National Park ServiceU.S. Department of the Interior
Mesa Verde National Park Cliff Palace, Mesa Verde's largest cliff dwelling


[With 15 plates.]


The Mesa Verde, or Green Plateau, is situated in the southwestern corner of Colorado and was set apart by Congress from the Ute Reservation for protection of its prehistoric remains. Its form is oval, measuring about 42,000 acres, with an average elevation of over 7,000 feet above sea level, rising abruptly on the north side to 8,700 feet, over 1,500 feet above the plain. Its surface is cleft by deep, almost parallel canyons opening into the Mancos Valley on the south, between which are spurs of the mesa sloping gradually southward. In the canyons (pl. 2) are located the most remarkable cliff dwellings of the Southwest. The top of the plateau is dotted with mounds of earth and stone. The present article deals with one of these mounds, which was excavated and the exposed ruins repaired by the Smithsonian Institution, during the months of July, August, and September, 1916, at the request of the Secretary of the Interior, following a recommendation of the writer in his report to the latter on field work at Sun Temple in the summer of 1915.

Clusters of mounds composed of artificially worked stones and earth situated on the surface of the mesa have long been known, and from indications these piles of stones were believed to mark the sites of buildings. None of these mounds, however, had been opened, or their contents investigated. The plan of operations was to determine, by excavations, the character of the buildings concealed in them, and to interpret their cultural relations and significance. A cluster of mounds known as the Mummy Lake group was chosen as promising and advantageously situated for this purpose. The excavation of one mound of this cluster revealed a large building of a type new to the plateau.

The importance of the results of the work and their bearing on southwestern archeology may be better appreciated after reading what immediately follows. A portion of the area now known as Arizona, Utah, Colorado, and New Mexico was inhabited in prehistoric times by Indians culturally unlike those of any other region of North America, and for that reason this unique territory bears the name Pueblo culture area. It is, in fact, the only aboriginal culture area where buildings have determined the name, being distinguished from all others mainly by architectural characters. This limitation of characteristic terraced communal houses to a geographical area leads us to associate climate or other conditions of that area with peculiarities of buildings, as cause and effect.

Of man when he first entered the Southwest, we know little save that his physical features show that he was an Indian. The time of his advent is in doubt. Considerable obscurity also exists regarding the direction whence Indian colonists entered this district, but there is no doubt regarding the geographical locality where Pueblo culture, judged from the character of buildings, originated. The immigrant clans that first peopled the Southwest are supposed to have come from people who built neither cliff dwellings nor pueblos, consequently this style of dwelling originated exactly where it is now found.

But the Pueblo culture must not be interpreted solely by peculiarities in buildings, for although it receives its name from architectural characters, there are influential factors that it shares with those of other tribes of Indians which are very important. One of the most noteworthy of these is the possession of maize or Indian corn as a reliable food resource. Agriculture is one of the cornerstones of the Pueblo culture, as masonry is another. When man first entered the Southwest he knew little of the advantages of stone as a building material, for he built his hut of mud, sticks, or possibly of skins of animals. The North American Indian became a good stone mason as a result of a life in caves. Nowhere outside of the Southwest were elaborate buildings1 constructed of dressed stone by the aborigines north of Mexico. Masonry and agriculture, then, are the primary factors that determined the essential peculiarities of Pueblo culture.

1Stone walls and vaults were, of course, constructed elsewhere by Indians; cf. Mr. Gerard Fowke's article, Bur. Amer. Ethnol. Bull. No. 37, et al.

The Mesa Verde was set aside as a national park on account of its prehistoric stone buildings and monuments. While it presents rare facilities for a study of aboriginal architecture, it shares with other regions of the Southwest the condition that imperishable aboriginal buildings have survived from prehistoric times. Evolution of masonry in this region is a development which occurred in prehistoric times, or before the advent of the white man. No European ever saw an inhabited cliff dwelling on the Mesa Verde, and no article of European manufacture has ever been found in the undisturbed débris of the rooms. These cliff dwellings were abandoned before the Spanish conquest.

