NPS: Archeological Excavations in Mesa Verde National Park, Colorado — 1950 (Excavation at Site 16 of Three Pueblo II Mesa-Top Ruins)

Mesa Verde National Park, in the southwestern corner of Colorado, is famous for the hundreds of cliff dwellings spectacularly located in high shallow caves and on narrow ledges of its many canyon walls. These cliff pueblos, with their well-preserved masonry, specialized ceremonial rooms (kivas) and varied structural features are a splendid tribute to the architectural craftsmanship of the people who built them. Such craftsmanship was not acquired in a day, a year, or many years. It was the outgrowth of centuries of trial, experimentation and progress by primitive people working with the simplest of implements, struggling all the while with a not too favorable environment to maintain existence. The story of this progress is as fascinating in its way as the ultimate achievement is spectacular.

Less well known, as they were often disregarded in the early exploration of the Mesa Verde, are hundreds of ruined structures dotting the mesa tops. Unspectacular as compared to the cliff dwellings in their dramatic settings, they are easily overlooked. Yet the story of the architectural progress which culminated in construction of the cliff dwellings of the late Classic Pueblo period (table 1) can be derived only through a study of these mesa-top ruins.

There are few clues to betray the location of the earliest structures. Later remains, however, are more obvious, ranging in size from a small area marked by a handful of sherds or a few pieces of fire-reddened sandstone, through innumerable low, nondescript mounds to large piles of tumbled masonry overgrown with brush and trees. Through excavation of these mesa-top ruins the archeologist is able to trace architectural progress from single family dwellings of the A. D. 400's to great communal structures of the A. D. 1200's.

It is regrettable in the interests of science and from the standpoint of interpretation that excavation of these mesa-top ruins was so long delayed. Comprehension of the progressive stages of advancement resulting in the Classic Pueblo period is needed for appreciation of the cultural achievement of the people, as well as understanding of the architectural features of the structures they built.

The aim and purpose of the National Park Service is a dual one: protection and interpretation. In order to interpret the cliff dwellings, and thereby create for the public a real appreciation for these features and a sympathetic understanding of the imperative need for their protection, it is necessary to be able to exhibit, and thus recreate, each successive step which led to the construction of the cliff pueblos.

Prior to 1948, the interpretive program of this park was handicapped by lack of excavated and stabilized exhibits of the architectural types intermediate between single-roomed, semisubterranean pithouses of Modified Basketmaker times (table 1), and the many-roomed multistoried pueblos of the Classic period. Visitors shown simple houses dating in the A. D. 500's and 600's, then taken to complex villages and ceremonial structures dating in the A. D. 1200's, could not readily grasp the relationship of the two. Without exhibits of the intervening steps, or stages, interpretation was difficult and while museum exhibits helped bridge the gap, nothing could take the place of actual excavated ruins.

For the benefit of the layman, the years between the Modified Basketmaker and Classic Pueblo periods are usually referred to as the Developmental Pueblo period. This simplified interpretive term adequately describes the span of approximately 400 years, from A. D. 700 to 1100, during which Pueblo architecture and culture evolved toward their climax in the Great or Classic period of approximately A. D. 1100 to 1300 (Roberts, 1935, p. 32). The archeologist, using the more precise Pecos Classification, refers to the early part of this period, A. D. 700 to 900, as Pueblo I, and the latter part, A. D. 900 to 1100, as Pueblo II (Kidder, 1927, pp. 554-561). Referring to these dates, it should be pointed out that excavation in the past few years indicates that Pueblo I probably did not start over a large part of the northern San Juan area until about A. D. 750, and while Pueblo II is considered as beginning about A. D. 900, there is a feeling it may not have ended until some time after A. D. 1100.

Many architectural changes characterize Pueblo I and II. Between approximately A. D. 750 and 900, pithouses were gradually deserted in favor of above ground rooms as living quarters. This 150 year interval saw the beginning of the Southwestern Indian "pueblo" (village), in the sense in which the term is used at the present time. Houses were no longer individual units, set apart one from another, for with the introduction of upright, above ground walls it was possible to join rooms together in contiguous rows or blocks, thus forming the pueblo. The people began to experiment with various architectural mediums, combining posts, poles, adobe, and stone in many different combinations.

Actually, the pithouses were never abandoned; they merely underwent a change in character and function. These chambers were made deeper, gradually developing into completely subterranean pitrooms. From about A. D. 750 to 900, one or more pitrooms were built to accompany each group of above ground living rooms. During these 150 years, the pitrooms became invested with more and more ceremonial character, finally evolving into true kivas with the beginning of Pueblo II.

The period from A. D. 900 to 1100 was characterized by extreme experimentation in architecture, especially in architectural style. By approximately A. D. 1000, the people had settled on the almost exclusive use of coursed stone masonry, and toward the end of the period there was a growing tendency to use double-coursed masonry in the construction of compact pueblos. The years A. D. 900 to 1100 saw the development of the kiva from a simple, four-post structure into a highly standardized ceremonial room, closely resembling in most respects the kivas of the Classic period.

The impossibility of attempting to describe, successfully and intelligibly, such a progression without field exhibits is immediately apparent. Interpreter and visitor alike were severely handicapped. Then came the proposal by the Gila Pueblo Archeological Foundation, of Globe, Ariz., to excavate a Developmental Pueblo site in Mesa Verde National Park during the summer of 1947. A permit for the work was granted by the Secretary of the Interior, March 26, 1947. This instituted the first of recent steps undertaken to provide for visitor use and interpretation a series of exhibits of excavated sites representing basic stages in the development of Pueblo architecture and culture.

Gila Pueblo excavated two ruins at the Twin Trees Site, located just beyond the shallow Modified Basketmaker pithouses on the Square Tower House-Sun Temple loop of the ruins road (pl. 1). One ruin at this site is a Pueblo I village of small, surface slabhouses associated with deep pitrooms, dating approximately A. D. 840, while the other is a Pueblo II village, consisting of a 2-room, crude masonry unit associated with a 4-pilastered kiva, dating approximately A. D. 950 (O'Bryan, 1950, pp. 28-43). These two ruins spanned the worst break in the interpretive series: that existing between earthen-walled, semisubterranean pithouses and masonry-walled, above ground pueblos.

These two ruins pointed up the need for other excavated sites which would illustrate further progressive stages toward the development of the multistoried pueblos and complex ceremonial rooms of the Classic period. As has been pointed out, Pueblo II was characterized by extreme experimentation in architecture. However, excavation in the areas surrounding the Mesa Verde has shown that most pueblos built between A. D. 950 and 1100 were, generally speaking, units of a few to several rooms constructed of fairly good masonry. These rooms were associated with one kiva, and sometimes more, which, while often developmental in style, was more advanced than the 4-pilastered kiva at the Twin Trees Site. A step urgently needed for the interpretive program was a pueblo such as described above.

In the spring of 1950, a survey was made of the area east of the Twin Trees site for a pueblo which would seem, on the basis of surface indications, to represent a step in advance of the Twin Trees village. Gila Pueblo Site No. 16 (the Gladwin Survey of Gila Pueblo Archeological Foundation, conducted in Mesa Verde National Park, 1929) was chosen, as a result of the survey and tests, as the best available example in the desired location of a small, masonry, unit-type pueblo which would probably date in the A. D. 1000's. (pl. 1 and 18).

Site 16, at an elevation of 6,875 feet, is located on the east slope of the ridge of Chapin Mesa between Square Tower House, in Navaho Canyon, on the west, and Sun Point, on the canyon rim at the junction of Cliff and Fewkes Canyons, on the east (pl. 1). This location is on the southern edge of an open sagebrush-covered area known as The Glades. The ruin lies 100 feet to the south of the Square Tower House-Sun Temple loop of the present ruins road system and is situated midway between the Developmental Twin Trees Site (O'Bryan, 1950, pp. 28-43), three-tenths of a mile west, and the early Classic Sun Point Pueblo, excavated in 1950 (Lancaster and Van Cleave, "Excavation of Sun Point Pueblo" in this volume), three-tenths of a mile east (pl. 1). The site, therefore, is in the correct chronological position for use in the present interpretive program of Mesa Verde National Park.

The location of the ruin is not a spectacular one (pl. 18), except for a view, to the northeast, of the rugged La Plata Mountains. This view, made possible by a forest fire of an undetermined date which created The Glades, breaks the monotony of the mesa top. A forest of pinon and juniper, the typical mesa coverage, extends south and east from the site, while the open area to the north and west is covered with low-growing sagebrush and yucca. The nearest water supply today, as well as in prehistoric times, is the spring at the head of Fewkes Canyon, one-quarter mile northeast (pl. 1). The land surrounding the ruin, cleared of trees and shrubs, would be adaptable to agriculture.

The site probably was known to the early cowboy explorers as it lay beside their trail up Chapin Mesa from the Mancos Canyon to the south. Either these cowboys or some later explorers paid particular attention to the ruin as the present excavation showed that the trash mound had been thoroughly pot-hunted sometime in the past.

The location of the ruin was noted by the 1910-11 topographic survey party which mapped the Mesa Verde, and the location is shown on the 1915 edition of the United States Geological Survey Topographic Map of Mesa Verde National Park.

The first actual record of the site was made in 1929 by the Gladwin Survey, of Gila Pueblo, Globe, Ariz. (duplicate record sheets and survey map on file, Mesa Verde National Park Museum). At that time it was designated on their survey sheets and map as Site No. 16, and marked with a metal stake bearing the number. A sherd collection was made and notes recorded to the effect that the site probably consisted of 15 to 18 rooms with kiva remains to the south and a possible kiva depression on the northeast corner.

In 1941, Dr. Deric O'Bryan, of Gila Pueblo, working under secretarial permit to collect charcoal and wood specimens for dating purposes, sank a test pit in the well-defined kiva depression at the site (the one "to the south" referred to above, as it was the only kiva depression at the ruin). Specimen material obtained was poor and did not yield dates with sufficiently good correlation to warrant their inclusion in Dr. O'Bryan's recent report on the Mesa Verde (O'Bryan, 1950, appendix A).

In the fall of 1948, [the senior author] tested the kiva. The test indicated the ceremonial room was a well-developed type and large for a Mesa Verde kiva. The structure was masonry-lined, had a southern recess, and was somewhat unusual in that it obviously had 8 instead of the usual 6 pilasters.

Following routing justification and planning, a permit for excavation was applied for on February 5, 1950, and was granted by the Director, National Park Service, February 24, 1950. Excavation was started April 21, 1950, and continued until June 13, 1950. During this period 3 separate Pueblo II ruins, superimposed 1 on the other, were excavated. These ruins were a post and adobe village, a small, single-coursed stone masonry unit pueblo, and a larger, double-coursed stone masonry unit pueblo with three towers and the kiva referred to above (pl. 18, lower photograph). The work was accomplished by two Navaho laborers under the direction and supervision of the senior author. One hundred and ten man-days were spent in excavation at this time.

In the fall of 1950, the trash mound was tested in an effort to establish stratigraphy. One test located a very early type kiva, which was excavated. The excavation of this structure revealed the location of a developmental type kiva, so it was excavated also. This work was done by the senior author and other members of the interpretive staff in November of 1950 and January of 1951. Two days were spent in the excavation of each structure.

The ruin was stabilized, protective shelters were erected over the three kivas, and graveled paths were constructed to and around the various features of the site during the late summer and fall of 1950 and winter of 1950-51 (pl. 56). Self-guiding signs and explanatory exhibits have been installed for the use of visitors unaccompanied by a ranger.

The first step in the excavation at Site 16 was to clear the low mound of a heavy stand of sagebrush and yucca (pl. 18, upper photograph). The appearance of the ruin, stripped of its vegetative cover, was somewhat discouraging as little was left in the way of standing walls, and scattered groups of building stones indicated several small buildings rather than a compact pueblo. The presence of three, masonry-outlined, circular depressions which might be towers, coupled with the fact the kiva was unusually large and apparently well-developed, aroused the suspicion that the ruin was a later type than desired for exhibit at this particular location. However, there were two other possibilities: first, this might be one of the odd structures for which the Mesa Verde area is noted, but which are of little use to the interpretive program unless preceded by a pueblo of more standard pattern; or, secondly, the apparent confusion was the result of more than one occupation of the site. The latter would be desirable as it would offer the possibility of graphic illustration of architectural advancement.

It was necessary to determine at once if the ruin was adaptable at this time to the planned interpretive sequence. Therefore, it was decided to begin excavation of a centralized building group surrounding one of the circular depressions. This method of approach, starting in the middle of a ruin, is not considered an advisable procedure normally, but excavation of ruins for exhibit often presents a unique set of problems and approved archeological methods must sometimes be adapted to circumstances.

The approach chosen proved fortunate. Excavation showed the circular structure to be a double-coursed masonry tower superimposed on a single-coursed masonry unit pueblo, with little original stone work left of either structure. The small, 3-room unit pueblo was superior in every way to the 2-room unit pueblo at the Twin Trees Site and, judging by similar pueblos excavated elsewhere, would date about A. D. 1000. The tower was, in turn, architecturally superior to the pueblo on which it was built and apparently was constructed late in the Pueblo II period, probably toward the end of the 11th century.

The site offered far more than was expected. The unit pueblo followed in type and time the exhibit just preceding it, the Twin Trees Pueblo referred to above. The double-coursed masonry tower indicated the presence of a pueblo immediately preceding in type and time the Sun Point Pueblo, the following exhibit in the interpretive series.

Excavation of the 3-room pueblo and superimposed tower was completed. From now on these structures will be referred to as Unit Pueblo No. I and Tower B (pl. 19). It had been decided previously not to excavate the kiva until all danger of any spring rains damaging its exposed walls would be past, so the next step undertaken was the stripping of the area surrounding Unit Pueblo No. I, as evidence encountered during its excavation indicated the building was superimposed on some earlier structure.

The evidence was as follows: occupational fill continued well below the floor level of the pueblo rooms, and two corrugated jars, apparently used as floor cists, were found in such a position that there could be little doubt as to their use prior to the construction of the pueblo. One was under the northeast wall of Room 1. A tree which grew on top of this wall had to be removed, and the removal exposed the jar directly under the base of the broken wall. The other jar was found below the floor in the southeast corner of Room 1, its mouth flush with a hard-packed clay surface. This evidence pointed to the former presence of post and adobe houses, or jacal structures, such as Dr. Brew found on Alkali Ridge (Brew, 1946); Dr. Martin located in the Ackmen-Lowry area (Martin, 1930; 1938); Mr. Morris discovered in the La Plata area (Morris, 1939); and Dr. Roberts excavated in the Piedra District (Roberts, 1930).

