DIVERSITY OF CALIF U JOLLA S Slttatthtt? nf Ammra PAPERS OF THE SCHOOL OF AMERICAN ARCHAEOLOGY number fifteen EXPLORATIONS IN SOUTHWESTERN UTAH IN 1908 ALFRED VINCENT KIDDER igio Institute of America EXPLORATIONS IN SOUTHEASTERN UTAH IN 1908 THROUGH the generosity of Colonel E. A. Wall of Salt Lake City the Utah Society of the Archaeological Institute of Amer- ica is enabled to send an expedition into the field each summer to study the archaeological remains which are scattered thickly throughout the southern part of the State, and to make collec- tions for the Museum of the University of Utah. It was my good fortune to be appointed Field Assistant for the year 1908, and put with Professor Byron Cummings of the University of Utah in joint charge of the summer's field campaign. We were instructed by Dr. Edgar L. Hewett, Director of American Archaeology, to choose a region in Southeastern Utah, spend a couple of weeks in its exploration, and then select a site to be excavated during the remainder of the sea- son. Having thus the freedom of action essential to successful field work, Professor Cummings and I met on June 11 at Monticello, Utah, and decided to explore the western tributa- ries of Montezuma Creek, in the southeastern portion of San Juan County, hoping not only to obtain some idea of the num- ber and distribution of the prehistoric remains of that region, but also to find a ruin suitable for excavation. We left Monticello on June 12, accompanied by Messrs. Neil Judd and Clifford Lockhart, students in the University of Utah, and James Hambleton, a cattleman, who was of great assistance to us as a guide. Later in the month, while we were engaged in excavations about Cave Springs, we were joined by Messrs. H. G. de Fritsch and Leavitt C. Parsons, both students in Harvard University. These gentlemen remained with us to the close of the season and, with Messrs. Judd and Lockhart, were constantly at the works, where they rendered valuable American Journal of Archaeology, Second Series. Journal of the 007 Archaeological Institute of America, Vol. XIV (1910), No. 3. 338 A. V. KIDDER assistance. I am indebted to Messrs, de Fritsch and Parsons for the map of the ruin given in Figure 2. Our work closed on August 1. Montezuma Canon (see Fig. 1), or Montezuma Creek, as it is locally called, is a deep and rather narrow valley, which heads in the eastern slopes (j ^ of the Sierra Abajo or ^T ^- v - x Blue Mountains, and ' flows in a southerly direction some 45 miles before emptying into the San Juan River. Its eastern tributaries drain the long mesa which separates it from the McElmo - Yellow Jacket system, while on the west its upper tributaries all head against the eastern and southern slopes of the Abajos. The work of the expedition was con- fined to these upper western tributaries. According to the cat- tlemen of the vicinity the lower western can- ons are short and con- tain few ruins. Montezuma Creek itself contains running water throughout its whole course only in wet seasons. At other times the stream sinks into the sand far above its mouth, and continues to the San Juan in the form of an underflow which reappears here and there in the form of " seep springs." Along FIGURE 1. SKETCH MAP OF WESTERN TRIBU- TARIES OF MONTEZUMA CREEK. SCALE : 1 INCH TO 12 MILES. (* Marks the ruin excavated.) EXPLORATIONS IN UTAH 339 the course of the river there is a considerable growth of cotton- woods, but apart from these there is little vegetation in the canon-bottom. The surrounding sandstone mesas, however, are thickly overgrown with dwarf cedar and pinon trees, replaced, as the country rises toward the Abajos, by spruce and yellow pine. The western tributaries are merely smaller replicas of the Montezuma itself, being, in most places, narrow, gorge-like canons with barren, sandy bottoms and abrupt, cliff-like sides. None of them contain live water in their lower reaches during the summer except after heavy rains in the Abajos. By digging in the stream-beds, however, a small supply of rather alkaline water may usually be obtained. A few fine clear springs are to be found, chiefly in Alkali Canon and its branches. The first considerable upper western tributary of Montezuma Creek is Long Caiion. It heads against the Abajos and flows in a southeasterly direction, gradually becoming deeper and more barren, until it debouches upon a wide " bench " above Montezuma Creek. Devil's Caiion, the next valley to the south, follows practically a parallel course. Alkali Canon, the largest western branch,. again heads in the Abajos, but instead of flow- ing east during its whole course, soon turns nearly south, thence