The inhabitants of the caves on the Mesa Verde were ignorant of hieroglyphs or letters, and therefore have left no written account of their origin and early history, although vague traditions are preserved by their descendants, especially among living Pueblos, as the Hopi. The most reliable data we now have to aid us in interpreting their culture are their buildings and archeological remains, or monuments, and minor antiquities, called artifacts, especially objects of burnt clay, that they have left behind. Their houses are the most significant.1 As pointed out by Westropp, in referring to prehistoric and historic cultures of other races: "Architecture is the external form of their public life; it is an index of their state of knowledge and social progress."

1Probably some writers on classical archeology would hardly consider cliff dwellings architectural forms.

One type of building characteristic of this culture is illustrated by Spruce-tree House and Cliff Palace, but this is not the only form. There are others, such as Sun Temple, brought to light in the summer of 1915, in which we find a building specialized for religious purposes.

Field work in the Mesa Verde during the summer of 1916 first revealed still another type differing considerably from the two preceding. This type, locally new, is known to ethnologists as a pueblo, commonly defined as a terraced community building constructed in the open or not attached to cliffs. It is a representative of many buried houses on Mesa Verde, and it is not too much to say that formerly there were as many buildings of pueblo type on top of the plateau as there were cliff dwellings in its canyons. Manifestly a knowledge of the Mesa Verde variety of pueblo is desirable, and a description of it will enlarge our conception of prehistoric culture in this locality. The object of the present article, then, is to make this known as a contribution to our knowledge of the aborigines of Mesa Verde.

The general condition and situation of mounds on the surface of the plateau will first be considered.


One of the best known group of mounds in the Mesa Verde National Park is situated south of a reservoir called Mummy Lake. There is no good reason for calling this prehistoric reservoir a lake, for it is not a lake, and no mummies have ever been found in or near it. The term "Moki Lake," by which it is sometimes designated, is equally meaningless, but both names are so firmly fixed in literature that it is difficult now to substitute others.

The best description of this reservoir is the following, quoted from Baron Nordenskiöld's "Cliff dwellers of the Mesa Verde" (p. 74):

A structure of considerable size, which was probably utilized for purposes of irrigation, lies on Chapin Mesa, some kilometers above the great ruins and not very far from the slope into Montezuma Valley. A large depression 30 meters in diameter is surrounded by a low, circular wall 4.5 meters thick. Water was probably conducted to this reservoir from some neighboring gulch. Traces of a ditch which formed the connection have been observed north of the reservoir by Richard Wetherill. A view of the reservoir is given in figure 43, which, however, shows only a part of a low, ring-shaped mound overgrown with bushes, all that is left of the thick wall. Quite near the reservoir we find the ruins of a considerable village, but the walls are now leveled with the ground, leaving only huge heaps of stone to mark the site.

Mummy Lake, or Moki Lake (pl. 7, fig. 1), is an artificial depression surrounded by an oval or circular ridge of earth, in places outlined by double walls of stones suggesting rooms. Excavations at the base of these stones, however, show that their foundations do not extend far below the surface; but work thus far has not been sufficient to prove conclusively that there were not rooms on the periphery. Mummy Lake lies on the northern edge of a group of mounds where the slope of the surface of the plateau would seem to indicate that water could be readily drawn from it. It is probable that the farms of the ancients were situated between the pueblos of the Mummy Lake group, and that these farms were irrigated by water drawn from this reservoir by means of irrigating ditches. In the time that has elapsed since the Mummy Lake pueblos were deserted the reservoir, like the ditches, has been filled with wind-blown sand or soil, so that its depth has greatly diminished, and at present water remains in it only a short time. Probably in prehistoric days it contained a perpetual water supply of a purer quality than now, when it is fouled by cattle excrement and made impure by mud washed into it from the surrounding banks; and if such were the case the reservoir probably supplied the neighboring pueblos with drinking water, since springs in this neighborhood are remote and very difficult of access. For instance, at the bottom of Soda Canyon there is an unpalatable soda spring, a climb from which to the pueblo is very arduous. Another spring, at the head of the same canyon, now used for watering stock, is over a mile distant, while a third possible source of palatable water is near the head of Navaho Canyon, even farther away. There was a small reservoir, possibly communicating with the larger by canals, now clogged with sand at each mound in the group. One of these minor reservoirs is indicated on the map near the mound excavated.