The first operation was to strip the area northeast of Unit Pueblo No. I. Careful troweling located a row of 10 post holes extending outward from under the northeast corner of the pueblo (pl. 19, row G—G), and these holes still held the charred and rotted butts of upright posts. The debris in the area surrounding the posts consisted of burned adobe and bits of charcoal, all that remained of the walls and roofs of post and adobe rooms.

Stripping this northeast area also located a section where occupational fill continued below what would have been the floor level of post rooms. The fill was tested (pl. 19, Test 1) and found to be composed of heavily burned adobe, mixed with which were shreds of Lino Gray pottery. This material continued to varying depths of 22 to 30 inches, indicating the possible presence of a Basketmaker III structure. The test was backfilled as there was no intention of excavating a Basketmaker pithouse at the site at this time.

The next step was to trowel the area southwest of the pueblo, where careful troweling again located two rows of post holes (pl. 19, row H—H—H), with the decayed ends of posts in place. The partial outline of one room could be traced. Considerable depth of occupational fill encountered below the floor level of this room again indicated an earlier occupation. The debris on the floor of the room and surrounding it was heavy chunks of burned adobe packed with chinking stones, showing the casts of fairly large posts.

The remnant of a single-coursed masonry wall was uncovered in this area southwest of Unit Pueblo No. I. This wall when traced, though incomplete, indicated the room had been D-shaped. As the room was undoubtedly contemporaneous with Unit Pueblo No. I, it was designated as Room 4 of that unit. This room was built apparently in what had been another room of the post and adobe village (pl. 19).

The areas to the southeast and northwest of Unit Pueblo No. I were strip-troweled. No more post holes could be located and nothing of importance was found.

Excavation of the two remaining circular depressions, now known to be towers, and a rectangular building to the northeast of Unit Pueblo No. I was undertaken. The tower southeast of the pueblo was designated Tower C, the one to the northwest Tower A, while the rectangular building was designated Unit Pueblo No. II (pl. 19). These structures, like Tower B, were constructed of double-coursed masonry. As Tower A was attached to the southwest corner of Unit Pueblo No. II, the pueblo walls were traced as soon as excavation of the tower was complete. Little was left of the massive walls of the pueblo or the towers, and very few building stones were encountered in the excavation. Everything indicated the structures had been robbed of all possible construction materials. Razing of abandoned structures seems to have been a common practice in prehistoric times.

The next step undertaken was excavation of the previously tested kiva, Kiva 1 as it was designated (pl. 19). This structure proved to be of late Pueblo II origin and somewhat large for a Mesa Verde kiva as it measures 23-1/2 feet in diameter and 10 feet in depth.

Kiva 1 posed quite a problem. Architecturally advanced as it was, it could not be associated with any village other than Unit Pueblo No. II. Where then was the kiva, or kivas, to accompany the earlier villages? The usual location for such a structure, or structures, the area directly in front of Unit Pueblo No. I, was occupied by Tower C. The walls of the tower had not settled, as would be expected had they been built on kiva fill. As the site was excavated for exhibit, it was not possible to carry on extensive tests which might destroy the tower or any other feature. Since there was no depression in the area south of the pueblos or anywhere else at the site to indicate the presence of another kiva, the answer seemed to be that Kiva 1 was a remodeled structure, the last remodeling having successfully obliterated any indication of previous use.

Provided this was the answer, why was the kiva located southwest and to the side of all three villages when the usual location for a kiva in Pueblo II times was to the southeast in front of the village? The only suggestion proposed the site of Kiva 1 as the former location of a deep pithouse or pitroom, possibly both. Rather than expend the effort necessary to dig a kiva, the occupants of the post and adobe village had enlarged and modified the pitroom and the people who built the later two villages followed their example. This might also explain the size of Kiva 1, each remodeling having enlarged the structure to get rid of previous features. While the above explanations were not altogether satisfactory, they would have to suffice, as there was certainly nothing to indicate the presence of another kiva.

With the completion of the excavation of Kiva 1, no further work was done at the site during the summer with the exception of erecting a protective roof over the subterranean structure.

Study of the material from Site 16 was instituted in the fall of 1950, the pottery receiving first attention. With the exception of a few sherds of early wares, the pottery from the site was typical Pueblo II Corrugated and diagnostic Mancos Black-on-white. Nevertheless it presented a problem, as analysis was not successful in establishing any stratigraphy for the three Pueblo II levels. It was felt that in the almost 200-year span of occupation indicated for the site, there certainly had been some advance in the techniques of manufacturing, finishing, and decorating pottery, but the sherds from the various structural units failed to indicate such an advance took place. However, in view of the various levelings and reoccupations, it was thought that the picture presented by the pottery might be distorted. Therefore, trenching of the trash mound was necessary to establish stratigraphy.

It was known the trash mound had been pot-hunted, but it was not until trenching was undertaken that it was realized how thorough the vandalism had been. Stray bones of adult, juvenile and baby burials were encountered, lying together in tangled masses. Holes continuing through the trash to the undisturbed soil beneath indicated the excavators had not been able to distinguish between the occupational debris and the native soil underlying it. There was every evidence of random, repeated and disorganized digging. Still the testing was continued in hopes of finding one undisturbed section. The area just south of Tower C seemed to be the least disturbed, so a test pit was sunk a few feet from the tower. This test struck occupational fill continuing well below what had been the bottom of the trash layer throughout the rest of the mound, indicating the possible presence of a subterranean structure. As this was the logical location for a kiva, the test was enlarged.

The enlarged test exposed three occupational levels in succession. Eighteen inches below the surface a crude bench, or structural ledge, apparently associated with some type of pithouse, was encountered. Thirty inches below ground level the test uncovered a narrow bench, much better defined than the one above, and at 51 inches the well-plastered banquette of a kiva was exposed. As this was the kiva to accompany one or the other, or both, of the earlier villages, its excavation was mandatory.

Kiva 2, as the structure was designated (pl. 19) proved to be a very early ceremonial structure. Posts were employed as roof supports, not pilasters, and as there were 4 this kiva was the prototype of the 4-pilastered kiva at the Twin Trees Site. Without doubt, the kiva was constructed by the occupants of the post and adobe village.

The south wall of the kiva was in poor condition and had slumped. The middle of the three ledges, or benches, described above, turned south where it met this wall, so it was apparent the kiva had cut the unstable fill of an earlier structure. It was decided to take a test through the upper kiva wall to determine what type of room was located to the south.

This test (pl. 19, Test 2) breached the kiva wall east of the southwest banquette post. Back of the banquette the test was carried down to 7-1/2 feet below ground level. At this depth a hard-packed floor was encountered. In order not to damage any room features, trenching was continued above the bench level along the west wall. Thirteen feet south of the kiva the wall turned east, so trenching was discontinued. A secondary test located the wall again at a point behind the kiva ventilator shaft. This testing demonstrated Kiva 2 had cut an earlier pitroom. As Kiva 2 is the earliest structure which can be classified as a true kiva, the pitroom it cut must necessarily be of Pueblo I origin, and is undoubtedly the type of structure found associated with crescentic rows of slabhouses. If the pit is of Pueblo I origin, the crude ledge which was first exposed by the test must belong to a Basketmaker III pithouse.

With winter setting in there was not sufficient time to excavate both the kiva and the pitroom. The kiva was the more important structure at the time, for it was associated with at least the post and adobe village, so the test which located the pitroom was backfilled to preserve the structure for future excavation. It is desirable that this room be excavated as soon as possible for it will add immeasurably to the architectural sequence at Site 16.

Excavation of Kiva 2 was continued, but the work was complicated by the location of Tower C as this later structure had been built partially over the north side of the subterranean room. Tower C was important to the story of the last occupation of Site 16 and had to be preserved, so the problem was solved by leaving a block of the kiva fill in place to support the superimposed walls of the tower (pl. 19, profile; also pl. 23).

The unexcavated portion of Kiva 2 presented another problem. The north side of the banquette of the kiva evidently had been remodeled with the result that it appeared to run straight between the two north banquette posts. However, so much of the north section of the banquette lay under the block of fill that it could not be demonstrated conclusively this was the case, nor was it known if the remodeling involved more than the banquette. It was felt the extent of the remodeling and the reason for its having been done might be solved by tunneling the block of fill.

The fill seemed stable enough to warrant its being undermined without endangering the walls of the tower above. However, such a tunnel would have to be carried for quite a distance. To forestall the possibility of anyone being caught in a cave-in, the floor of Tower C was first breached to relieve the overload and expose as much as possible of the back wall and banquette of the kiva. In this way short tunnels could be brought in from either end of the block of fill to the exposed section and a long tunnel under the heavy fill need not be risked.

The excavation through the tower floor located the back wall, the banquette and the northwest roof support post of the kiva. It demonstrated that only the banquette had been remodeled, not the wall, but no logical reason could be discovered for the change.

The excavation through the floor of Tower C exposed a masonry pillar of roughly shaped stones laid in adobe mortar, arising from beside the banquette of Kiva 2 and ending squarely under the wall of the tower. This was an odd thing to encounter in a kiva, but despite its strange location it looked suspiciously like the ventilator shaft of another kiva. Tunneling the block of fill along the top and in front of the banquette definitely proved this column was the masonry-lined ventilator shaft of a third kiva. It had been constructed in Kiva 2, starting at floor level directly in front of the banquette (pl. 19, profile). The kiva to which it belonged lay to the north and under the other side of Tower C.

During the excavation of Kiva 2 it was noticed that the fill was material from some excavation, and it was supposed the material, despite its drift, probably came from Kiva 1. Instead the material was from the excavation for a third kiva which undoubtedly was built in connection with Unit Pueblo No. I, since the 4-post kiva was of the same constructional period as the post and adobe village, and Kiva 1 was too well advanced to be associated with any village other than Unit Pueblo No. II.

It was never suspected that a kiva to accompany the 3-room pueblo would be built so close to its front wall, and there was nothing other than the newly exposed ventilator shaft to indicate its presence. In view of the location of Kiva 2, this was the only space left, cramped as it was, for the building of another ceremonial room.

Excavation of the third kiva, designated as Kiva 3 (pl. 19), was undertaken since it was an integral part of the second Pueblo II village at the site. Again the work was complicated by the location of superimposed Tower C, so it was necessary to leave a block of fill in this kiva, this time in the south half, to support the tower walls. Kiva 3 was smaller than Kiva 2 and the tower covered more of this structure than it had of the preceding one. For this reason the block of fill had to be cut straight to the floor in order to expose any floor features at all (pl. 19, profile). This so weakened the tower support that it was not deemed advisable to try to tunnel the fill to expose the banquette, for it was feared tunneling would not only endanger the lives of the excavators, but might result also in the collapse of the tower. Instead of tunneling, the breach in the floor of the tower was widened to locate the south wall of the kiva and as many of the southern features as possible.

Kiva 3 proved to be intermediate in type. Pilasters had replaced posts for roof support. As they were 6 in all, this kiva was the next step following the 4-pilastered kiva at the Twin Trees Site. Architecturally the kiva was of the same period as Unit Pueblo No. I, leaving no doubt as to its construction by the occupants of the first masonry village.

With the completion of the excavation of Kivas 2 and 3, the picture as regards the Pueblo II occupation of Site 16 was fairly complete; three villages in superimposed position, each with its own ceremonial chamber. It is known that these villages are superimposed in turn on even earlier structures, and the evidence indicates there were at least five different occupations of Site 16, occupations which covered a period of several hundred years.

An observer cannot help but wonder why this particular location was chosen as a homesite. What was its appeal to the first settler, or settlers, and even more baffling, what made it attractive to successive groups of people? As previously pointed out, the location is not a spectacular one, nor is there a particular view which would offer an irresistible appeal to man's aesthetic sense. A short distance to the north and east are canyon rim locations which offer far more exciting vistas, so appreciation of natural beauty obviously played no part in selection of the site.

The site offered nothing that could appeal to man's sense of comfort. Located as it is on the ridge of the mesa, it is fully exposed to the elements, to the biting winds and drifting snows of winter, to the long hot days and punishing sun of summer. There is little drainage, so the accumulating moisture of melting snow, and of prolonged rains in the spring or fall, would penetrate walls and roofs, bringing discomfort and forcing the expenditure of considerable effort in the repair of homes and kivas.

It is impossible to see that the site offered any particular convenience. The nearest water supply is, today, a quarter of a mile away, and geologic conditions are such that it is impossible for a spring to have existed any nearer than the present one in the head of Fewkes Canyon at any time during the occupation of Site 16. There is no depression in the surrounding area to indicate the people ever constructed a reservoir or catchment basin to conserve rain water or melting snow. The fields which surrounded the village were close at hand, it is true, but this convenience is negligible since a location closer to the head of Fewkes Canyon would offer just as easy access to the farm lands. Furthermore, a site nearer the canyon would be close to a good supply of readily available building materials.

There is, today, no apparent logical reason why the site was chosen by the first settlers. Even more unfathomable is its appeal for succeeding groups of people. Why did people choose to build on the accumulated trash and debris left by their predecessors? A short distance away they could have settled on undisturbed ground where there would be no ruined walls to tear down, no debris to level, and no man-made obstructions with which to contend. What caused, or prompted these people to deliberately ignore better locations in favor of living on someone else's trash?

These questions cannot be answered. Such behavior is illogical, but throughout the history of man there has been a common, worldwide tendency on his part to build and live where others have built and lived before him. Sites in Asia are known, for example, which have been reoccupied, time and again, century after century since the paleolithic age until the last occupation—in some cases a modern village—is situated on top of an artificial hill composed of manmade debris. Epidemics, fires, floods, earthquakes, and wars have driven men from their villages and cities and have, as in the case of the recently bombed-out towns and cities of Europe and Asia, completely destroyed entire settlements, yet men always return to the same places to build anew. Why they should choose to do so is beyond comprehension, yet history, backed by archeology, demonstrates the trait is, and was, peculiar to man the world over. In this respect, the ancient inhabitants of the Southwest were no different from their fellowmen, and Site 16 is a graphic illustration that this trait was present in the Mesa Verde.

Component parts—Remains of post and adobe rooms to the northeast and southwest of Unit Pueblo No. I (pls. 19, 21, and 22). Kiva 2 (pls. 19, 20, and 23).

Taxonomic position—Pueblo II, Pecos Classification (table 1, p. 6). Developmental Pueblo; Roberts' Classification (table 1, p. 6). Mancos Mesa Phase; Gila Pueblo Phase System (table 1, p. 6).