running parallel and close to Montezuma Canon for some 15 or 20 miles before eventually entering it. This leaves a narrow mesa, called Alkali Ridge, between the two systems. Because of the narrowness of this ridge, all the canons empty- ing into Montezuma Caiion between the mouths of Devil and Alkali canons are short. The majority of them are little more than draws and probably contain few ruins. The country in that region is, however, so split up and broken, and presents such a tangle of steep gullies, cliffs, and precipitous ravines, that many weeks would be necessary for its complete exploration. We confined ourselves, therefore, to a study of Long, Devil's, and Alkali canons ; and also examined the ruins about the heads of Rustler's and Ruin canons. The former is a small western branch of Alkali Caiion ; the latter one of the largest of the short draws which drain Alkali Ridge and run into Montezuma Caiion. The prehistoric remains of the region fall into three well- defined groups : (1) Cliff-dwellings, (2) Canon-head dwell- 340 A. V. KIDDEE ings, and (3) Pueblos. Cliff-dwellings are built in caves or on ledges of the cliffs. Caiion-head dwellings are loose aggre- gations consisting of a considerable number of separate small houses, which are always formed about the abrupt ends of small canons. Pueblos are more or less compact settlements built in the open, either on the mesa-tops or in the canon-bottoms. Cliff-dwellings were found scattered thickly throughout the whole region explored, from the heads of the tributary canons to their mouths, and all along the course of Montezuma Creek itself. They were, indeed, the only buildings found in the greater part of Long and Devil's canons and in the headwaters of Alkali Canon. The two or three pueblos which we noticed in those regions were very small, and had every appearance of having been merely temporary affairs. Two canon-head groups were found in the branches of Ruin Canon, one at the head of Rustler's Canon, and two in small western tributaries of Alkali Canon. There are many large pueblos about the middle and lower reaches of Alkali Canon and along the side branches of Ruin Canon. There seems to have been a great centre of population along the whole middle portion of Alkali Ridge. The majority of these large ruins are situated on the mesa, the few which are to be found in the bed of Alkali Canon itself being rather small. CLIFF-DWELLINGS Cliff-houses are built in every conceivable sort of situation, and therefore cannot, of course, be classified by shape or loca- tion. The simplest type is a small natural cave made into a single room by the addition of a wall closing in the front. Between this and such buildings as Cliff-Palace and Sprucetree House on the Mesa Verde, which are really great pueblos built in caves, and hardly dependent at all on the cliffs, there is an endless variety of types. None of the cliff-houses in the region here under discussion, however, are of any great size. The largest of them do not contain over eight or ten rooms, while the majority are merely single- or double-room structures, their roofs and back walls usually being supplied by the cave in which they are built. EXPLORATIONS IN UTAH 341 In spite of their small size these cliff-dwellings are neverthe- less interesting in that they show great ingenuity of construc- tion and bear testimony to the adaptability and resourcefulness of their builders. Their protected situations, also, have shielded them from rain and snow, and so preserved for our study cer- tain architectural features, such as doorways, beams, and roof- ing, which, in the pueblo ruins in the open, have long since disappeared. The masonry of the cliff-dwellings of the Montezuma Canon district is much inferior to that seen on the Mesa Verde to the east and in Grand Gulch to the west. The building-stones were here simply cracked out or picked up at random, and at best very rudely shaped. Many of the walls consist merely of small, irregular stones set in adobe with no attempt at coursing. The surfaces, both inside and out, are usually coated with adobe roughly laid on with a wooden implement or with the hand. Beyond this there is little plastering, although all cracks and crevices in the back or cliff walls of the houses are carefully plugged up with small stones or corn- cobs set in adobe. The floors of the rooms are seldom levelled or filled in to do away with irregularities, the natural rock being left without modification. Wooden beams are sometimes incorporated in the masonry, usually along the foundations of walls, for the purpose, probably, of bridging spaces which it would be difficult to span