkes' first project was the excavation and stabilization of Spruce Tree House in 1908. He returned to the park in 1909 to excavate and repair Cliff Palace. This took from May through August. His report on Cliff Palace was published as a bulletin of the Bureau of American Ethnology (Bulletin 51, 1911). This report is brief and primarily descriptive. He felt that there was no way to reconstruct the life of the people who had lived in the village. He blamed this on the lack of undisturbed deposits and the site's poor state of preservation, no wood remaining and many walls broken down. The site was mapped, possibly working from a 1907 map by Kidder, Morely and Nusbaum (Kidder 1960:24-26), with rooms numbered and kivas identified by letter. He concentrated on repairing wall foundations and reconstructed the corner of the four-story painted "tower" (actually a four-story roomblock) at the south end of the site. A volunteer, R.G. Fuller, photographed and surveyed the site during this project (1911:plates).

In the text of his report he mentions finding a great deal of broken pottery which he classifies as coiled and indented, or smooth polished. The illustrated artifacts from this excavation (1911:plates) include black-on-white bowls, two black-on-white kiva jars, a corrugated jar, two mugs, a dipper, stone ax with handle, six ax heads, three pot rests, a basket, numerous sandals, numerous bone awls, planting sticks, woven tumpline. He found two pieces of turquoise, a fragment of a Tchamahia which in a footnote he states are used on an alter at the Hopi village of Walpi. He states that the objects uncovered at Cliff Palace were similar to those found at Spruce Tree House. Some or all of these objects were sent to the National Museum (Smithsonian). He identified two main types of rooms, ceremonial and secular, with two subtypes of ceremonial rooms and six subtypes of secular rooms. Of the ceremonial rooms, the presence or absence of roofing was used to make a distinction between types. Most of the kivas (19) had pilasters which supported the roof. Kiva M was an exception but since it was roofed (the upper walls undoubtedly supported the roof), it was placed in this category for a total of 20 kivas with roofs. Three circular, subterranean rooms did not have pilasters and there was no evidence they were ever roofed, Kivas O, R (the two kivas also noted by Nordenskiold as being unique), and W, and so these were placed in his second subtype. Fewkes mentions kihus in a footnote (ibid. 49, footnote b) and suggests using the term for "ceremonial rooms above ground." There are no kihus in Cliff Palace but Rohn (1971: 84-86) identified one on the upper ledge in Mug House, Room 9, which had the following special attributes: decorated plaster on the walls, double course construction, lack of chinking stones in wall joints, hearth, possible loom loops in one wall.

Fewkes classifies the secular rooms based on function, at least in part: living rooms, milling rooms, storage rooms, rooms of unknown function, towers, and round rooms. He mentions that it is difficult to distinguish secular rooms on the basis of function because so much of the occupants' work took place in plazas and house-tops. He assumes that large rooms and rooms with banquettes were living rooms and owned by women as at Hopi. He does not mention what criteria he uses to identify storage rooms nor does he discuss what he believes to be the function of towers and round rooms. Fewkes discusses each secular room singly or as a group of rooms. Unfortunately he does not place each of these rooms into one of the above categories.

Fewkes had lived with the Hopi and was much taken by their culture. As with Nordenskiold, he is convinced of the relationship of the cliff dwellers to the Pueblo Indians and throughout the publication compares what he found at Cliff Palace with the situation in the Hopi villages. Although interested in the placement of Cliff Palace in time, he can only speculate on the temporal relationship of Cliff Palace to that of neighboring pueblos, such as Spruce Tree, and the ruin across the canyon on the mesa top (Sun Temple). He questioned whether the kiva/room ratio differed through time. At Cliff Palace there are seven secular rooms to every kiva while at Spruce Tree House there are fifteen rooms to every kiva. Comparing this to modern Pueblos where there are many more secular rooms than kivas, he concludes that Cliff Palace is older than Spruce Tree House. Fewkes feels that the final abandonment of the Mesa Verde was the result of an unfavorable environment. He found no evidence of a cataclysm at the end of the occupation. It is interesting that Fewkes found a great deal of evidence of what he thought was cremation. He even felt that at the north end of the refuse area, which must be the area behind Rooms 20 to 40, there is an area where the cremations took place (Fewkes 1911: 23, 39). Another interesting comment is made in a footnote (ibid. 38). He states that the rooms for storage were of the earliest construction and that this earlier use of alcoves for storage may have lead eventually to the choice of alcoves for habitation.