Comparative dating—On the basis of architecture, presumed to have been constructed early in the Pueblo II period. Considered on the basis of two dated ruins at the Twin Trees Site which precede and follow this village in architectural type (O'Bryan, 1950, pp. 28-43), to have been built around A. D. 900. Probably not constructed much before A. D. 900, or later than A. D. 925. For further discussion see Dating of Site 16, pages 77, 78.

This village apparently was composed of several vertical walled, flat-roofed, single-storied post and adobe rooms, arranged in a rectangle, or "L," facing southeast toward Kiva 2. The village burned, as evidenced by the charred butts of wall posts and the fire-hardened adobe used in wall and roof construction. This burned adobe shows the method of wall construction. Heavy layers of the material were packed between upright posts and further compacted by shoving small and large chinking stones into the mud while it was still plastic. Several pieces of this burned adobe show casts of fairly sizable posts.

Only one room outline can be partially traced. This room, beside and just south of the D-shaped Room 4 of Unit Pueblo No. I, is indicated by rows of posts forming the southwest and northwest walls (pls. 19 and 20). The fireplace, in what appears to have been about the center of the room, is the only firepit other than the one in Kiva 2 which can be assigned to this occupation. It is a shallow depression in the floor and, when excavated, was full of ashes. A corrugated jar used as a floor cist was located northeast of the firepit (pl. 20, No. 1). As reference to plate 19 will show, Room 4 of Unit Pueblo No. I appears to be constructed in what was once another room of the post and adobe village. These two rooms indicate either a double row of rooms, or an "L" extension at the southwest end of the village (pl. 24). The small slabs located by the posts on the southwest side of the village do not seem to serve any particular purpose. It is suggested they were used as wedges to force and hold the posts upright until the adobe packed around them hardened.

The former presence of rooms to the northeast, and also beneath Unit Pueblo No. I, is indicated not only by the line of posts extending outward from under the pueblo, but also by two corrugated jars used as floor cists. One jar, found under the northeast wall of Room 1 of the unit pueblo (pls. 19 and 55, lower right), obviously was in position prior to the construction of Room 1. The other jar was found in the southeast corner of Room 1 (pls. 19 and 55, lower middle). However, the mouth of the jar was below the floor level of the masonry room and flush with another floor level, no doubt that of a former post and adobe room. These jars show as Nos. 2 and 3, plate 20.

The post and adobe construction used in this village was the direct successor to the slab, post and adobe construction of Pueblo I times. Large slabs were no longer used as the bases of walls, but in other respects construction was similar. The important difference is seen in the floor levels of rooms: slabhouse floors are slightly depressed, being a few inches below the surrounding ground level, while post and adobe house floors are flush with ground level.

There are two significant differences between slabhouse and post and adobe villages. The rows of rooms in a slab house village are arranged in a crescent, while post and adobe rooms are aligned in rectangular blocks. The other and most significant difference between the two villages is the ceremonial chamber. Slabhouses are accompanied by deep pitrooms located in the area on the south side of the village between the arms of the crescent rows of rooms. All evidence points to these pitrooms being partially domiciliary, as well as partially ceremonial, in character. Post and adobe villages are associated with true, circular kivas, ceremonial in character, located in front and south of the house block.

The best preserved examples of post and adobe, or jacal structures are those excavated by Dr. Roberts in the Piedra District, just east of the Mesa Verde (Roberts, 1930). Dr. Roberts found three types of post houses which he designates as Classes A, B, and C. House construction in the Piedra District apparently followed somewhat different lines of development than in the Mesa Verde, since Class A and B houses, though having certain features comparable to Pueblo I slabhouses of the Mesa Verde, are nevertheless so different as to constitute a separate architectural type. Class C houses, on the other hand, are very similar to post and adobe houses of this region. These houses are usually associated with one or more masonry rooms, however, and the rooms are often isolated units instead of being incorporated in a rectangular house block. The great difference between Class C villages and post and adobe Villages is the ceremonial room, or kiva, of the latter. No Class C houses have been found associated with kivas. Instead, in the location where one would expect to find a kiva, the existing depression marks the site of a reservoir.

Because of the subsequent occupations of Site 16, little remains of the post and adobe structures. After they burned, the site was leveled preparatory to the construction of the first unit pueblo, and most of the debris from the burned rooms was carried away. Plate 21 shows the village as it looked after excavation.

Plate 22 shows the village after stabilization. Post holes in the open immediately fill in and all evidence of their presence is lost. Furthermore, they are never obvious to the average visitor. In order to retain the outline of the former post rooms and thus be able to demonstrate the presence of the village, it was necessary to place short posts in the original post holes. Since the village originally burned, these replacement posts were charred so they would be in keeping with the story. Visitors are always told these posts are not the original ones, but modern substitutes, and the necessity for the substitution is fully explained.

Plate 24 shows the artist's reconstruction of the village as it is thought to have looked when occupied.

General features: earliest type; circular; 18 feet in diameter and 7-1/2 feet deep; 4 posts in the banquette for roof support; a partially masonry-lined ventilator shaft located 2 feet back of south wall; no southern recess; no deflector; deep ashpit; slab-lined, box-type firepit; sipapu in alinement with firepit and ventilator tunnel, no banquette niches; banquette cist between 2 north posts; small amount of crude masonry; no evidence of method of roof construction; orientation 28-1/2° east of south (magnetic reading).

This structure is an excellent example of the earliest type of kiva. Its predecessor, a subterranean pitroom with four roof support posts arising from the floor, was used as a dwelling and a ceremonial room. In Kiva 2 the four support posts have been moved from the floor to the banquette, thus clearing the floor space for the performance of ceremonial rites. Earlier rooms usually contain a certain number of household furnishings, but nothing was found in Kiva 2, such as metates, cooking vessels and the like, to indicate the room had any domiciliary function.

While the structure is crude in many respects, several basic kiva features are present: banquette, ventilator shaft and tunnel, firepit and sipapu, with the last four in a general north-south alinement (pl. 20). Certain features considered typical of well-developed kivas are lacking: pilasters, southern recess, deflector, masonry lining and a niche in the north face of the banquette in alinement with the sipapu, firepit, deflector, and ventilator tunnel.

This kiva has a small amount of crude masonry, apparently added in an effort to stabilize weak sections. The south part of the kiva was constructed in the fill of a Pueblo I pitroom. This fill was unstable, so the builders found it necessary to place a partial lining in the ventilator shaft to keep it from slumping. For this same reason they incorporated a small block of masonry above the mouth of the ventilator tunnel to keep it from collapsing. The face of the banquette between the two north roof support posts is masonry lined to a depth of 1 foot below the edge. The banquette was remodeled during the time the kiva was in use, and the builders probably found it necessary to use masonry to hold the banquette edge in place. Possibly the remodeling of the banquette was done to accommodate the large corrugated jar which was used as a cist (pl. 20, No. 4).

One other strip of crude masonry possibly was incorporated as a stabilization measure. This is a narrow, tapering band, at its widest only a foot high, extending just above the banquette around the north side of the upper kiva wall. The band starts 6 feet to the north of the southwest post and continues around the kiva to a point 3 feet from the southeast post. The test which located Kiva 2 struck a crudely constructed earthen ledge 18 inches below the present ground level. This tapered off to the north, following somewhat the general kiva contour. Presumably this ledge, possibly a bench, indicates the presence of a Basketmaker III structure. If the kiva was partially excavated through another structure earlier than the pitroom to the south, the band of masonry above the banquette may have served to hold back the fill of a large pithouse.

The floor features are few in number. No deflector as present, nor is there any indication of there ever having been one. No doubt the tunnel opening was closed with a movable slab when need arose to cut off the down draft of cold air from the ventilator shaft. A deep cist filled with ashes is located in front of the firepit and between it and the ventilator tunnel (pl. 20, C). A small, shallow cist to the east of this ashpit may have been used as a pot rest (pl. 20, No. 6). The box-type firepit is lined with slabs. The sipapu, 8-1/2 inches deep, is the only other floor feature showing. The block of fill left to support the superimposed walls of Tower C of Unit Pueblo No. II, precludes the possibility of determining any other floor features which might be present (pl. 23).

The banquette, 3 feet 6 inches high and varying from 1 to 2 feet in width, is well plastered. Two upright slabs are set in the banquette between the wall and the northeast post (pl. 20), while a flat slab is embedded in the banquette beside them (pl. 20, No. 5). The deep banquette cist (pl. 20, No. 4) was covered with a flat slab. This cist was formed by sinking a large, oval corrugated jar into the banquette. The jar is illustrated on plate 55, upper right.

The method of roofing Kiva 2 could not be determined as the roof was torn from the structure prior to the construction, to the north, of Kiva 3. It is presumed that construction was similar to that used in roofing deep pitrooms. Excellent illustrations of this method, which employs a basic framework of four stringers laid across support posts, which are in turn crossed by smaller poles, may be found in Morris (Morris, 1939, p. 70, fig. 13); Roberts (Roberts, 1930, p. 44, fig. 9); and O'Bryan (O'Bryan, 1950, p. 43, fig. 16).

When the following kiva, Kiva 3, was to be constructed, the builders were hampered by lack of space between their new masonry pueblo and the abandoned kiva 2 in which to completely construct their ceremonial chamber. Not having space for a ventilator tunnel and the shaft, the tunnel for the new structure was dug under the north side of the banquette of Kiva 2, and the connecting ventilator shaft was completely constructed of masonry within the old kiva. With the ventilating system arranged for, Kiva 3 was excavated and the dirt was thrown into Kiva 2, effectively burying the ventilator shaft and filling the kiva.

Structures similar to Kiva 2 have been excavated in the area northwest of the Mesa Verde by Dr. Paul S. Martin. One of these, House Kiva 1 at Little Dog Ruins, 30 odd miles northwest of Cortez, Colo., is very much like Kiva 2 (Martin, 1930, pp. 27-29, and fig. 2). However, due to the presence within the structure of a number of cooking vessels and baskets of shelled corn, Dr. Martin felt the structure was still used to some extent as a dwelling place, so he called it a "house kiva." Another structure called a house kiva, Feature 1, at Site 1, Ackmen-Lowry area, also some 30 miles northwest of Cortez, seems to be earlier than House Kiva 1, for the banquette is not complete (Martin, 1938, pp. 242-243, and map 6). Both of these house kivas, as well as two others at Little Dog Ruins, are associated with post and adobe villages of Pueblo II origin, and together with Kiva 2 are proof that the true kiva, in the sense in which the term is now used, made its appearance in the Mesa Verde area early in Pueblo II.

MANOS: 2. One flat with single grinding surface, found with corrugated jar under the northeast wall of Unit Pueblo No. I. One convex end-to-end and side-to-side, with single grinding surface, from the firepit of Kiva 2. Both illustrated on plate 35, see upper row, left and lower row, right.

AXES: 1. Single-bitted and side-notched, from the floor of Kiva 2 (pl. 38, upper row, left).

HAMMERSTONES AND PECKING STONES: 1. With corrugated jar under northeast wall of Unit Pueblo No. I (pl. 37, middle row, middle).

BONE AWLS: 4. Three from banquette of Kiva 2, one from floor of Kiva 2 (pl. 41, Nos. 3, 4, 7, and 8, upper row).

POTTERY: 4. Wide-mouthed, narrow-rimmed corrugated jars. Three used as floor cists in rooms of the village and one used as a banquette cist in Kiva 2 (pl. 55, upper left, upper right, lower middle, lower right).

Component parts.—Rectangular, single-coursed masonry pueblo of three rooms, Rooms 1, 2, and 3 (pls. 19, 25, and 26). Remnant of a detached, D-shaped, single-coursed masonry room, Room 4 (pls. 19, 25, and 26). Kiva 3 (pls. 19, 25, and 26).

Taxonomic position.—Pueblo II; Pecos Classification (table 1, p. 6). Developmental Pueblo; Roberts' Classification (table 1, p. 6). Mancos Mesa Phase; Gila Pueblo Phase System (table 1, p. 6).

Comparative dating.—Presumed on the basis of findings at similar sites excavated elsewhere in the Mesa Verde area to date about A. D. 1000. Probably not before A. D. 975 and not later than A. D. 1025. This pueblo is architecturally superior to the Twin Trees Pueblo, Site 102, dating approximately A. D. 950 (O'Bryan, 1950, pp. 32-35), but is not as well advanced architecturally as the Site 1 pueblo, located 400 yards south of the Twin Trees, which dates approximately A. D. 1025 (O'Bryan, 1950, pp. 44-51). The masonry, arrangement of rooms and type of kiva, all point to construction of the village in the middle of Pueblo II, late in the 10th or early in the 11th century. See Dating of Site 16, pages 77, 78.

This unit is a typical example of the hundreds of small farming communities which dotted the mesa tops and valley floors of the Mesa Verde area in the 10th and 11th centuries A. D. These pueblos were the homes of a few families, perhaps a small clan, and are accompanied by a community ceremonial room, a kiva situated outside and usually south of the house block. Modern counterparts are our own small rural communities, a few farm houses nestled close to a country church.

The main house block, 21 to 23 feet in length and 16 to 16-1/2 feet in width, contains 3 rooms. The large room across the front, Room 1 (pl. 25), undoubtedly served as living quarters since the only inside firepit, a slab-lined box, is located in this room. The two small rooms at the rear, Rooms 2 and 3, quite possibly were used for storage. The one isolated room—the somewhat D-shaped Room 4—may have been built after the main pueblo was constructed, possibly when need arose for another room.

Single-coursed stone masonry was used in the construction of the walls. The courses are fairly even but the stones used vary greatly in size. The method of dressing these stones is characteristic of this stage of the period. Rather flat sand stone slabs were selected, roughly shaped by chipping or knocking the edges from both sides, with the result the inner and outer faces of the blocks appear V-shaped in cross section. Because of the V-shaped edges it was necessary to employ considerable mortar in laying up the courses, and the walls of pueblos of this period usually were not too strong. The builders sometimes incorporated upright posts in a wall to give it greater stability, as they did in the front wall of Room 1 of this pueblo (pl. 25).

It is doubtful if the walls of the pueblo were ever more than one story high. Certainly Room 1 was only one story in height for the front wall of the room fell into the kiva sometime after abandonment, and not enough stone was present in the kiva to account for more than a portion of a one-story wall. Furthermore, the walls appear too thin and too weak to have supported a second story in any section. A positive statement cannot be made for the ruined pueblo was leveled by the last occupants of the site and any remaining building stones were disposed of elsewhere.

Nothing is known of the method of roof construction but it certainly was no different than that used in other masonry pueblos. One or two cross beams were placed in the wall at the desired ceiling level, and several slender poles were laid at right angles across these vigas. These in turn were covered with smaller poles or shakes (shingle-like boards), and bark, grass, and the like laid on top. This roof would then be covered with a thick layer of mud tamped in place.