A much worn trail extended from Mummy Lake to Spruce-tree House, just east of the house excavated, and between it and the rim of Soda Canyon. This trail, used by horsemen before the Government road was constructed, was probably an old Indian path of great antiquity, connecting the various pueblos of the Mummy Lake group with Spruce-tree House and Cliff Palace. A steep branch trail descends from it over the rim of Soda Canyon to the spring above mentioned, near which are mounds of ruins sheltered by Steamboat Rock. This trail may have been used by water carriers in prehistoric times.

The position of the mounds on the plateau near Mummy Lake were first designated on an excellent map of Mesa Verde, published by the United States Geological Survey. It is evident from this map that the cluster of mounds near Mummy Lake is only one of several groups; for instance, from the rim of Soda Canyon, looking north and east, four similar clearings can be seen, in each of which are several artificial mounds, all of which have the same general form and are covered with sagebrush.1 No regularity is noted in their arrangement (pl. 1), but they vary in size and shape, all appearing to have, as a common feature, a central depression, which, judging from that excavated, indicates a large kiva. We find superficial evidences of rectangular and oval houses, and in one instance the building under the mound may have a D shape. Fragments of walls projecting above the ground are absent in all cases, but in one or two instances the direction of the buried wall can be followed for a few feet by surface indications.

1This relation of mounds to sagebrush covered clearings is discussed by Dr. Prudden, Amer. Anthr., Vol. 16, No. 1, 1914.

As these communes or clusters of small pueblos are more conspicuous in clearings than among the thick cedars, the question naturally arises whether they were built before the cedars grew or whether man burnt or otherwise removed the trees of the forest before he laid their foundations. The author inclines to the belief that the clearings were made by the hand of man, and that cedars were growing on the mesa when man appropriated it for his habitation or for planting. When once removed the constant tramping of people would certainly prevent trees from again growing on the cleared areas. At the time the buildings were inhabited they were surrounded by farms cleared of underbrush, and it appears from the amount of sand and soil filling the rooms of the pueblo that the wind played a great role in transporting sand to the mound from the surface of the bare earth. Sagebrush or trees would tend to anchor the soil and prevent its blowing away, which implies that the sagebrush has grown since the fields were no longer cultivated. As shown on the map (pl. 1), one or two of the smaller mounds of the group lie outside the clearing or in the cedars, a few of which trees also appear on top of the mounds. It may be mentioned that there is no evidence of a sagebrush clearing in the area about Sun Temple, which supports the theory that it was unfinished and uninhabited. Had Sun Temple been a domicile we would expect what we find in the neighborhood of Mummy Lake, some evidences of cultivated fields.

The sagebrush clearings are very fertile and throughout the summer months are carpeted with flowers, the most abundant of which is the "Indian paint brush"; later these plants, rare or unknown among the cedars, are succeeded by various species of asters. On account of the large number of flowering plants in the sagebrush clearings, unusually tame humming birds are very common, but with the advent of autumn they likewise vanish and the leaves of the scrub oaks change their colors and the mesa top is brilliantly painted with bright yellow and red. Almost everywhere, especially over the surface of the mounds, fragments of pottery are abundant, and here and there on the level surface between the mounds are remains of low stone walls, suggesting pit-houses,1 gardens, or irrigating ditches.

1It would be futile in the present state of our knowledge to speculate on the number of the inhabitants of these buildings long ago fallen into ruins, if simultaneously inhabited. There is no doubt it was large, much greater than suspected by early investigators. We are on the threshold of a great research and every year's field work will advance us a step in deciphering the history of this interesting race.

There are several clusters of mounds visible from near the Mummy Lake group. On the side of Soda Canyon there is an elevated outcrop, called Steamboat Rock, which protects a cluster of mounds, with sunny southern exposure, from the north winds. In a clearing on hills near the head of Soda Canyon there are also mounds or sites of former pueblos. It is important to note that these groups of mounds always occur in sagebrush clearings; their occurrence among cedars, where they are smaller, is common but less conspicuous. The many flowers blooming in these localities show that the land is rich, and it is probable that Indian corn could still be grown on the Mesa without artificial irrigation.