A big frustration of all early archeologists was their inability to place their observations in real time. When was Cliff Palace occupied, for how long and was this before or after the occupation of such sites as Aztec Ruin and Pueblo Bonito? It was less than twenty years from the initial idea of tree-ring dating to the creation of a real time chronology covering Pueblo prehistory but those must have been frustrating years for southwestern archeologists. Finally in 1929 the gap in the tree-ring sequence was filled. Cliff Palace was sampled for tree-ring material by the First Beam Expedition in 1923 and by the University of Arizona Tree-Ring Laboratory in 1931-2 (Robinson and Harrill 1974:53). The dates were first published in 1935. Even with a limited number of dated samples, it was clear that the site dated to the 13th century. Robinson and Harrill mention that there are few dates, i.e. little wood in relation to the size of the site. Nordenskiold and Fewkes also mention this. It still is not clear why this is the case. Is it due to the early disturbance by unauthorized excavation or did this happen prior to the complete abandonment of the site?

From 1909 to 1934 there is no record of any research or repairs to the features of Cliff Palace. A drainage trench was excavated into the rimrock above the site in 1932 (Horn 1989:23). Then in 1933 Public Works Administration (PWA) funds were allocated for the mapping, documentation and repair of Cliff Palace among other cliff dwellings. Earl Morris was in charge of the project but it was Al Lancaster, a local handy-man, failed bean farmer, who did the on-site supervision. Lancaster went on to an illustrious, forty plus-year career in the excavation and stabilization of sites in Mesa Verde NP. During this project the north end of the site was mapped in detail by NPS architect Stanley Morse, the boulder upon which the Speaker Chief's complex is built was stabilized with a steel I-beam and concrete, with the repairs covered with a veneer of stone. Wooden roof beams from Balcony House across the mesa and Aztec Ruin in New Mexico were added to the rooms making up the Speaker Chief's complex to stabilize walls (Horn 1989:23). Also the corner of the four-story painted tower was rebuilt due to the poor quality of the Fewkes' reconstruction (ibid. 23).

By the 1940s, moisture is mentioned as a serious problem in Cliff Palace (ibid. 24). It is not completely clear when these problems started or where they occurred within the site. Fewkes, as mentioned above, states that the cliff Palace alcove is dry. Soil/sand was removed from the upper ledge to reduce moisture in 1948 (ibid. 24). Then the collapse and settling of retaining walls/terrace walls convinced the park to excavate a tunnel "under" the site to intercept the water at the shale/sandstone contact. This was done in 1961. It was successful as the north end of the site is now relatively dry and there is a steady stream of water going through the tunnel. The tunnel is actually north of the site with its mouth below the site between Kiva W and the main portion of the site.

Cliff Palace has been the centerpiece of park interpretation since the park's founding. Hundreds and now thousands of visitors enter the site annually. Initially visitors walked through the site and a lot of time was spent by park staff controlling dust. Then in 1939 the visitor trail was moved to the front of the site and after World War II visitors were restricted to this trail (ibid. 28). This reduced the amount of physical contact the visitor had with the resource and lead to better preservation of such fabric as plaster and wall mortar. An enlarged parking area and water pipes for drinking water and flush toilets over Cliff Palace was part of the Park's Mission 66 plan. Almost immediately after piped water and wet sewage were in the area, water problems were noted in the middle and south end of Cliff Palace. The water problems in the north end of the site were solved by the tunnel at about the same time as water appeared at the middle and south end of the site. Various efforts were made to control the problem but nothing was successful until the water was turned off in 1995.

In fact it is a major problem with water in the fall and winter of 1994/95, with gallons of water flowing down a fault and into the back of the Cliff Palace alcove, that lead to the most recent research on the site (Nordby 2001). The sections of the site damaged by water were the most intensively studied, Courtyards J and M, during this research. But also a basic construction sequence for the entire site was developed and all wood in the structure was documented and sampled if of appropriate size for dating. The focus of the recent research has been on architecture based on a model of construction that starts with individual stones and builds up through rooms, room clusters, villages and communities of villages. Based on work done at other cliff dwellings in the Pueblo area, Nordby has identified five types of secular rooms, living rooms, granaries (for food), storage rooms (for non-food storage, with no evidence of a sealed doorway) and mealing rooms. He also has open areas (unroofed areas), and miscellaneous structures. Kivas are classified separately but with a note that they may have had a habitation as well as a ceremonial function (ibid 2001:14). Then following Rohn, Nordby feels that rooms of various function are clustered into room suites, which he equates with households. Then these suites of rooms are clustered into courtyard complexes (ibid.14-15) which include one or more room suites associated with a kiva. Nordby (2001:89) concludes that Cliff Palace had a permanent population of 25 households based on the number of living rooms (large rooms with hearths)/room suites.