The only inside firepit is in Room 1. A deep ashpit, located beside this firepit, contained ashes, manos, and broken stone slabs. An outside firepit (pl. 25) is located against the front wall of the pueblo at the southeast corner. One floor cist was found which was made by sinking a corrugated jar into the floor of Room 2 (pl. 25). The mouth of this jar was covered with a sandstone slab.

Unit Pueblo No. I was abandoned and fell into ruin before the last reoccupation of the site. The roof timbers of the kiva rotted and the roof caved in, perhaps when the front wall of the pueblo fell on it. The kiva filled and the fill was well-packed before construction was started of the massive walls of Tower C, of Unit Pueblo No. II (pl. 19). It appears that most of the stone used in building this pueblo was carried away for use elsewhere prior to the last occupation as few stray building blocks were encountered during excavation. The stones were not reused in the walls of Unit Pueblo No. II or its towers, for the stones employed in the construction of these last buildings are squarish blocks, quite different from the flat slabs used in the single-coursed walls of Unit Pueblo No. I.

During the last occupation, Tower B was built on the remains of the ruined walls of this earlier pueblo (pls. 19 and 26). Plate 26, upper photograph, shows the pueblo after excavation.

Plate 27 is the artist's reconstruction of the pueblo as it is thought to have looked when occupied.

General features: intermediate type; circular; 16 feet in diameter and 8 feet 3-1/2 inches deep; 6 block-type pilasters for roof support; masonry ventilator shaft located 3 feet back of south wall and inside Kiva 2; no southern recess; presence or absence of deflector unknown; very deep, partially slab-lined, box-type firepit; sipapu; north banquette niche in alinement with sipapu, firepit and ventilator tunnel, banquette cist; face of banquette masonry lined and plastered; walls above banquette of native earth; roof constructed by cribbing; oriented 34° east of north (magnetic reading).

Kiva 3 is a typical example of a developmental kiva and comparison with Kiva 2 will indicate the advances which have taken place. The roof support posts have been replaced by pilasters, and, as usually is the case in most later kivas, these are six in number. Pilasters show cribbing was the method used in roof construction. Masonry lining of the face of the banquette has been introduced. Another feature characteristic of later kivas is now seen—the niche in the face of the banquette in alinement with the sipapu, firepit, and ventilator tunnel.

Comparison with Kiva 1, of Unit Pueblo No. II (pl. 30), will emphasize the developmental characteristics of this structure. The pilasters are narrow, thin blocks, not at all like the massive pilasters of Kiva 1. The southern recess is lacking, a feature which is usually present in later kivas. The wall above the pilasters is not masonry lined, but this is a variable feature in mesa top kivas.

Kiva 3 may or may not have a deflector. This could not be determined due to the block of fill left in the south half of the structure to support the superimposed walls of Tower C. Very likely the kiva does have a deflector, and in all probability it is an upright slab. Because of the supporting column of fill, the only floor features it was possible to uncover are the firepit and sipapu. The firepit is very deep, 26-1/5 inches, box-shaped and partially lined with sandstone slabs. The sipapu, 9-1/2 inches deep, is in the usual location north of the firepit.

The banquette, averaging 32 to 33 inches in height and 18 to 19 inches in width, is faced with stone masonry of rather good quality, better in all respects than that used in the pueblo. The walls of the kiva were covered with many layers of plaster, most of which has peeled off since exposure. A cross section of this plaster shows that tan, brown, white, gray, and reddish muds were used at different times to coat the walls.

The narrow pilasters, 20 to 21 inches in height, are set an average of 5 to 6 feet apart on the banquette. With one exception these are made of thin, narrow slabs of sandstone laid one on top of another to form a pillar. The exception is a remodeled pilaster which looks more like those of later kivas. The remodeling was done to incorporate a small cist in the pilaster.

A cist, 28-1/2 inches deep, is located in the banquette on the east side of the kiva. It is presumed this occupies most of the banquette between two pilasters but due to the block of fill it was impossible to determine the actual length of the cist, even though it was traced back for 49 inches under the fill. A mass of food bones had been tossed in the cist. These are discussed under Foodstuffs, p. 77.

The method of roofing the kiva was easily determined. Six pilasters with rotted timbers lying in position across them and rotted timbers lying on the banquette demonstrated the roof was cribbed. Those interested in this method of roofing kivas will find excellent illustrations in Nordenskiöld (Nordenskiöld, 1893, p. 57, fig. 31) and Fewkes (Fewkes, 1920, pp. 53-55, figs. 53, 54, 55), based on actual kiva roofs still in position in Mesa Verde cliff dwellings.

As previously described, the ventilator shaft for Kiva 3 is constructed as a masonry column in Kiva 2 (pl. 19). The connecting ventilator tunnel passes beneath the banquette of the earlier kiva and enters Kiva 3 under its banquette between the two southern pilasters. In order to determine the presence or absence of a southern recess, the excavation through the floor of Tower C, which located the masonry ventilator shaft in Kiva 2, was enlarged. While it was possible to show the kiva does not have a southern recess, and to locate the two southern pilasters by limited tunneling, it was unfortunately impossible to go further into the fill to locate the deflector or any other features which might be present.

MANOS: 7 (4 complete, 3 broken; 6 with single grinding surface, one used on bosh sides). Four from floor of Kiva 3; 1 flat; 1 flat and wedge-shaped in cross section; 2 are slightly convex side-to-side. Two from the banquette of Kiva 3; 1 flat and wedge-shaped in cross section; 1 convex side-to-side and wedge-shaped in cross section. One from ashpit in floor of Room 1 (pl. 35 upper row, right).

HAMMERSTONES AND PECKING STONES: 7. Five from the floor of Room 4, 1 from the banquette of Kiva 3 and 1 from the floor of Kiva 3 (pl. 37, upper row, entire; middle row, left and right; lower row, left and middle).

PROJECTILE POINTS: 1. Lateral and base notched, from the banquette of Kiva 3 (pl. 39, upper row, left).

PROBLEMATICAL BONE OBJECTS: 3. From the floor of Kiva 3 (pl. 41, Nos. 5, 6 and 7, lower row).

TUBULAR BONE BEADS: 2. Both from the floor of Kiva 3; 1 broken (pl. 42, lower row, right).

POTTERY: 1. Wide-mouthed, narrow-simmed corrugated jar, used as a floor cist in Room 2 (pl. 55, lower left).

Component parts.—Remains of a rectangular, double coursed masonry pueblo (pls. 19, 28, and 29). Three circular, double-coursed masonry towers, Towers A, B, and C (pls. 19, 28, and 29). Kiva 1 (pls. 19, 28, 30, 31, and 32).

Taxonomic position.—Pueblo II; Pecos Classification. Developmental Pueblo; Roberts' Classification. Mancos Mesa Phase, Gila Pueblo Phase System. (See table 1.)

Tree-ring dates.—Two bark dates, both A. D. 1074, from Kiva 1. Material from the 1950 excavation dated by Dr. Edmund Schulman, Tree-Ring Laboratory, University of Arizona (Schulman, 1951, pp. 28-29; Smiley, 1951, p. 23, No. 88z). Eighteen of the 67 specimens Dr. O'Bryan secured from Kiva 1 yielded dates, but so far these dates have not been published (Gila Pueblo Tree-Ring Laboratory, Globe, Ariz., "Tree-Ring Dates for Ruins in Mesa Verde National Park"). List based on the final review of specimens completed in March 1949, furnished the Federal Government, on file at Mesa Verde National Park Museum.

Comparative dating.—On the basis of pottery, predominantly Mancos Black-on-white and allover, Variable Corrugated, the ruin is assignable to Pueblo II. While architecturally approaching Pueblo III in some respects, the majority of features still point to the pueblo, towers, and kiva being of Pueblo II origin. As no Mesa Verde Black-on-white was found at the site, occupation did not extend into Pueblo III. See Dating of Site 16.

So little is left of the demolished pueblo that no conclusions could be reached concerning its original size or outline. Plate 28 shows the extent to which it was possible to trace the walls and the existing outline suggests there may have been two rectangular blocks of rooms, as only one other explanation, which will be dealt with later in this section, would seem to explain the eastern extension of the southeast wall. The area to the north of this wall is strewn with building stones but no section of wall outline could be picked up. The outlining walls of the northwest room block were easily traced up to the point shown, then disappeared completely. One cross-wall in the block is fairly distinct, but there is only a bare suggestion of a second.

There is a feeling that the people who built this pueblo were preparing for the possibility of attack. They seemingly were not at all familiar with double-coursed masonry and they did not bother, or were too hurried, to use any care in alining the walls or paralleling the inner and outer courses of masonry. Plate 28 illustrates how the walls bulge in one place, pinch-in in another, varying erratically from 14 inches to almost 5 feet in width. The lopsided outline of the pueblo could be due to an inability of the builders to follow a straight line, but in view of the care they used in the construction of Kiva 1, this seems illogical. They apparently were concerned with the defensibility of their pueblo, not its appearance. There is the possibility, of course, that Unit Pueblo No. II is a good example of the extremes to which experimentation was carried in Pueblo II. However, the evidence of the double walls, the obviously hurried construction, and the presence of the three towers, all seem to point to defense construction, which in itself was experimental.

The masonry used in the construction of the pueblo and the towers associated with it is the forerunner of that used in the Sun Point Pueblo (Lancaster and Van Cleave, "Excavation of Sun Point Pueblo," in this volume) and later in the walls of Sun Temple (Fewkes, 1916a), structures of the Classic Pueblo period. Heavy blocks of sandstone were selected, roughly squared, and the inner and outer faces of the walls were then built of these stones. The space between was filled with dirt and broken blocks of stone. The faces of some of the stones used in the masonry, especially in Towers B and C, were pecked to even the surface, but the stones showing the "dimpling" resulting from this pecking appear crude in contrast to the well-finished blocks used in the Sun Point Pueblo tower. Undoubtedly Unit Pueblo No. II and its towers are some of the earliest structures built which employed pecked building stones, and it is interesting to find this trait appearing late in Pueblo II. Dr. O'Bryan lists the trait as new for the Montezuma Phase (late Pueblo III) (O'Bryan, 1950, pp. 110-111), but Site 16 proves the method was introduced in Pueblo II, and the Sun Point Pueblo shows the technique was well established and fully developed in early Pueblo III times (Lancaster and Van Cleave, "Excavation of Sun Point Pueblo," in this volume).

A suggestion of a masonry wall was found adjoining the house block on the northeast side, and a similar wall once adjoined Tower C. Between the towers and the kiva, a curving row of stones (pl. 28, row X—X') may be the outline of a single-coursed masonry wall, or it may be the remains of a small terrace built as a retaining wall to divert drainage from the kiva roof.

Towers A, B, and C are considered the most significant structures excavated at Site 16. Two opposite theories are advanced in explanation of the circular towers of the Mesa Verde region and there are excellent arguments to support either theory. One theory explains the towers as lookouts and defensive structures, while the other assigns a ceremonial usage to these buildings. These two explanations are not incompatible when one considers that most Mesa Verde towers are located close to kivas, and in many instances connected to them by means of an underground passageway. It is entirely possible that the towers were introduced as defensive structures and with continued use and because of location they became invested also with ceremonial significance.

The towers at Site 16 are considered to be lookouts and defensive structures, and arguments for their use as such will be presented here. Inasmuch as the later kiva-tower complex, not present at Site 16, was encountered at the Sun Point Pueblo, a structure of the Classic period, a discussion of the ceremonial significance of kiva-tower units is presented in the report on that site (Lancaster and Van Cleave, "Excavation of Sun Point Pueblo," in this volume).

It seems significant that towers appear with the introduction of double-coursed masonry walls and just prior to the construction of compound pueblos. (In a compound pueblo the court in front of the house block, where the kivas are usually located, is bounded on the east, south and west sides by a wall or a row of rooms, thus making it an integral part of the pueblo. Enclosing the court served to wall in the previously exposed and unprotected kivas.) If the towers are considered as watchtowers and defensive structures, then their sudden appearance in the Mesa Verde area no doubt heralds the beginning of the troubled times which led to the construction of compound pueblos early in the Classic period and resulted finally in the retreat, sometime around or shortly after A. D. 1200, to the shallow caves and narrow ledges of the precipitous cliffs of the Mesa Verde, and the rocky canyon heads of the McElmo and Montezuma drainages to the north and west. The change in the architectural pattern, followed by a withdrawal to less desirable but more defensible home sites, certainly seems to point to enemy pressure. Whether the pressure came from inside the pueblo group or was applied by alien peoples, the fact still remains that the people apparently were forced to change their manner of living to adapt their lives and their homes to the possibility of attack.

The location of towers by kivas seems particularly significant. A raid launched when the men of the village were in the kiva would have a good chance of success, for men rushing from the subterranean structure to the defense of homes and families would present perfect targets as they came through the hatchway in the roof; or if caught in the underground room, they would be helpless. Probably because of this, Pueblo II kivas were connected frequently by tunnels to rooms in the pueblo prior to the construction of circular towers; a step taken, no doubt, to give the men a chance of escaping the kiva trap. Later, by building a tower close to the kiva and posting a lookout, the men could be given sufficient warning of an impending raid. Still later, kivas were connected to the towers by tunnels, thus not only providing a means of escape, but also access to a more defensible position.

This use of towers for lookout and defense has been discounted on the basis that the amount of fallen stone around most structures indicates the buildings were seldom more than a story in height, and that such structures would not afford much of a view and offered little advantage as a defense position. This argument does not take into consideration the pueblo practice of robbing abandoned structures of their building stones, so that lack of any appreciable quantity of stone today is not necessarily indicative of low buildings originally. Secondly, it does not take into consideration the willful destruction of these towers in historic times, an outstanding example of which is the "Triple Walled Tower" of the McElmo. Visiting the ruin in 1875, Mr. W. H. Holmes found the walls of one section standing to a height of 12 feet (Holmes, 1878, p. 399). Dr. Fewkes, examining the ruin in 1917, found the tower so destroyed he was unable to determine whether it had been circular or D-shaped (Fewkes, 1919, p. 22). Neither of these men would recognize the structure today; the outline is difficult to trace and there is nothing to suggest that the walls ever stood more than a foot or two in height. Since ruined towers are known which have walls standing 10 to 25 feet, it is reasonable to suppose they once had considerable height.

It is felt the argument against the use of towers for lookout and defense may be influenced somewhat by the modern conception of what constitutes a watchtower or defense structure. It is not argued that the towers were constructed to give a lookout an uninterrupted view over great distances. Instead, their purpose was to afford a position which commanded a view of the immediate area surrounding the village or kiva, and for this purpose no great height was necessary. It is also argued that, to meet an attack on the village, towers afforded the best possible defense position under the circumstances. Probably the towers, like those in the Hovenweep District, had parapet walls, so that men stationed on the roof of a tower, even though it might be but little more than a story in height, had a decided advantage over the raiders approaching on foot.