with masonry. This use of wood is very uncommon on the Mesa Verde and McElmo, but I have been told that farther to the west, in White Canon and Grand Gulch, it is a very common style of building, and that some of the houses are almost entirely constructed of logs and adobe. Another feature which is comparatively rare on the Mesa Verde, 1 but of which we found a fine example in a small house in Devil's Canon, is the wattle-work wall. This wall was be- gun after the manner of a picket fence by placing upright and about a foot apart a number of slim cedar poles. These poles were then wattled together with twigs and osiers, making a fairly close and basket-like surface, which was then coated inside and out with adobe until the whole had a thickness of about three inches. This construction appeared to form an 1 It occurs in a ruin in Fewkes Canon and in Long House. 342 A. V. KIDDER addition to the house, and to have been built at a later time than the other walls. The cliff-dwellings of the region are, as a rule, very small and for the most part placed in caves so low that their roofs are also the roofs of the rooms. For this reason artificial roofs were seldom necessary, and as all traces of the roofs of the pueblos in the open have long since rotted away, our study of this feature of ths architecture was limited to a single example offered by a cliff-dwelling in Devil's Canon. This house, built on a ledge some 15 feet above the talus slope, consists of a series of seven or eight rooms, the westernmost of which are partly protected from the elements by a projection of the cliff overhead. The last room of the house is excellently preserved. It is 8 feet long by 4 feet wide, the back wall being formed by the cliff. At a height of 7 feet 6 inches from the floor a cedar beam 8 inches in diameter at the small end runs the length of the middle of the room par- allel to the cliff. Its two ends are set into the masonry of the walls without projecting through. At right angles to this main beam and resting upon it are four smaller beams about 2 inches thick. Their outside ends are set in the masonry of the outer wall, the inner ends resting against the cliff, where they are held in place by daubs of adobe. Upon this second series, and at right angles to it, or parallel to the main beam, are laid slabs of split cedar of about the length and thickness of ordinary "shakes." They cover the entire roof, and a layer of adobe some 3 inches thick is placed directly upon them. There is no coat of cedar-bark between the adobe and the wooden part of the roofing, such as usually occurs in Mesa Verde houses. The top of the roof is carefully levelled off, the walls rising a few inches higher than its upper surface. A trapdoor leads from the room below to the open housetop. This door is 20 inches long by 15 inches wide and was coped about by flat stones, one of which is still in place. The other rooms of the building seem to have been covered in the same way, although they are in a so much more advanced state of ruin that the beams have nearly all rotted away. For statistics as to doors also we are dependent on the cliff- dwellings, for in them alone are the entrances sufficiently well EXPLORATIONS IN UTAH 343 preserved for measurement and study. We found them to differ little from those of the Mesa Verde and McElmo. They are rectangular, and an average of the many examples that we examined gives the following dimensions : height 22 inches, width 15 inches, thickness of wall 11 inches, height from floor of room 20 inches. They are usually fitted with a single large slab of sandstone for a sill, while the lintels are made either of a similar slab or of several small wooden rods sunk in the ma- sonry of the jambs. A single rod about an inch below the middle of the lintel served as a rest for the stone slab which was used to close the door. The Tau-shaped door does not, so far as I know, occur in the western tributaries, although in Montezuma Creek itself we noticed one example. What the purpose of such large numbers of cliff-dwellings could have been is more or less a puzzle. That they were of the same culture as the pueblos seems proved by the potsherds found in both. With the exception of two houses in Devil's Caiion, one in Ruin Canon, and one or two in Alkali Caiion, none of them contain more than two or three rooms, while the great majority are nothing more than single chambers hardly large enough to hold a man, and usually built in caves so low that one cannot sit upright in them. Even the larger exam- ples just mentioned contain only six or eight rooms, and even these groups do not contain the kiva or ceremonial