The mound in the Mummy Lake group chosen as a type for excavation to determine the character of Mesa Verde pueblos is situated four miles and a quarter due north of Spruce-tree House, and is one of sixteen scattered at intervals on both sides of the Government road. It stands about an eighth of a mile east of this road, a few steps from the rim of Soda Canyon. This pueblo (pl. 4) might be called Far View House, for the distant southern outlook from it is very fine and has been commented upon by almost every visitor.2 From its highest rooms the corners of the four States of Utah, Arizona, New Mexico, and Colorado—the only case in the country where four States meet in one point—can be seen far to the southwest. Sleeping Ute Mountain, Ship Rock, once called the Needles, and distant mountains of Arizona, rise on the horizon to the south and west. In the less distant foreground, beyond a forest of cedars, one can trace several important canyons of the Mesa Verde, among which may be mentioned Navaho, Mancos, and Soda. When the wind is favorable, the flag at Spruce-tree camp can be seen as a speck waving above the trees; the course of Spruce-tree Canyon can be traced without difficulty through its whole length. The surface of the land south of the ruins is covered with a dense forest of cedars and pin˜on trees sloping to the south. Looking back from the well-known tower at the head of Navaho Canyon or across country from the fine ruin, Spring House, one could make out the workmen on the ruin, with a good glass. Not many feet (80) from the southwest corner of the court there is in view a large mound pleading for excavation which may have an interesting story to impart regarding aboriginal culture. Two mounds in the group are situated in the cedars beyond, and a third, of large size, lies just south of the edge of the sagebrush clearing. The site of the pueblo is the most prominent one in the southeast corner of the area, and in a way this pueblo may be said to dominate the others. It was probably the largest, the most populous and important.

2The name Far View House, which calls attention to this fact, was suggested by one of my workmen, Mr. Jason Myers.

When excavation work was begun, the entire surface of this mound, like all of the group, was covered with sagebrush (pl. 3) and, like them all, showed a deep circular depression in the interior strewn with stones and débris. Some seeker after curiosities had dug a shallow trench on the highest point of the north side, revealing a fragment of a well-made wall and the sides of a doorway. From this a trench had been dug across the mound to what was eventually found to be the south side. This excavation had not determined the form, size, or height of the building, and probably did not reward the workmen with the small objects they sought.




Almost every visitor to the pueblo while the excavation was in progress remarked on the quantity of débris that filled the rooms and naturally asked whence it came. Many visitors were sure it indicated a great age; that a long time had elapsed to fill the rooms. The writer has also given much thought to this condition and concludes that this alone does not prove a great antiquity. It is difficult to explain this condition and to draw conclusions therefrom, but an examination of the arrangement or stratification of débris in the rooms is significant. For several feet below the surface the débris consists mainly of fallen stones mixed with adobe, resulting from the overturned tops of the walls. Penetrating deeper or below this stratum, soil free from stones was found. This material, identical with the sand of the plateau, appears to have been blown into the rooms or brought from the surrounding fields by wind storms. The accumulation of débris due to falling walls and the addition of wind-blown sand would progress very rapidly as long as the wall projected above the ground, but would then cease. That time might be measured by centuries, certainly not by millenniums. Lower still occurs a layer of ashes with fragments of charcoal, a "pay dirt" in which artifacts are common. This is mixed with adobe, evidently remnants of plastering. Deeper sometimes follows another layer of sand or an aeolian deposit; a sequence not uniform and not the same in thickness in all rooms. Evidently some of the rooms had been deserted and the sand accumulated to a depth of 2 or 3 feet, after which they were reoccupied and foundations laid on the sand, the new walls having been constructed to the height of the adjoining walls on this foundation.

FIG. 1.—Ground plan of Far View House. Mesa Verde National Park, Colorado.


The arrangement of rooms reduced to a ground plan is seen on figure 1. The lowest story of the main building has 40 secular rooms and four ceremonial chambers or circular kivas. A few of the secular rooms have not been excavated to their floors. The majority of these are arranged in two tiers on the north and west sides. They are two-storied; the floor beams of the second story, which are rafters of the first, were found and left in place. The row of rooms north of kiva A likewise show evidences of the existence of a third story, so that it may be said there were about 50 secular rooms in the building.