In 1928, Dr. Paul Martin excavated two pueblos in the Ackmen-Lowry area northwest of Cortez, Colo., which it is felt prove the sudden appearance of an enemy threat, and illustrate the steps taken to meet that threat. These villages, Units I and III on the Herren farm, were originally unit type structures built in Pueblo II times, changed to compounds during occupancy and abandoned sometime after the beginning of Pueblo III (Martin, 1929, pp. 11-25, and pls. IIb and VII; 1939, pp. 474-476, and maps 27 and 28a).

The Pueblo II kivas which were excavated are connected to rooms in the pueblo by means of underground passageways. Evidence throughout the Mesa Verde area points to the fact that construction of tunnels to connect kivas and rooms preceded the introduction of circular towers. The ground plans of the pueblos indicate the towers were built prior to the addition of the compound walls, and the towers also are connected to the kivas by tunnels. The compound walls, roughly constructed, were built after the rooms and towers as evidenced by the joining of the masonry (Martin, 1929, p. 22). These walls were built without regard to plan, but in such a way as to connect the separate structures of the village, leaving the space between the houses, towers, and kivas an open court.

Entrance to the houses apparently was gained through hatchways in the roofs as Dr. Martin was unable to find any evidence of doorways in the pueblo rooms, whereas the towers all had doorways opening onto the courts (Martin, 1929, p. 23; 1939, pp. 475-476). It is felt that both doorways and tunnels to the towers gave the men access to the structures from either the kivas or the courts, as well as giving them a chance to leave the towers for the houses if the need arose. With the only entrance to houses being through the roofs, the ladders giving access to the roofs could be drawn up in case the court was invaded, providing an added measure of security for women and children, and for the men in case they were fighting from the roof tops.

The question has been asked, if the towers were used for lookout and defense, why were they always round like the kivas, not rectangular like the ordinary rooms? This is a good point which raises the obvious question, were the towers always round? Certainly they were not always round in the Hovenweep District to the west, where some of the most striking structures are the high-walled, rectangular towers which occur, sometimes side-by-side with D-shaped and circular towers. If square or rectangular towers were built in the Hovenweep it is logical to presume they were constructed elsewhere in the Mesa Verde region. Why have they not been reported? The answer seems to be that since their ruined walls would differ in no way from those of other rectangular structures, they have been overlooked. In case of the poorly preserved structures of the mesas and valleys of this region, no one seems particularly impressed with the lack of fallen building stones around a rectangular structure. Whereas they are immediately impressed with the lack of stones around a circular structure.

As mentioned above, a kiva of a unit pueblo is sometimes connected by a tunnel to a room in the village. It is possible this room was a tower room, a few feet to a story higher than the other rooms of the pueblo. A rectangular tower room, incorporated with other rectangular rooms would not be obvious, whereas a circular tower room, such as seen at Pipe Shrine House on the Mesa Verde (Fewkes, 1923, p. 98 and fig. 90), or at Unit III on the Herren farm (Martin, 1929, pl. VII; 1939, map 27, Tower B), is immediately apparent. It is felt the first towers of the Mesa Verde area probably were rectangular, and possibly were 2-story structures an an otherwise 1-story pueblo. As these did not offer an unobstructed view, detached structures soon supplemented them. Someone may have seen the advantage of a round room over a rectangular one with its corners and four distinct sides, so the idea of constructing circular towers was introduced and caught hold.

There is nothing to prove or disprove the use of rectangular rooms as towers, but if circular rooms connected to kivas are considered to be towers, it is just as plausible to regard rectangular rooms connected to kivas as towers.

The towers at Site 16 are considered to be lookouts and defensive structures. The massive walls of the pueblo lend weight to this theory as they give every evidence of having been constructed by a group of people who were either too harassed to use care in building them, or who were employing, with an eye to defense, an architectural style with which they were totally unfamiliar. Yet Kiva 1 is proof the builders were capable of careful work if they so chose.

The grouping of the towers may be significant. As plate 28 shows, only Tower A is attached to the pueblo, and like the pueblo is poorly constructed. The other two towers are more uniform in shape and are constructed of better masonry employing a number of pecked building stones, indicating they were built at a later date. If the work area was an front of the pueblo, as is usually the case, the construction of two extra towers in that area may point to an effort to furnish protection for a number of people should a hurried need arise.

It may be, though it is impossible to demonstrate, that this pueblo, like those to the northwest on the Herren farm (Martin, 1929; 1939), was converted to a compound. Little is left of the pueblo and tower walls (pl. 29) as the village was demolished for its building materials. If a compound wall existed, it could have been completely torn out since there would not be the accummulation of roof and wall filler material around the base stones which would bury them, as there would be in the case of the roofed pueblo and towers. It is not proposed this pueblo was a compound, but if such a wall existed it might explain the otherwise puzzling extension to the east of the southeast wall, previously mentioned, and might also explain the stub ends of walls against the northwest side of the pueblo and Tower C.

General features: large, well-developed type; circular; 231-1/2 feet in diameter and 10 feet deep; 8 heavy, flaring masonry pilasters used for roof support; ventilator shaft with small amount of masonry lining located about 3 feet back of south wall; southern recess; slab-type deflector; slab-lined, box-type firepit; sipapu; north banquette niche in alinement with sipapu, firepit, deflector and the ventilator tunnel; 2 oblong and 2 circular floor cists; 6 cylindrical floor holes; 1 potrest (?); 1 post hole; 3 banquette niches (including the north niche); masonry lining below banquette, and above banquette to height of pilasters with the exception of the southern recess; roof constructed by cribbing; oriented 32° east of south (magnetic reading).

Kiva 1 is a well-developed structure and comparison with Kiva 3 will demonstrate the architectural strides which have been made. The southern recess is present, bringing the kiva to the pattern favored in Classic times. The pilasters have developed into substantial columns showing little resemblance to the flimsy blocks used in Kiva 3. The masonry lining has been extended to include the upper walls to the height of the pilasters with the exception of the southern recess. Plate 30 illustrates these features.

However, as comparison with Classic Pueblo III kivas will show (pl. 34, step 6), Kiva 1 still exhibits certain developmental characteristics. It is a good deal larger than later kivas, a circumstance which required the use of 8 instead of the usual 6 pilasters. The ventilators of late kivas are usually built directly in back of the southern recess, using the south wall of the kiva as the north side of the shaft. The greatest difference is the number of floor features. Classic kivas have few if any floor features other than deflector, firepit and sipapu. They do not exhibit a variety of floor features such as are present in Kiva 1.

These floor features in Kiva 1 are interesting (pl. 30). The two oblong cists, Cists 2 and 3, one on either side of the firepit, are located in the same position as the floor vaults, or boxes of Chaco-type kivas, but there the resemblance ends. These shallow cists are merely hollowed out of the caliche floor and plastered with adobe; and Cist 3, east of the firepit, was sealed over during the time the kiva was in use. Later, two cylindrical holes, Holes E and F, were sunk into the sealed cist. Besides the 2 oblong cists, there are 2 circular ones—Cists 1 and 4—both of which were filled with brown sand.

There are six cylindrical holes in the floor of the kiva which vary from 3 to 7-1/5 inches in depth and from 3-1/2 to 6 inches in diameter. There is no formal arrangement and all but one of the holes were filled with brown sand. The one hole was filled with small chunks of roof adobe.

Other floor features are not unusual. One post hole was found just west of the sipapu, with the rotted end of the post in place. The firepit is like those of the preceding two kivas, a slab-lined box. The slabs forming the box stand 4-1/2 inches above floor level. The groove in the floor which held the upright slab deflector is very shallow and as no remains of the deflector were found, a substitute slab was placed in the groove when the kiva was stabilized for exhibit. This makes it possible to demonstrate the type of deflector to visitors, who are told of the substitution.

The banquette averages 41 inches in height and varies from 2 feet in width on the north side of the kiva to 1-1/2 feet in width toward the southern recess. The 8 masonry pilasters, averaging 28 to 29 inches in height, are set approximately 6 feet apart on the banquette. These pilasters flare from front to back so that the "side walls of the pilasters . . . (are) . . . as the outer sections of radii of the circle" (Brew, 1946, p. 210). They average 19 to 20 inches in width at the banquette edge and 2 feet in width where they abut the outer wall of the kiva. They are set back approximately 2.4 inches from the edge of the banquette. There are three banquette niches, the one on the north being in alinement with the sipapu, firepit, deflector, and ventilator tunnel.

The masonry lining of the kiva is entirely different below and above the banquette, as plates 31 and 32 illustrate. The banquette is faced with fairly well shaped, sometimes surfaced, stones laid in regular courses in a minimum of adobe mortar chinked with small spalls. The masonry above the banquette to the height of the pilasters is crude by contrast. Broken, unfaced, irregular stones are laid in a quantity of adobe mortar. The walls were originally heavily plastered, and the same brown plaster used on the walls was applied in a thick layer over the caliche floor.

The southern recess is not masonry lined, though it was plastered. In the back of the recess, directly above the ventilator tunnel, is a block of masonry 33-3/5 inches high and 20-1/2 inches wide. An examination of the ventilator tunnel and shaft proved that the south kiva wall had not been cut in constructing the shaft, for the underlying caliche is still in place, so there is no apparent structural reason for the masonry.

The walls of the ventilator tunnel are native earth. The stones of the banquette above the mouth of the tunnel were held in place by small poles which rotted, letting the masonry slump, as seen in plates 31 and 32. The rest of the tunnel under the southern recess apparently was roofed over with shakes, or split timbers. The ventilator shaft, located approximately 3 feet back of the southern recess, has a small amount of masonry on the south side.

The roof of the kiva was built by cribbing, and, as the kiva is large, 8 instead of 6 pilasters were needed to support the roof. The roof was torn from the kiva, as evidenced by the adobe casts of the heavy timbers encountered on the kiva floor. When the logs were wrenched from place the small poles, twigs, bark, etc., used to fill spaces between the timbers, slid onto the banquette and piled up over the pilasters. Later this material burned, possibly when the kiva was used as a dump as evidenced by the fill, and it was this burned filler material which yielded the dates cited above. The material obtained by Dr. O'Bryan came from the banquette between two pilasters on the northeast side. The material dated by Dr. Schulman came from the banquette on the northwest side.

While the significance of the dates is discussed in detail later on (see Dating of Site 16), it should be pointed out the architectural evidence is in keeping with the dates secured by Dr. Schulman. Very likely the two 1074 bark dates, though from small specimens, reflect very closely the construction dates of the kiva.

One detail, the importance of which is often passed over, may have considerable significance. This is the filling of floor features, such as cists, holes, and sometimes the sipapu, with sand. That the filling was deliberate and not the result of rain water deposition after abandonment, is proved in the case of Kiva 1 by the fill of Cist 2, Hole B and the sipapu (pl. 30). These features did not contain sand but were filled with roof adobe.

The use of sand to fill floor features was a common practice in the San Juan Pueblo area from Basketmaker III through Pueblo II, and continued occasionally, as at least one instance shows, into Pueblo III. Two explanations of the practice are usually accepted: (1) holes so filled were used as pot rests, the soft sand supporting the rounded bottoms of cooking vessels, thus keeping them upright; (2) the holes were used infrequently (use unknown), so when not in use were filled with sand to add to the already somewhat restricted floor space of the room (Brew, 1946, pp. 156-157).

It is true that these sand-filled holes were used as pot rests in some cases, for Mr. Morris has found vessels sitting in sand-filled cists (Morris, 1939, p. 60). A logical explanation for the large, circular, sand-filled cists located to one side of the firepit in so many Basketmaker pithouses would be their use as pot rests. It is difficult, however, to visualize a hole 3 inches in diameter, such as found in Kiva 1, serving such a purpose. It is felt there is a definite reason for filling the holes with sand, and it was not that of adding floor space when the holes were not in use.

If the various authorities are correct in assuming the partitioning of most pithouses and later pitrooms was done to set aside the southern part of the chamber for domestic functions, a women's section so to speak, and the northern part was reserved as general living quarters, there may be a good explanation for the floor features of the north section of the room. Coincident with the explanation for the partitioning is the assumption that the male head of the Basketmaker family was also its religious leader. Having no family or village ceremonial room, he conducted such rites as he deemed necessary for the well-being of his family in his home, using the northern, or general, living section. The theory is given support by the presence in so many pithouses of a small hole in the floor north of the firepit which is believed to be analogous with the sipapu of the later kivas. On the basis of this and later evidence, it is felt the Basketmaker man used the sand-filled holes in arranging his religious paraphernalia, to hold prayer sticks, fetishes, wands, etc., in an upright position.

Later, when above ground rooms replaced pithouses as living quarters for most of the people, the pitrooms began to assume added ceremonial significance. It is noteworthy, in this connection, that the floors of pitrooms continued to be characterized by a variety of floor features, many of which are filled with sand, while the floors of living rooms of the same period seem bare by comparison. If the floor features served a domestic function, why did they almost entirely disappear from the domestic quarters, yet continue to characterize the partially ceremonial pitrooms?

About A. D. 900 distinct ceremonial rooms, the kivas, replaced the deep pitrooms, and post and adobe, and later masonry houses replaced the rows of slabhouses and storage rooms. Sand-filled holes continued to be used in the ceremonial rooms, but they disappeared entirely from the floors of domestic quarters. This supports the theory given above that the floor features of the pithouses and pitrooms were connected with the religious observances of the people and, for the most part, did not serve any domestic function. The sand-filled holes of the partially and fully subterranean chambers would provide excellent supports for such slender articles as listed above, prayer sticks, wands, etc.

The practice continued into Pueblo III, as was discovered in 1951 at Fire Temple, a ceremonial structure which is considered to be a Great Kiva built in a cave. It became necessary to stabilize the bases of the walls of Fire Temple to forestall the possibility of any collapse. The senior author, in exposing the bases of the walls, discovered that Dr. Fewkes, who originally excavated Fire Temple, had never reached the actual floor level of the structure (Fewkes, 1916b; 1921). With the assistance of Temporary Ranger Archeologist Francis Cassidy, the senior author completed the excavation, which disclosed for the first time the floor features of this unique building. In addition to the large, circular firepit and two rectangular floor vaults, or boxes, only the tops of which had been uncovered by Dr. Fewkes, there are several other features of interest.