room. The relation of the kiva to the religious and tribal life of the pre- historic people is as yet not clearly understood, but it neverthe- less seems probable that no permanent dwelling-place could exist without it. For this reason the absence of the kiva from even the largest of the cliff-dwellings makes it seem improbable that these buildings were ever continuously inhabited. Whether they were lookout places, granaries, or shelters from which to watch the cornfields, are questions which it is better to leave open until more complete data as to their exact topographical situation and their relations one to another and to the larger pueblo groups can be collected. CANON-HEAD GROUPS The canon-head groups differ, as has been stated above, from the pueblos chiefly in being scattered aggregations of small 344 A. V. K1DDER houses, rather than many houses or groups of rooms brought together to form a more or less compact, or at least contiguous, whole. The component buildings, none of which are individu- ally of any great size, form, nevertheless, collectively a consid- erable group. In at least three cases, i.e. at the head of Rust- ler's Canon and in the two branches of Ruin Canon, they are only a few hundred yards distant from large pueblos. This fact suggested to us that they might have been block-houses or watch-towers to guard the springs which at one time certainly existed directly below the buildings and which must have formed the chief water supply of the near-by pueblos. A fourth group, although a small one and not directly at the head of a canon, guards Cave Springs, a locality which, from its abundant supply of water and its propinquity to several large ruins, must have been strategically very important. Further evidence that tends to strengthen the theory that these structures were fortifications rather than regular dwelling- places is offered by the fact that they do not often seem to con- tain kivas, have no well-defined burial places, and are almost all built on the edge of the rim-rock, on the tops of large boul- ders, or in other easily defensible places. They are now so badly ruined, however, that little can be said of their original ground plan or architecture. Excavation would conclusively prove whether or not they are of the same culture as the other remains of the region. From the potsherds found about them there seems no reasonable ground for supposing them to be the work of a different people or a different period. PUEBLOS The larger pueblos are nearly all to be found on the tops of the cedar-covered mesas between the canons. We mapped over twenty good-sized groups in a small section of Alkali Ridge alone, as well as a very large settlement above the head of Rustler's Canon. These pueblos are so badly ruined that they are now merely low mounds thickly strewn with fallen building- stones and heavily overgrown with sagebrush and greasewood. They are usually situated on the crest of a ridge some distance back from the ruins of the canons, thus occupying the highest EXPLORATIONS IN UTAH 345 ground in the immediate vicinity with a view out over the cornfields that must once have surrounded them. The smaller pueblos always seem to have consisted of a sin- gle or double row of rooms running roughly east and west, with one or more kivas, which appear as shallow circular depressions 15 or 20 feet across, lying just to the south of them. To the south of the kivas again is found the cemetery, a low mound thickly covered with potsherds. The larger ruins are merely multiplications of the unit just described, with a corre- spondingly greater number of kivas and cemeteries. As the burial mounds are unfortunately quite obvious, they have been much pillaged by " pot-hunters," relic-seekers, and other vandals, who, digging carelessly, have broken fully as much as they have recovered, and who have also entirely de- stroyed the skeletal remains. We were fortunate, therefore, to find for our excavation a large ruin with two burial mounds, one of which had been only partially dug over, while the other one was practically untouched. This pueblo, quite typical of the larger examples of its class, is built upon a cedar-covered ridge some 200 yards from the head of one of the terminal branches of Ruin Canon. At that point there is a large canon-head group, which must have protected the water supply for the community and was perhaps built for that purpose. The pueblo itself is a strag- gling structure of many wings and additions, 500 feet long and, at its widest part, about 300 feet across. The wings, it will be noticed (Fig. 2), run for the most part east and west; the kivas lie to the south of the