All the secular rooms on the west side were completely excavated, and earth was removed from all kivas. The court or dance plaza is situated south of the main structure and is inclosed by a low wall measuring 110 feet on the south side, 37-1/2 feet on the east and 34 feet on the west side.

The peculiarity of this pueblo consists of a large central circular kiva, around which are grouped secular rooms, to which are added smaller circular kivas. This central room recalls a tower, but, unlike some of the towers, this and the smaller kivas have pilasters attached to the walls for support of a vaulted roof. The great size of the central kiva suggests that the room was not limited to one clan; it points rather to a fusion of clans forming so intimate a union of several families that the room may no longer be considered as limited to men of one clan, but the meeting place of a fraternity of priests, drawn from several clans. The formation of such a fraternity is an advance, sociologically speaking, upon what we find indicated by the small clan kivas of cliff dwellings and implies, more recent construction.

The regularity of the secular rooms, as shown in plate 5, strikes the observer at first sight. The partitions separating these rooms run north-south and east-west and are continuous through the pueblo. No such regularity is found in cliff dwellings, although it is a marked feature of pueblo ruins along the Chaco and elsewhere. Inhabited pueblos as Zuñi, Walpi, and others show this character only to a limited extent.

The rooms of this pueblo are consolidated into a rectangular form with straight walls broken on the south. The building is oriented approximately to the cardinal points, and terraced to secure, sunny exposure on the south side. The method adopted by the Mesa Verde people in orienting their buildings, as Sun Temple, seems to have been followed at this pueblo, and reveals a knowledge of solstitial sun rising which is instructive. The sun priests of the pueblos had, of course, no compass and probably the polar north was unknown to them. Their north, west, south, and east, as with the Hopi, are not the same as ours and the line of the south wall of Sun Temple was determined by the position of the sun. It was not made haphazard, but was carefully thought out and determined by astronomical observation before the foundation was laid down. At the autumnal equinox theoretically the sun rises in the east and sets in the west. In other words, sighting from the shrine along the so-called south wall on that date we ought to see the sun rise on a continuation of that line, if the wall extended exactly east and west. Observation shows that such is not the fact; the south wall does not extend exactly east and west. On the morning of the 21st of September, in company with several others, the writer determined this by observation; and found the line of the south wall if extended would touch the point of sunrise on the horizon a little more than 20° north of the extended line of the so-called south wall.

The same is true at sunset as viewed in the opposite direction as observed by my friend, Mr. T. G. Lemmon. The sun on that date sets about 20° from the extended line of the so-called south wall, which, if projected, would touch the point of sunrise at the summer solstice. This is so exact that the builders of Sun Temple probably determined the direction of the wall by observation of the sun as seen from the sun shrine at that solstice. The point of summer solstitial rising of the sun, as observed on the horizon, as well as sunset in mid-winter were cardinal points among them, as among the Hopi, and determined the lines of their temple devoted to sun worship. It seems to have been in somewhat the same way that the orientation of the pueblo at Mummy Lake was determined, but as the south wall is more irregular and the building more patched upon this side, it was not as easy to make observations there as at Sun Temple, but it was possible to use the north wall for that purpose.

Inasmuch as some of the highest walls had been reduced in altitude by the fall of their tops there had accumulated around their foundations a mass of detached fragments. The remains of fallen walls were especially extensive along the north wall, and the removal of this material was a work of considerable magnitude. Scrapers and stone boats were used for that purpose but the wall itself was laid bare by hand. The funds appropriated for this work were insufficient to permit the removal of this mass to a considerable distance to make the desired grading, but an automobile road was constructed around the ruin so that it can be visited with little inconvenience.

The excavation was begun on the northwestern corner of the mound (pl. 6), which later proved to be a small square room (pl. 6, fig. 3) annexed to this angle of the building. It was found that the greater part of the northern wall had been reduced to about 6 feet in height, and that the partition walls of several rooms formerly attached to it had been shattered. The east wall (pl. 6, fig. 4) was in somewhat better condition, but inclined so much outward that it was considered advisable to construct a buttress to hold it up. The south wall (pl. 8, figs. 1, 2) was irregular and much broken down; here (fig. 1, a, a, a,) buttresses were necessary. Near a recess situated midway in the length of the wall props (aa, aa) had been constructed by the aborigines to hold it from falling while the building was still occupied. The east wall also leaned considerably and had to be repaired. There is fine masonry in certain portions of all these walls (pl. 9, fig. 2; pl. 10, fig. 2), but on the whole it was inferior to that seen at Sun Temple.