The floor plan of Fire Temple is rectangular, with a high banquette extending across the north side, or back wall of the cave. The firepit is in the approximate center of the floor, and to either side of it lie the floor vaults, 1 to the east, 1 to the west (Fewkes, 1921, p. 79, fig. 95). Between the west floor vault and the west wall of the room lies a large, circular floor cist, never uncovered by Fewkes. In the northwest corner of the structure, back of the circular cist and the western floor vault, a narrow groove or slot is cut into the floor. This starts at the west wall, passes in back of the circular cist and the western vault, then curves back to meet the north banquette wall on a line parallel with the east side of the vault. This narrow groove or slot is not continuous, but is broken by two openings. Where it passes back of the circular cist, and again where it passes in back of the vault, it is discontinued for a few inches, leaving openings which face the center of the cist and center of the vault. The groove is plastered and filled with sand. This sand-filled slot could well have served to hold a row of prayer sticks upright around an "altar," such as is used by the Hopi and Zuni today. The openings gave access to, or a view of the "altar," or accommodated other items of religious paraphernalia, such as medicine bundles, fetishes, etc.

A report of the work carried out at Fire Temple the summer of 1951 will be made by Mr. Cassidy, who assisted Mr. Lancaster in the excavation and stabilization. As the plans of the structure published by Dr. Fewkes are inaccurate, Mr. Cassidy will include surveyed ground plans of Fire Temple with his report.

The following artifact is the only one definitely assignable to this occupation:

MANO: Convex end-to-end, side-to-side and wedge-shaped in cross section. From the floor of Kiva 1. The only other object found on the floor of the kiva was a sandstone concretion, shaped like a ball (pl. 40, middle row, left).

Plate 33 illustrates the artist's reconstruction of Unit Pueblo No. II, its associated towers and kiva.

Evidence—1. Evidence was encountered in Test No. 1, northeast of Unit Pueblo No. I (pl. 19), of what may be a burned pithouse of Basketmaker III origin. As there was no intention at this time of excavating a pithouse at Site 16, the test was backfilled.

3. The test which located Kiva 2, the four-post kiva accompanying the post and adobe village, encountered a section of a roughly constructed ledge, or bench, about 18 inches below the present surface. It appears that the kiva, as well as the Pueblo I pitroom in which it partially is constructed, cut a large, circular structure. There can be little doubt that this subterranean room is of Basketmaker origin since it had been cut by the pitroom prior to the construction of the kiva.

4. Evidence was uncovered that Kiva 3, the six-pilastered kiva associated with Unit Pueblo No. I, also cut an older structure. The northwest wall of the kiva had slumped, indicating it was originally cut into unstable occupational fill. This fill is full of charcoal which shows plainly on the kiva wall. The depth of fill indicated a Basketmaker pithouse and not a Pueblo I pitroom.

5. Pottery analysis shows 10 percent of the sherds from Site 16 to be of Basketmaker III-Pueblo I origin. For discussion of the situation relative to Basketmaker III-Pueblo I pottery in the Mesa Verde, see subsection so headed under Pottery, p. 69.

Evidence.—1. Evidence was uncovered that Kiva 2, the four-post kiva associated with the post and adobe village, was constructed partially in the fill of an earlier structure. Test No. 2 (pl. 19), demonstrated this structure was a deep pitroom of the type associated with crescentic rows of slabhouses in Pueblo I times.

Excavation of this pitroom is imperative. This is the only site encountered to date in the Mesa Verde where a superimposed order of structures so conclusively demonstrates actual progressive steps in architectural advancement, and one of the few sites so far excavated in the Southwest presenting this sequence in such indisputable order.

2. Occupational fill underlying the floor level of the post and adobe rooms southwest of Unit Pueblo No. I probably indicates the location of former slabhouses which would have accompanied the deep pitroom referred to above. The outline of slab rooms was later obliterated by construction of post and adobe rooms in the same location.

3. Analysis of the pottery from Site 16 indicates the use of the site in Basketmaker III-Pueblo I times. See discussion >under Pottery, p. 69.

In the excavation of the large, eight-pilastered Kiva 1, associated with Unit Pueblo No. II, the dirt was removed in layered blocks. This revealed the presence, 6 feet above the kiva floor, of a well-defined firepit. Apparently the partially filled kiva was used temporarily as a camp site long after it was abandoned by the last Pueblo II occupants. It would be interesting to know the identity of the people who camped here, and whether it was during the Pueblo III occupation of the Mesa Verde or sometime after A. D. 1300.

One outstanding development of Pueblo architecture was the small ceremonial room known today by the modern Hopi term "kiva." Actually, two types of kivas were in use in the Mesa Verde area in prehistoric times: the Great Kiva, a specialized ceremonial chamber, the function of which is not clearly understood, and the small kiva, which was the common ceremonial room, the center of ritual observances for a religious society or a small village.

Great Kivas were being built as early as Basketmaker III times and there is good evidence to support the theory that these structures represent a borrowed trait, that the idea of their use did not originate with the Basketmakers. Construction of Great Kivas apparently was discontinued with the abandonment of the San Juan Pueblo area at the end of the Classic period, approximately A. D. 1300.

The small kiva, on the other hand, is considered indigenous to the region, making its appearance at the beginning of Pueblo II as a direct outgrowth of the shallow pithouse of Basketmaker III. Many small kivas continue in use today, still the centers of religious life of several groups of Pueblo Indians. The concern here is with the small kiva of the northern San Juan area, known as the Mesa Verde Kiva.

Three Pueblo II kivas at Site 16, one early Pueblo III kiva at the Sun Point site (Lancaster and Van Cleave, "Excavation of Sun Point Pueblo," in this volume) and two late Basketmaker III deep pithouses (Lancaster and Watson, "Excavation of Two Late Basketmaker Pithouses," in this volume) were excavated by the National Park Service in 1950. Each structure constitutes a different stage in the development of the Mesa Verde Kiva. Because of this excavation, which filled the existing gaps, structures are now available for interpretive use to illustrate the basic changes which took place in the gradual development of the stylized kiva of Classic Pueblo III.

In this interpretive sequence are the evolutionary stages preceding the construction of the first true kiva, structures dating approximately A. D. 600, 700, and 840, respectively. Then comes the first kiva, a Pueblo II structure, built around A. D. 900. Next are three developmental kivas of Pueblo II times which see the addition, one by one, of all the basic features found in ceremonial rooms of Pueblo III. The first of these kivas dates about A. D. 950, the second approximately A. D. 1000, and the third dates A. D. 1074 (tree-ring dates). These structures are followed by a well-developed mesa-top kiva of the early Classic period, dating late in the 1100's. At the end of the sequence are the beautifully constructed kivas of the late Pueblo III cliff dwellings, structures which resulted from the slow evolution throughout the centuries. These late Classic kivas date in the 13th century.

This is the first time such a series has been available for public presentation and while the sequence does not present all the infinite variations to be found in kivas, it does present the basic stages and fundamental changes. The dictates of medicine men, individual preference, radical differences or changes in the ritualistic patterns of various religious societies, could influence the final design of any particular kiva. However, kivas adhere generally to a pattern and that pattern evolved along certain lines. The series now on exhibit and used in the interpretive program constitutes a unique historical sequence.

It is not necessary to describe the stages preceding the construction of the first kiva. Most archeologists are convinced the kiva was the outgrowth of the Basketmaker pithouse and the evolution was: first, shallow pithouse with antechamber; second, deep pithouse with ventilator crawl tunnel; third, deep pitroom with ventilator, but a roof hatchway entrance; fourth, the four-post kiva. Excellent summaries of the development are found in Brew (Brew, 1946, pp. 203-214); Morris (Morris, 1939, pp. 36-38); and Roberts (Roberts, 1929, pp. 81-90). The most significant points are presented by Mr. Morris, who states: "Several diagnostic features of the kiva, such as fireplace, deflector, and ventilator are present in the very oldest dwellings thus far found" (Morris, 1939, p. 36). It might be pointed out that many pithouses are characterized also by a banquette and sipapu.

This discussion will deal with the development of the kiva from its four-post beginning to its final Classic stage, following the development as it is presented in the interpretive program of this park. Plate 34 shows the ground plans and profiles of six kivas now used in the program and these will be referred to in describing the steps or stages. Step 1 is the four-post kiva at Site 16; Step 2, the Twin Trees kiva; Step 3, Kiva 3 at Site 16; Step 4, Kiva 1 at Site 16; Step 5, the Sun Point kiva; Step 6 is Kiva C, Cliff Palace, a highly standardized Mesa Verde ceremonial chamber.

Step 1.—Four basic kiva features are present: banquette, firepit, sipapu and standard ventilating system, consisting of a vertical shaft joining a horizontal tunnel which enters the kiva from under the banquette on the south side.

The first kivas are simple structures, and though reminiscent of Pueblo I pitrooms, certain significant changes have taken place. The basic change is immediately apparent: the partitioning of the room has disappeared and the four roof support posts have been moved from the floor to the banquette. The floor space is cleared of major obstructions of a nonreligious nature, freeing the space for the performance of religious rites. Domestic arrangements have been dispensed with. Another change is seen in the banquette: this now circles the structure instead of being confined to the east, north, and west walls, as was customary in the earlier pit structures.

Kiva 2, Site 16, illustrating this first step, has no deflector. This feature may or may not be present in any given kiva, but it appears that few of the early kivas had a built-in deflector. Instead, a movable slab was used to close the opening of the ventilator tunnel when need arose to cut off the down draft of air. The firepit is a slab-lined box, somewhat south of the center of the kiva. The sipapu is located to the north of the firepit and on a line with it and the ventilator. Other floor features consist of an ashpit, often present in early kivas, and a small cist, or pot rest.

The kiva has the standard ventilating system mentioned above, and it may be noted that the ventilator shaft is located some distance back of the south wall. The only other feature that need be mentioned is the banquette cist, apparently a common feature of early kivas.

The roof probably was constructed by laying 4 horizontal beams, or stringers, across the forked tops of the 4 support posts. Smaller poles across the stringers formed the top of the roof, while other small poles spanned the space between the stringers and the top of the earthen kiva walls to form the sides of the roof. These were covered with reeds, bark, etc., and a thick layer of adobe. A hatchway in the roof served as smoke hole and entrance.

Step 2.—Five basic kiva features are present: banquette, pilasters, firepit, sipapu, and intersecting shaft-tunnel ventilator.

The second step is almost identical with the first except for one important substitution: stone pilasters replace posts as roof supports, thus paving the way for a major architectural change in Step 3. Dr. O'Bryan, in describing the Twin Trees kiva, which is used to illustrate Step 2, stated the roof probably was cribbed (O'Bryan, 1950, p. 34). This seems unlikely as there has been no basic change in the number of roof supports which would reflect a major change in construction and design. If the roof were cribbed, it would have been necessary to carry the cribbing straight up in order to allow any headroom. It seems more logical to presume this 4-pilastered kiva was roofed in the same manner as the 4-post kiva.

The same basic features, with the exception of the pilasters, are present in Step 2 as were found in Step 1. The firepit in the Twin Trees kiva is a shallow, circular basin instead of a slab-lined box. The sipapu is north of the firepit and in a line with it and the ventilator. Those familiar with Dr. O'Bryan's report on this structure will note that he does not show a sipapu in his ground plan of the kiva, and in commenting on this feature he states, ". . . a shallow depression at the southwest base of the circumference wall . . . may have served as the opening to the underworld (as no 'sipapu' was found in a more conventional place)" (O'Bryan, 1950, p. 34). When a ground plan of this kiva was being drawn for inclusion in this report, the question of the presence or absence of a sipapu arose. Feeling that the kiva had a sipapu, and it was not the cist against the west wall, the floor area north of the firepit was carefully swept with a stiff whisk broom. The circular outline of a plugged sipapu was located a little over 2-1/2 feet north of the firepit, so the feature was cleaned out. It is standard in all respects.

This kiva, like the first, does not have a deflector, and like the first, there is no indication on the hard-packed floor of there ever having been one. There is a well-made groove in the mouth of the ventilator tunnel, and a shaped slab found leaning against the wall by the tunnel opening fits the groove, showing the down draft was cut off by placing the slab in the tunnel opening.

Step 3.—Five basic kiva features are present: banquette, pilasters, firepit, sipapu, and ventilator. Undoubtedly, though it could not be demonstrated in this particular kiva (Kiva 3 at Site 16), a sixth basic feature, a built-in deflector, also is present as most kivas of this type have one.

Three new features make their first appearance: a cribbed roof, masonry lining of the face of the banquette, and a banquette niche in alinement with the sipapu, firepit and ventilator. These features are considered to be typical of later kivas, but not basic.

Step 3 is marked by the first really basic change in kiva design and construction. Six stone pilasters indicate the introduction of a cribbed roof. By using this number of pilasters, or more if the kiva was larger, the builders were able to hold the cribbing close to the outer circumference of the structure, thus not interfering to any degree with headroom or floor space.

The six pilasters are developmental in character, being like those employed in Step 2. Narrow slabs were laid one on top of another, forming small rectangular blocks. These are usually 6 in number in Mesa Verde kivas, but occasionally, as in the following step, 8 were needed to support a roof.

While it is not known if Kiva 3, illustrating this step, has a deflector, most 6-pilastered kivas do, and a built-in deflector is a basic characteristic from this stage on. The deflectors in these early kivas are, for the most part, a thin slab set on end in a shallow groove between the firepit and the ventilator tunnel, and cemented firmly in place with adobe.

The firepit is a slab-lined box, like that seen in the first kiva. The sipapu is in the usual position to the north of the firepit and aligned with it and the ventilator.

Step 3 is marked also by the introduction of coursed masonry. This is confined to the face of the banquette, and to judge by the early kivas with their earthen banquettes, there is in most cases no structural need for this lining. It is interesting to note the masonry lining makes its appearance coincident with the first major change in kiva design. It is felt this indicates the growing importance of these structures and reflects the attitude of the people toward their religious edifices.

Another new feature makes its first appearance—a niche in the north face of the banquette in a direct line with the sipapu, firepit, deflector (if present), and ventilator tunnel. This niche makes its appearance with the introduction of the masonry lining of the banquette and, while not always present, it becomes a common feature of later kivas. Beside the niche, a deep banquette cist also is present.

Step 4.— Seven basic kiva features are present: banquette, pilasters, southern recess, deflector, firepit, sipapu and ventilator.

The final basic change in kiva design takes place in Step 4: the southern recess is added, thus giving the kiva the pattern that becomes more or less standard in Pueblo III. Actually, any kiva with pilasters is characterized by the same number of recesses as it has pilasters, the recesses occurring between each set of pilasters. But in the latter part of Pueblo II, the recess occurring between the two southern pilasters was deepened, giving the kiva a "keyhole" outline. In most kivas the ventilator tunnel enters under this recess. The reason for deepening the southern recess has never been understood.