two largest lateral wings. We made our camp at Cave Springs, a mile and a half west of the ruin. The water of that spring is fresh, cold, and abun- dant, and there is ample feed for a few horses in the canon be- low. We remained at this ruin for nearly five weeks, laying bare seventeen rooms and three kivas. We also completely dug over the two burial mounds and spent several days on the mounds of two small ruins to the east. The task of excavation was slow and arduous, as we were unable to secure proper picks to pry out the quantities of tightly packed fallen stones which filled the rooms and kivas; and having only one wheelbarrow we were often forced to handle 346 A. F. KIDDER FIGCRE 2. PLAN OF RUIN ON ALKALI RIDGE. EXPLORATIONS IN UTAH 347 our back dirt two or even three times. The results were, nevertheless, fairly satisfactory; we procured about four hun- dred museum specimens, among them thirty pieces of unbroken pottery, besides many pieces in fragments, which Professor Cummings has since successfully restored. We also recovered a considerable series of crania and other skeletal remains. A report on these last by some competent somatologist will, we hope, be presented at an early date. Digging was begun at the east end of the northernmost wing (Fig. 3), and here we occupied ourselves for nearly two FIGURE 3. SOUTHEAST END OF NORTH WING. weeks in clearing rooms and kivas. We not only emptied the rooms themselves, but ran trenches all along the outside of the house, laying bare the walls to their foundations. Even in the best preserved sections these walls did not stand, when exca- vated, to a height of over 4 feet, but such large quantities of fallen building-stones were present that it seems safe to assign a height of two stories to the entire building. On the other hand, its narrowness throughout argues against the former pres- ence of more than two stories; so that we seem to have here a fairly low and much spread-out structure which must have been 348 A. V. KIDDER quite different in appearance from the more common, terraced type of pueblo, which was compact in ground plan and rose to a considerable height. Such a village as the one under dis- cussion could not have been easily defended. The need for defence, apparently not so keenly felt here, was met in the Mesa Verde district, in the Caiion de Chelly, and elsewhere, by building in caves and on ledges difficult of access and easily defensible ; while the pueblos of the McElmo were placed upon the edges of precipitous rim-rocks, their otherwise unprotected mesa or back wall being high and without ground- floor doorways. 1 In the Chaco Caiion, as well as in other parts of the Southwest, pueblos, where built in the open, are made safe from marauders by their compact form. In this case, how- ever, the buildings are in no way protected by the configuration of the land, and the various component wings are so loosely strung together that no combined resistance to a sudden attack could have been made. Living-rooms. The excavation of the living-rooms gave very little insight into the minor features of the architecture of the pueblo. As only the lower courses of the walls were standing, we were unable to recover any evidence as to the system of roofing or the method of door construction, while the floors could only here and there be made out. They seemed to be, as usual, of hard-packed adobe. The plastering, too, had almost entirely disappeared from the walls. The rooms were fairly uniform in size, averaging about 10 feet long by 5 feet wide. The easternmost chamber, however (Fig. 4), which had apparently been used as a granary, was longer than any other that we observed (23 feet). A violent conflagration had raged in this room, oxidizing a large quantity of corn on the cob. The heat of the fire had been great enough to vitrify, arid in some places even to turn into a sort of irides- cent slag, parts of the adobe of the walls and ceiling. Frag- ments of black-and-white pottery had been burned to a reddish yellow color, the black paint becoming a rich brick-red. The body of the ware was greatly hardened and in spots vitrified. Such conditions as this may possibly account for the rumors 1 S. G. Morley, ' The Excavation of Cannonball Ruins,' American An- thropologist, Vol. X, N.S. 1908, p. 597. EXPLORATIONS IN UTAH 349 which one hears in the Southwest of the finding of cliff- dwellers' remains imbedded in volcanic ash or lava, rumors that are sometimes quoted to prove the immense age of the prehistoric period. The masonry throughout the pueblo is much inferior to that of the buildings of the McElmo and the cliff-dwellings of the Mesa Verde. Little