As the number of rooms is greater than in Sun Temple, the work of excavation was more laborious than in the preceding summer, the shattered walls necessitating more repair work. The walls of a few rooms had been constructed on sand foundations, indicating that these rooms had been deserted and reoccupied. Other walls showed evidences of having been repaired while rooms were still occupied. The writer had no doubt that the building was a habitation, as many objects of household use were found at all depths from the very inception of the work.

The main north wall, exclusive of a small room of unknown use on the northwest angle, measures 113 feet from the northeast to the northwest corner, and was formerly about 20 feet high. The east wall extends 50-1/2 feet and the west wall 64-1/2 feet, both averaging about 10 feet high. There is a court surrounded by remnants of a wall rising a foot out of the ground on the south side. This wall rises highest where it joins the southeast and southwest angles of the main building. About midway in its length there is a recess in the south wall, evidently intended to hide the entrance ladder, resembling a similar recess at Sun Temple and Cliff Palace. The angles of this recess and the accompanying wall show good masonry; the corners inclined slightly outward, not being properly bonded to the remaining wall. The masonry throughout is fair but shows all the faults of cliff dwellers' work; joints unbroken, corners not bonded or properly tied to the other walls. The adjoining surfaces of the superposed stones were not flat, the mason relying upon slivers of stones, set in mud, to fill the intervals between them. He so multiplied the number of these stones that it weakened the walls, for the mud in which these were inserted easily washed out and the walls became unstable in course of time, notwithstanding they are thick, though in some cases the walls are narrow, not more than a few inches wide. Marks of human hands, and in a few instances impressions of corncobs were seen in the pointing of the walls—the latter perhaps accidental; no marks of a trowel were found.

Large, flat, thin, unworked stones set on edge occur at the southwest inner corner, where the wall surrounding the court joins the south wall of the main building. These stones are of such a size that they may be called megaliths, as it would require three men to handle one of them. Their insertion in the wall is regarded as a survival of a stage in southwestern masonry antecedent to the employment of hewn stones.1 As a rule there were no stones in the wall construction that could not be carried by a single pair of hands.

1Jackson describes an extensive wall of a ruin in Montezuma Canyon constructed in this manner.

A majority of stones show evidences of artificial pecking or dressing on their surfaces, a few being smoothed by attrition. Plastering as a rule is absent, but appears in layers over the surfaces of the small kivas. Its absence on the rectangular rooms and presence on the kiva walls suggests that it was protected by the vaulted roofs of the latter, which fell long after those of the former.

There are many stones with incised decorative figures, possibly "mason marks," different from those of Sun Temple set in the inner walls of the building. The spiral (fig. 2), representing the serpent of the water, occurs several times. The same figure was noticed on a round room lately discovered a mile from Spruce-tree House and on a round tower in Cannon Ball ruin, near the McElmo. One of these spirals was accompanied by radiating, peripheral, parallel lines, suggesting a figure of the feathered snake. Several of the more striking figures from stones fallen or still remaining in the walls are shown in the accompanying figures (fig. 3). They resemble designs on black and white pottery ware.

FIG. 2.—Serpent symbols incised on rocks in masonry.

FIG. 3.—Incised figures on masonry.

Various interpretations have been suggested to explain these figures, some of which are fanciful; there is no reason to doubt that they were primarily decorative, but they may also be symbolic. The complicated form of several incised figures suggests something more than meaningless efforts at embellishment, but it is too much to hope that they have any value as inscriptions. Although these designs are regarded as decorative, the limitation of the spiral to round rooms, towers, or kivas hints at a deeper significance. There is an obscure legend among the Hopi that circular kivas are connected in some way with snake ceremonials, and the association of the spiral sign and circular rooms seems to support, in a way, this idea.