Step 4 is characterized by another change, a constructional improvement that also involves design. As a glance at plate 34 will show, the pilasters in Kiva 1, Site 16, Step 4, are very different from those used in the Twin Trees kiva and Kiva 3, which illustrate Steps 2 and 3. The heavy masonry pilasters of Kiva 1 are no longer rectangular but flare from front to back. This change, as pointed out by Dr. Brew, ". . . brings the previously rectangular pilasters into the circular plan of the kiva. The side walls of the pilasters become as the outer sections of radii of the circle" (Brew, 1946, p. 210). The pilasters are much more substantial and capable of bearing a heavy load. It is interesting also to note that these pilasters are set slightly back from the edge of the banquette. Expanding masonry pilasters, set back from the banquette edge, are characteristic of Mesa Verde kivas from. this time on.

The deflector is still a sandstone slab set on edge in a narrow groove between the firepit and the ventilator tunnel. The firepit has not changed, being a slab-lined box as in earlier kivas. The sipapu and north banquette niche are present, and are alined with the firepit, deflector, and ventilator tunnel. Other banquette niches make their appearance and become a common feature. The variety of floor features of Kiva 1, while not too common, are in no way unusual for a developmental kiva.

The masonry lining has been extended and now appears above the banquette to the height of the pilasters, with the exception of the southern recess. This is a feature which varies so much from kiva to kiva, in mesa-top ruins, that it cannot be said to be either typical or atypical. The masonry is not essential to construction, as evidenced by the southern recess which is unlined. Occasionally it was necessary to use masonry in mesa-top kivas to stabilize the earthen walls, but on the whole the use of a lining seems to indicate a desire on the part of the builders to improve the appearance of their ceremonial rooms. Certainly from now on the finest craftsmanship was often expended on the kivas. Kiva 1 is an excellent example of this trend: the kiva is well built, but the village with which it is associated is poorly constructed.

Step 5.—The seven basic kiva features are present: banquette, pilasters, southern recess, deflector, firepit, sipapu, and ventilator.

Except for one slight change, Step 5 is much like Step 4, as far as basic pattern is concerned. In Step 5 the ventilator shaft becomes an integral part of the kiva, being constructed as a part of the chamber, so that the back side of the south wall of the southern recess becomes the inner side of the north wall of the shaft. In Step 4 the ventilator shaft was located 3 feet back of the southern recess. This change in the position of the ventilator shaft does not show too well on the ground plan of the Sun Point Pueblo kiva, used to illustrate Step 5, as in excavating it was not possible to completely open the top or full length of the ventilator shaft. Constructed as it is in native earth, the wall between the shaft and the southern recess would have collapsed if cut too thin. However, it was possible to determine that the back of stones used in the southern recess of the kiva formed the north side of the ventilator shaft.

The pilasters of this kiva are like those used in Step 4, and since the kiva is considerably smaller than Kiva 1 at Site 16, only six pilasters, the usual number, are present.

Step 5 sees the introduction of a masonry block deflector, replacing the thin slab deflector of earlier kivas. The firepit is no longer a slab-lined box. A circular basin is used instead, as is the case in most classic kivas, and the slab-lining has been dropped. The sipapu and north banquette niche are present and in alinement with the firepit, deflector and ventilator tunnel.

While the Sun Point Pueblo kiva has little masonry lining, that employed in finishing the upper one to two-thirds of the face of the banquette is exceptionally fine. The lower part of the banquette of this kiva cut the caliche layer underlying the red top soil of the mesa. This caliche is sometimes quite firm, as is the case here, so the builders did not bother to extend the masonry below the top of the layer. The upper part of this banquette may have been unstable, demanding the use of masonry to hold the edge in place. As pointed out before, the use of masonry varies from kiva to kiva, and as it is known the Sun Point kiva was excellently plastered, perhaps the builders did not see fit to expend much energy on giving their ceremonial room a complete lining. A block of crude masonry between the two north pilasters of this kiva is undoubtedly a stabilization measure. The masonry bulges, indicating pressure from behind. The southern recess of the kiva is lined with slabs and some masonry.

This kiva has three features which are not shown on the ground plan (pl. 34, Step 5). These features are unusual but not unheard of in a Classic kiva, though similar features are more often encountered in Pueblo II kivas. There are three dome-shaped cists cut into the face of, and built under, the banquette. Two of the cists are quite large and are located to either side, 1 to the east and 1 to the west, of the ventilator tunnel and southern recess. The smallest of the three cists is located toward the back of the kiva on the east side. The openings to the cists extend from floor level up into the face of the banquette in the caliche layer, but the bottoms of the two larger cists are below floor level. For a complete description of the cists, see the report on Sun Point Pueblo (Lancaster and Van Cleave, "Excavation of Sun Point Pueblo," in this volume).

Step 6.—The seven basic features are present: banquette, pilasters, southern recess, deflector, firepit, sipapu, and ventilator.

There is only one basic difference between Classic kivas built early in the period, as in Step 5, and those built later in the period, as in Step 6. Because of the location of the late structures in the sandy fill or loose trash of caves, or at the front edges of sloping cave floors or on rocky ledges, they necessarily had to be constructed of masonry, so one is immediately impressed with the seeming excellence of the cliff dwelling kivas. It should be mentioned in this connection that early Pueblo III kivas are built of masonry when they are incorporated within the floor levels of a house block, as at Far View House, an early Classic compound pueblo located 4 miles north of headquarters, Mesa Verde National Park. These kivas are referred to as intramural, and there is seldom anything to distinguish an intramural kiva of the A. D. 1100's from a cliff dwelling kiva of the A. D. 1200's. Since no intramural mesa-top kivas are now used in the interpretive program, they are not considered here. This discussion follows kiva development as it is now illustrated in the present interpretive program.

A question often arises concerning the extent of masonry in the outer walls of the kivas built in caves. Did it extend downward back of the banquette to what would be the floor level of the kiva? Did it extend above the tops of the pilasters to the level of the surrounding court, or was it merely confined to the space between the top of the banquette and the tops of the pilasters? The appearance of the stabilized kivas in Spruce Tree House, Cliff Palace, and other excavated cliff dwellings in which the outer kiva wall extends to the level of the surrounding court, has sometimes been questioned, the feeling being that the walls never extended above the tops of the pilasters.

When Kiva C in Cliff Palace was chosen as a typical example of a Classic kiva in a cliff dwelling and surveyed for inclusion in this report to illustrate Step 6, the question of the actual extent of masonry used in the outer wall immediately arose. Did it extend downward back of the banquette? Was there any justification for showing it extending above the tops of the pilasters to the surrounding court level? Old pictures of Cliff Palace and other cliff dwellings were studied. These pictures, taken in 1891 and 1893, many years before the excavation and stabilization of a Mesa Verde cliff dwelling, show several kivas in which the masonry wall extends above the tops of the pilasters, and in two kivas where the masonry is intact the photographs show the walls continuing to the level of the surrounding court. These pictures were taken to Cliff Palace where the two kivas are located and compared with the walls as they appear today. In several instances individual stones could be spotted in exactly the same location as shown by the photographs. Further study of other kiva walls above the pilasters proved that parts of them are definitely original and not reconstructed. Not all kivas had the upper kiva wall, it is true, but it is now demonstrated that some of them did and still do. It is not known, however, if the masonry of the outer wall ever extends down behind the banquette to floor level. Investigation of Kiva C showed that it, at least, does not have such an extension of the outer wall.

In all respects Step 6 is the ultimate refinement of Step 5. The basic as well as the typical kiva features are all present and in the usual standard position. A glance at plate 34 will show that Step 6 incorporates all features and is the result of the steady, gradual trend to standardization.

A description such as given above is of necessity more or less limited and generalized, since it deals with only six structures. A given kiva may contain any or all of the basic features and the arrangement of the features may differ a great deal from one kiva to the next, leading to infinite variations. The complexity of the situation, as regards not only kivas but Pueblo architecture in general, is recognized by all, but best described by Dr. Brew, who states, 'In my own observations of both ancient and modern examples of Pueblo architecture, if I have found any rule at all it has been one of uncompromising irregularity" (Brew, 1946, p. 215). For this reason, and in order to summarize the discussion of the various stages, it may be well to consider the time of appearance and variations of the several kiva features referred to above.

Roof support posts.—Characteristic only of Step 1; use in later kivas is uncommon but not unknown (Kiva F, Lowry Pueblo, is a case in point. Martin, 1936, pp. 40-42, and pl. LVI).

Pilasters.—Replace posts for roof support and are characteristic of Steps 2, 3, 4, 5, and 6. Usually lacking in kivas which do not have a banquette, though this is not invariable, as seen in Kiva H, Lowry Pueblo, which has no banquette but does have pilasters rising from the floor (Martin, 1936, pl. XLIX. This kiva is pictured, not described). Pilasters are narrow, rectangular blocks in Steps 2 and 3. Flaring masonry pilasters are introduced in Step 4, and characterize Steps 4, 5, and 6. Coincident with the introduction of improved pilasters is the practice of setting these members slightly back from the edge of the banquette. Dr. Brew mentions that narrow block pilasters are sometimes set back from the banquette edge (Brew, 1946, p. 210), but by Step 4 the practice becomes common, though not universal.

Step 2 is characterized by 4 pilasters; Steps 3, 4, 5, and 6, usually by 6 pilasters. However, the number may vary and Classic kivas sometimes have only 4 pilasters, and large Pueblo II and III kivas have 8 pilasters, like Kiva 1 at Site 16.

Pilasters are not always functional, as kivas are known which have pilasters that never served to support a roof. Mr. Morris excavated 2 kivas at site 41, La Plata Valley (southeast of the Mesa Verde), each of which has 4 pilasters, the tops of which are finished with flat slabs. The walls of these structures are smoked black, and the smoking extends in an even coat across the tops of the pilasters, indicating nothing ever rested on them (Morris, 1939, pp. 89, 94).

Southern recess.—Present in Steps 4, 5, and 6, being introduced late in Pueblo II and becoming more or less standard in Pueblo III. There is great variation in the depth of the southern recess, and like any other feature, it may be lacking in individual kivas. In some kivas where it is lacking, the standard recess between the two southern pilasters is obliterated by filling the space with masonry. An odd variation of the southern recess is seen in Kivas A and B, at Lowry Pueblo. The banquette is missing between the two southern pilasters, but there is a deep southern recess which starts at floor level (Martin, 1936, pls. LI, LVIII, pictured but not described). A similar floor level recess was encountered in Kiva 9, Site 41, La Plata Valley. However, in Kiva 9, short shelves extend into the recess from in back of each flanking pilaster (Morris, 1939, p. 90, and pl. 62).

No satisfactory explanation of this feature has ever been advanced. It has been suggested they were used as storage compartments, spectators benches, altars, etc. If the recess was used for storage, why is it invariably in the same position? Even when it was necessary to bring the ventilator in from another direction in order to promote a proper draft, the southern recess is often present in the customary place. This factor, and the usual alinement of the recess with the deflector, firepit, sipapu and north banquette niche, indicate more significance was attached to this feature than would be accorded a storage compartment. The size of most southern recesses in the Mesa Verde discounts their use as a spectators' bench. It is impossible to prove or disprove their use as "altars," but there is at least one kiva in the Mesa Verde, described below, in which the southern recess never served such a purpose.

There are southern recesses in two different kivas in cliff dwellings in the Mesa Verde which, when considered with Kivas A and B at Lowry Pueblo, and Kiva 9 at Site 41 (mentioned above), indicate the difficulty of trying to advance any logical theory as to the purpose of these structural features. The kiva beside the so-called Square Tower, in Square Tower House, has an original cribbed roof in position. In this structure the southern recess is roofed along with the kiva proper, which makes it a dark alcove. Another kiva, in Long House, has half of the original roof in position. This roof is not cribbed; instead, four long poles are laid across the north-south axis of the kiva and crossed with smaller poles. The half of the roof still in position shows that the hatchway entrance to the kiva is located directly above the southern recess. So many variations of this feature probably indicate a symbolic and not a functional use.

The explanation is sometimes offered that the southern recess represents a survival of the southern antechamber of pithouses, or the south end of partitioned pithouses and pit rooms. In view of the kiva development, such an explanation is farfetched. Any suggestion of a southern antechamber disappeared in Pueblo I, and any partitioning of pithouses or pitrooms by the end of the period, prior to the beginning of Pueblo II. The southern recess was not introduced until 150 or more years later, so it obviously could not be a survival of anything connected with pithouses or pitrooms.

Deflector.—Present in all steps, but it may be lacking in individual kivas. It is usually lacking in Steps 1 and 2, and the evidence shows that, in this case, a movable slab was used to close the opening to the ventilator tunnel to cut off the draft. A built-in deflector characterizes Steps 3, 4, 5, and 6. Deflectors in Pueblo II kivas, Steps 3 and 4, are usually sandstone slabs set in a groove between the firepit and the ventilator tunnel, and cemented firmly in place with adobe. Deflectors in Steps 5 and 6 are often constructed of masonry. However, many types of deflectors are found in Classic kivas: masonry blocks, sandstone slabs, upright posts set in adobe, and low, enclosing masonry walls curved to meet the banquette on either side of the ventilator tunnel. Evidence shows that a kiva with a built-in deflector may also have a movable slab to close the tunnel opening, as seen in the outer court kiva at Far View House.

Firepit.—Present in all steps and in all kivas. Those in Pueblo II kivas are often slab-lined, rectangular boxes. Beginning with Step 5, the circular firepit replaces the slab-lined box to a great extent. These circular firepits may be shallow basins with no distinguishing features, they may be clay-lined, with the lining brought up to form a raised lip around the edge of the basin, and occasional firepits in Classic kivas are masonry lined.

Sipapu.—Present in all steps though it may be lacking in an individual kiva. Usually, though not invariably, located in the floor north of the firepit and alined with it, the deflector, ventilator tunnel, and north banquette niche, if the latter is present.

North banquette niche.—Present in Steps 3, 4, 5, and 6. While not an invariable feature, this niche, alined with the sipapu, firepit, deflector, and ventilator, is common enough to constitute a characteristic of later kivas.

Other banquette niches.—Present in Steps 4, 5, and 6, these are considered to be characteristic of Classic kivas.

Banquette cists.—May be present in any step, either as pockets built into the banquette or, especially in later kivas, as pottery jars set in the bench.