attempt had been made to shape the stones, Plan of Excavated , s Portion of \ Alkali Ridge Ruin I Scale /"= 30' N FIGURE 4. PLAN OF EXCAVATED PORTION- OF RUIN. the rough blocks and fragments, quarried in the near-by canon, being merely hammered or cracked out and laid up in adobe with scarcely a semblance of coursing. In several places cedar posts were incorporated in the lower parts of the walls. They were driven several feet into the ground, their upper ends sunk in 'he masonry. The lower parts of these stakes had been sharpened by fire, their charred portions and the marks of their bark in the adobe of the walls being usually the only evidence of their former presence. All roof beams and other objects of wood were reduced, unless charred, to mere reddish streaks in the earth. The finds in the living-rooms were very meagre. Because of their extreme dilapidation and from the fact that no "mano" or " metate " were unearthed in them, it seems not unlikely that 350 A. V. K1DDER that portion of the pueblo had been deserted and all such utensils moved to some other place. Kiva. It will be seen by consulting the plan of the pueblo (Fig. 2) that its forty or more kivas are very evenly distrib- uted among the rooms in a proportion that may be roughly estimated at one kiva to seven ground-floor chambers. The majority of the kivas lie to the south of, and immediately con- tiguous to, the groups of rooms to which they belong. In making the plan, only large and well-marked circular depres- sions were called kivas, and it is possible that there exist many small examples of the intramural type (Figs. 4, 6, arid 7), which, before excavation, could not be distinguished from an ordinary dwelling-room. The greater number of the ceremo- nial rooms of this ruin, and, so far as we could determine, of all the sites in the Montezuma drainage, are structurally quite independent of the buildings to which they belong. They are not enclosed in a square or rectangular walled-up space, as was found by Mr. Morley to be the case on the McElmo, 1 nor are they set among the rooms as they are in the cliff-dwellings of the Mesa Verde. 2 This fact is quite in accord with the straggling and loose-knit plan of the Alkali Ridge ruin. As is the rule in the San Juan, the kivas are subterranean. They are round, and, like those of the neighboring regions, have a plain lower wall some 3^ feet high, surmounted by six pilas- ters which divide the space above the lower wall into six niches (Fig. 5). These pilasters also served to carry the entire weight of the roof. 3 They are usually about 2| feet high, thus making the ceiling of the chamber approximately 6 feet above the floor. The outside of the roofs of these subterranean rooms apparently formed a kind of plaza, which was on about the same level as the floors of the living-rooms. There is a slight inward trend of both walls and pilasters. The kivas here average about 16 feet in diameter. Of the six niches or recesses, which are divided from each other by the pilasters, the south one is always the deepest and 1 Loc. eft., p. 600, and pi. XXXVII. 2 J. W. Fewkes, Bulletin 41, Bur. Am.Ethnol. 3 W. J. Fewkes, 'Ventilators in Ceremonial Rooms,' American Anthro- pologist, N. S. Vol. 10, 1908, p. 385. EXPLORATIONS IN UTAU 351 FIGDRE 5. NORMAL KIVA. broadest (Fig. 5, d d). The other five are almost exactly of the same width and depth. Under the south recess there runs a horizontal passage 2| feet high by 1 foot wide, roofed with planks of split cedar (Fig. 5, 6). Its floor is a continuation 352 A. V. KIDDER of the floor of the kiva. At a distance of 7 feet from the lower wall, this passage turns upward at right angles and, rising ver- tically just behind the back wall of the south niche, it emerges from the ground at the level of the kiva roof. The vertical passage grows quickly smaller as it rises, until at its mouth it is less than a foot square (Fig. 5, a). In front of the opening of the horizontal passage into the kiva and about 2^ feet from its mouth, there is an upright slab of stone 2 feet wide, 2|- feet high, 2 inches thick. 1 (See Figs. 5, e, 8, and 9.) A line drawn from the entrance of the passage through the centre of the slab and continued across the floor would bisect the other two principal features of interest; namely, the fireplace and the "sipapu." The firepit is a round depression in the floor 3 feet in diameter and 8 inches deep, filled with tightly packed wood ashes. The " sipapu " is a small hole in the floor, barely large enough to admit the hand and 5 or 6 inches in depth. It lies about midway between the fireplace and the back or north wall of the kiva. (See Figs. 5,