Just as Fewkes was concerned about the low room to kiva ratio found at Cliff Palace, so are modern researchers. Long House on Wetherill Mesa is the only other large alcove village in Mesa Verde with a similar low ratio. Nordby sees the permanent residents as caretakers of the village. But, at certain times of the year, during special ceremonies, the village was used by a much larger population. Kivas would have then been used as sleeping quarters. This is based on the fact that at Cliff Palace there are a number of kivas without associated storage and living rooms, and areas in the site with a-typical rooms such as the rooms of the Speaker Chief's complex and Rooms 59 and 64 which are large but do not contain hearths. Nordby (ibid. 94, 95) considered the presence of a hearth as the most significant attribute to identify a living room. Attributes such as size and plaster varied widely in the two courtyard complexes studied in detail. He did find that all living rooms were on the ground floor and that the size of rooms of all function were smaller in one courtyard complex (Courtyard Complex M) than the other (Courtyard Complex J). Doorway attributes were tracked and bear a complex relationship with room function. This is probably due to the fact that the function of rooms often changed over time. Courtyard Complex J contained from two to five room suites or households through time with construction from A.D. 1260 to 1278. One concern of the author is that there are no storage rooms dating before A.D. 1271 in this courtyard. Courtyard Complex M was constructed from 1268-1272 and evolved from one suite to two suites, or households. In both of these courtyards the room walls outlining the open area (courtyard) were plastered and decorated and in fact become areas of research, documentation and treatment for architectural conservators as discussed elsewhere.

A very interesting discovery made in the recent research is the division of the village of Cliff Palace into two discreet areas. The dividing line can be seen on Figure 1 and extends from the back of the alcove north of Room 62 to the north wall of Room 58. One of Fewkes' a-typical kivas is on each side of the line and one large featureless room is on either side of the line. Recently dated wood places construction of these a-typical kivas to A.D.1278 and 1280. This is consistent with what has been found at Spruce Tree House and Balcony House where tree-ring dates place construction of walls which restrict access from one part of the village to another in the late 1270s. The a-typical kivas and the large featureless rooms are not found at Balcony House or Spruce Tree House. As mentioned above, Mug House also has a wall dividing the village. The date of construction of this wall is not established. Cliff Palace also contains the unique area, Speaker Chief's complex and Nordby (ibid. 108) would attach to this complex Kiva Q. Nordby (ibid. 108) believes that "this kiva (Kiva Q) truly represents the point at which the dual divisions come together, integrating two major subdivisions." The plaster on the lower walls of Kiva Q varies with the north half of the kiva wall a slightly different color than the south half. This will be dealt with in great detail below. But here it is enough to note that a physical division and a visual division are being used by researchers to conclude that this represents a social division. Ethnographically, many Pueblo groups are known to be organized into two overriding groups, or moieties. These social groups are not always based on kinship and are not spacially grouped in the physical layout of many of the modern villages (Dozier 1970:155). It is important to note that at both Mug House and Cliff Palace, the physical division is actually not dividing the village in half. At both villages, one side of the dividing line contains many more rooms and kivas than the other. Nordby (ibid.109) also feels that the Speaker Chief's complex is earlier than the dual division and was the focal point in the village by at least A.D. 1264. This conclusion is based primarily on a tree-ring date from a room that is an addition to the main complex, Room 80.


In the thirteenth century, on the periphery of the area then occupied by the Pueblo Indians, the local Pueblo farmers changed the preferred location for their homes from open areas on the mesa tops and in the broad canyons to the natural alcoves found in the sandstone cliffs. For five centuries most of their ancestors had lived near their farm fields, now in a matter of a few generations cliff dwellings with living and ceremonial architecture were built, modified, and then left. By the end of the thirteenth century construction had stopped. An area that had been farmed for 700 years was left unoccupied. The last occupants moved to areas more central to the Pueblo world at that time. Mug House is a typical example of a mid-size cliff dwelling on Mesa Verde. Its major period of occupation was A.D. 1204 to 1276. In that period a minimum of thirteen suites and four courtyard units were built and modified with the peak activity in the years A.D. 1250 to 1266. Cliff Palace, the largest village on the Mesa Verde, was built on a different mesa east of Mug House, with most of the extant construction built at the end of the construction period for Mug House. Cliff Palace as it presently exists dates to A.D. 1268-1280. This village contains the usual kivas, living rooms, storage rooms, mealing rooms. Just like Mug House, there is a wall dividing the village into two areas. There is also a complex of rooms associated with a walled open area, Speaker Chief's complex. There is nothing comparable at Mug House. Possibly even more interesting and again unique to Cliff Palace, there is on either side of the dividing wall an atypical kiva and a very large room. These special areas were noted by Fewkes and they still intrigue. Cliff Palace seems to be special. The park visitors certainly think so based on its size and beauty, archeologists think so based on the layout and contents of the village.

<< previous
© 2002 University of Pennsylvania, All rights reserved.