Ventilating system.—Present in all steps. Composed of a vertical shaft joining a horizontal tunnel which passes under the banquette, usually on the south side. Shaft. located one to several feet back of the south wail of the kiva in Steps 1, 2, 3, and 4. In the last two steps, Nos. 5 and 6, incorporated within the actual plan of the kiva, the center section of the south wall of the southern recess, or kiva, being the north wall of the shaft. Tunnel: dug into native earth in mesa-top kivas and seldom lined. Usually roofed over with poles or wooden planks where it passes under the banquette or southern recess in mesa-top kivas, though sometimes roofed with large slabs, as in the Sun Point kiva. Constructed of masonry in intramural mesa-top kivas and cliff dwelling kivas, and roofed over with large slabs or timbers.

Masonry lining.—Seldom encountered in Steps 1 and 2 (any masonry found in these earliest kivas is crude and usually incorporated as a stabilization measure). Lining of the face of the banquette introduced in Step 3, of the entire kiva in Step 4. While a variable feature in any step except Step 6, when the location of the kiva necessitated extensive masonry construction, it would seem that the growing tendency to use masonry in kivas coincides with the improvement of masonry and its standardized use in house architecture. It would appear also that kivas became more and more important with the passing of the years, and that every effort was made to make most of these ceremonial structures as fine as possible. Indicative of the growing importance of kivas is the structural excellence of so many of these chambers compared with the crudity of some of the villages with which they are associated.

Roofing.—Most 4-post and 4-pilastered kivas are presumed to have been roofed like pitrooms of Pueblo I, with a rectangular framework resting on the posts or pilasters. The framework supported smaller poles covered with reeds, bark, etc., which were covered in turn with a thick layer of adobe. Evidence shows that most 6 and 8 pilastered kivas were roofed by cribbing, the bottom logs of the dome-shaped crib resting on the pilasters and spanning the recesses between them. Shakes or small poles embedded in the outer kiva wall above the height of the pilasters served to fill the space between the straight logs and the circular kiva wall. Poles, twigs, and bark were used to fill spaces between logs, then the roof was topped with adobe. Kivas in which the pilasters were not functional, and kivas lacking pilasters, posts, and/or banquettes were roofed, in all probability, like the kiva in Long House described above.

Entrances.—The standard entrance was through the hatchway left in the roof, this hatchway also serving as a smoke hole. A ladder through this hatchway gave access to the room beneath. A number of kivas have been excavated which have holes, boxes or slots in the floor to hold the bases of ladder poles, such as seen in the Twin Trees kiva (pl. 34, Step 2). Secondary entrances are present in some kivas. The kiva at the Sun Point Pueblo (pl. 34, Step 5) is connected by a crawl tunnel to the circular tower located beside it, a situation often encountered in kiva-tower units in the Mesa Verde region. Occasional kivas, both on the mesa tops and in the caves, have a tunnel entrance to a nearby room in the pueblo.

Outline.—Nothing has been said so far about kiva outline, as a typical Mesa Verde kiva is regarded as circular. While this is true, on the whole, there are modifications in shape, just as there are variations of all other features. A standard Classic kiva with southern recess is actually "keyhole" shaped, rather than circular, even though the basic outline is round. A glance at the ground plan of any excavated ruin of size will convince one that the shape of the kivas, while tending to be circular, depended to considerable extent on the location of each individual structure and the space available for its construction. Occasionally a kiva is somewhat flattened on the southern side, giving it the appearance of a modified "D." Other kivas are more oblong than round, while still others are squarish with rounded corners. And there are square kivas in the Mesa Verde. These are not common, but enough examples are known to indicate such structures are not too unusual. One cliff dwelling, Bone Awl House, located on the east wall of lower Soda Canyon, has four square kivas, and an occasional square kiva is encountered in a cliff dwelling in which the rest of the ceremonial rooms are circular.

Orientation.—It is interesting to note, though it may or may not be significant, that the mesa-top kivas discussed here are oriented east of south. Kiva 2, Site 16, is oriented 28-1/2° east of south; the Twin Trees kiva, 35°; Kiva 3, Site 16, 34° Kiva 1, Site 16, 32°; while the Sun Point Pueblo kiva is oriented 31-1/2° east of south. These are all magnetic readings. This orientation is exactly like that observed in the pithouses and pitrooms of the Basketmaker III and Pueblo I periods, and is typical of almost all mesa-top and valley kivas.

Kivas in cliff dwellings vary greatly in their orientation, depending upon the direction in which the cave faces. There seems to have been an effort whenever possible, however, to orient these structures to the south. Kiva C in Cliff Palace (pl. 34, step 6), for example, does not face the front of the cave, which would orient it to the west, but faces the south end of the cave, which gives it an orientation of 6-1/2° west of south.

Few tools, implements or utensils were found at Site 16, but this is not surprising in view of the story revealed by excavation. No one occupation of the site was of prolonged duration and each village excavated was the home of, at the most, only a few families. The post and adobe village burned, undoubtedly destroying most household furnishings and personal possessions. Remains of the burned village were cleared away when the first unit pueblo (Unit Pueblo No. I) was constructed. This pueblo was afterwards abandoned and undoubtedly the people took their possessions with them. Later it too was leveled, this time by the last occupants of the site to provide building space for Towers B and C. Finally this last pueblo (Unit Pueblo No. II) also was abandoned. Either on abandonment or some time later it was razed and the salvaged building materials carried away. The small number of artifacts recovered represent what was normal loss over a period of years and objects discarded as having no further usefulness.

No discarded artifacts or burial offerings were taken from the trash mound, usually a rich source of material culture. Excavation of the thoroughly pot-hunted debris would have been a waste of time and effort since the indiscriminate digging of the past completely destroyed any story the mound might have contained.

Some stone and a few shell and bone artifacts were recovered. The bulk of the excavated material consists of broken pottery and, considering how an earthen vessel may break into several hundred pieces, even this material is scarce. Specimens from Site 16 are considered, with the exception of certain pottery types mentioned below, to be typical of the Pueblo II period.

While occupation in Basketmaker III and Pueblo I times is indicated, no Basketmaker III pithouses or Pueblo I slabhouses or pitrooms were excavated. Material representative of these earlier periods consists only of potsherds of Lino-like plain wares, La Plata Black-on-white and La Plata and/or Bluff Black-on-red.

The stone tools found at Site 16 are in no way exceptional as a glance at the illustrations will show. Plates 35 through 40, immediately following this section, picture most of the stone artifacts recovered from the site.

Manos—Thirteen manos, handstones used in grinding corn, etc., on a metate, were found and four are pictured and described on plate 35. Eleven of the specimens have a single grinding surface; 2 have been used on both sides. Six are convex end-to-end, 2 are convex side-to-side, and 2 are convex both end-to-end and side-to-side; the other 3 are flat. Four are wedge-shaped in cross section. No metates were found, but the use of both convex and flat manos indicates metates were both trough-shaped and flat-slabbed.

Rubbing stones or small manos—These are small stones of various shapes which conveniently fit the hand and they could have been used for grinding many different materials in a basin or on a slab. The ends of several specimens show they were used also for pecking and hammering. All the small manos from the site are pictured on plate 36.

Pecking and hammer stones.—For the most part these implements are reused fragments of axes, mauls and small manos. From the chipped and battered ends of these specimens it is obvious that they were used in shaping other stone artifacts such as metates, manos, axes, door slabs, building blocks, etc. Specimens found are shown on plate 37.

Grooved and side-notched axes and hammers, and grooved mauls.— Of the 14 implements grooved or notched for hafting, only 3 axes and 2 hammers are complete. Specimens show that side-notching and full-grooving were both used in hafting axes and hammers; full-grooving only was used in hafting mauls. Several implements are illustrated on plate 38.

Miscellaneous stone artifacts.—Plate 39 pictures miscellaneous items of stone from the ruin. Only one projectile point, large and crude, was found. Five polishing pebbles were recovered (pl. 39, middle row). These small, hard, smooth pebbles were used, as they are today, to polish pottery. The miniature side-notched axe, which looks like a modern curio store item, is an oddity (pl. 39, upper row, middle). Made of the material used in the manufacture of tchamabias (fleshers), it is no doubt a child's toy as its battered edges would not be in keeping with a ceremonial object. The sandstone niche cover shown (the cover for the north banquette niche of Kiva 3) is the only complete specimen of its type from the site (pl. 39, bottom). Numerous flakes, spalls and chips were found but, as none of these showed any evidence of use, they are not considered as artifacts. Stone beads and pendants are discussed under ornaments.

Concretions.—Concretions were picked up in the past, as today, because of their strange shapes. Children no doubt treasured them, medicine men may have valued them, and an occasional one served a utilitarian purpose. Those illustrated in the top row of plate 40 are amusing because of their shapes and interesting in that they were found together as a cache.

Bone artifacts are few in number and plate 41 pictures all but two of the specimens. A bone bead, considered under ornaments, is illustrated on plate 42, and a bone die (or gaming piece) is pictured on plate 44.

Awls—The eight awls from Site 16 are made of mammal bone, mostly deer, and though few in number represent a wide range of types.

Flakers—Two bone implements can be classified as flakers. One of these (pl. 41, No. 1, lower row) is unusual in that it is made of a piece of massive, dense and very heavy bone. It was sent to the Chicago Natural History Museum in hopes the bone could be identified. Identification proved impossible, but Dwight Davis, curator of vertebrate anatomy of that institution, suggested bison or elk (personal letter from Dr. Paul S. Martin, Chicago Natural History Museum, to Park Archeologist Don Watson, January 17, 1951).

Fleshers or scrapers—Two implements found are classified as fleshers or scrapers, though similar specimens are sometimes considered to be weaving tools. The two specimens from Site 16 do not exhibit any polishing such as would be expected in a weaving tool.

Objects of unknown use—The last three objects pictured on the right in the bottom row of plate 41 are of unknown use. These are sometimes encountered in excavation of ruins in the Mesa Verde region and have been the subject of considerable speculation. Dr. Martin (Martin, 1939, pp. 424-425, fig. 122, upper left corner) and Dr. Roberts (Roberts, 1930, p. 147, and pl. 43a) classify these objects as whistles. Dr. O'Bryan suggests they were used as tinklers (O'Bryan, 1950, p. 87, and pl. XXXVIIIb). Dr. Reed illustrates what appears to be an incomplete specimen and suggests it possibly was used as a flute or whistle (Reed, 1943, p. 176, and pl. XXIX, specimen 1/11).

These artifacts were manufactured only of the tibia bones of small mammals: jack rabbits, cottontails, foxes, and a fairly consistent pattern was followed. The proximal end of the bone was ground down to expose the marrow cavity. Just below this ground surface a shallow notch was often cut across the bony ridge extending down the front of the shank. A hole was drilled through the shank to the marrow cavity somewhere between the proximal and distal ends. The distal end of the bone was never worked.

A specimen in the Mesa Verde Museum collections, which came from a cliff dwelling, has a fragment of cord attached in the following manner: the cord is inserted through the hole in the shank, knotted to keep it from slipping through, and run up the marrow cavity to emerge at the proximal end. The cord is broken just beyond the point where it emerges from the cavity so the original length is not known. It is this specimen which gave Dr. O'Bryan the idea the objects were used as tinklers. It is difficult to visualize the use of these long, slender and fragile bones for such a purpose. There are numerous other bones better adapted to such use, and it seems odd that such a distinctive bone, and always the same bone, would be chosen for a tinkler.

The specimen with the cord through the hole in the shank definitely proves the objects were not used as whistles, an untenable idea in the first place since no one has ever been able to produce the slightest noise on one. The placement of the hole on most specimens, in the trough between the ridges for muscle attachment, makes it impossible to use these artifacts as whistles.

It is quite possible that these bone artifacts which, so far as can be determined, are found only in the Mesa Verde and closely allied Piedra areas, have some ceremonial significance. Five of the 9 specimens in the Mesa Verde Museum collections are known definitely to have come from kivas: the 3 discussed here, from kiva 3, Site 16; 1 from an abandoned and backfilled kiva at Far View House; the fifth from Kiva V, Site 34, Soda Canyon Pueblo excavated by Gila Pueblo in 1948 (O'Bryan, 1950, p. 73, pl. XXXVIIIb; also M. V. Museum catalog). There is little doubt but that two others in the collection, from Bone Awl House, were found also in kivas, as only the kivas in this cliff dwelling were excavated and they yielded a number of bone implements (excavated by Superintendent J. L. Nusbaum, 1927; no report). It is one of the two specimens from Bone Awl House that still has the cord attached. The original cataloging of the 2 specimens indicates that both were found with yucca cord strung through the perforations, though only the 1 still retains the cord.

It is interesting to note that most of the specimens found by Dr. Martin at Sites 1 and 2, Ackmen-Lowry area, were in deep pitrooms of Pueblo I origin (Martin, 1939, p. 430). Most archeologists feel these pitrooms were the forerunners of the Mesa Verde kiva and were partially ceremonial in function. The only complete specimen obtained by Dr. Roberts came from a trash mound (Roberts, 1930,. p. 175). Baron Nordenskiöld pictures a specimen found in Spring House, but the exact provenience is not given (Nordenskiöld, 1893; sketched on the caption page for pl. XLI).

Manufacture of these objects apparently began in Pueblo I, since to date none has been reported from Basketmaker III sites. They continued to be made through Pueblo III, but so far as is known were not manufactured after the abandonment of the Mesa Verde area. Dr. O'Bryan lists these objects as a new trait for the McElmo Phase of Pueblo III (O'Bryan, 1950, p. 109). This is difficult to understand since he states they were found at both Site 102 ruin (Pueblo II) and Site 34 (Pueblo III), and pictures specimens from both sites (O'Bryan, 1950, p. 87, and pl. XXXVIIIb). Their presence in Pueblo I pitrooms is proof of much earlier origin.

Bone die or gaming piece.—One elliptical-shaped, flat piece of polished bone, seven-eighths of an inch long, came from the floor of Kiva 3 (pl. 44). Objects such as this are often referred to as dice.

Archeologists working in the Mesa Verde area are often impressed by the fact that few articles classifiable as ornaments are encountered in excavation. To all intents and purposes the former inhabitants of this region were "poverty struck" as regards possession of items for personal adornment. It is not often that ornaments are found interred with the dead and, considering Pueblo Indian burial customs in general, this factor alone is indicative of their actual scarcity. A review of items reportedly taken from the cliff dwellings in the early days shows that jewelry rarely was found.

Considering the usual dearth of such articles, and how few artifacts were recovered from Site 16, it is surprising any ornaments were found. Yet one of the finest necklaces taken from a Mesa Verde ruin was discovered here (pl. 43).

Pendants.—Four pendants were taken from Site 16: 1 of hematite, 1 of lignite, 1 of shell and 1 made from a reworked potsherd (see pl. 42). The bird pendant (pl. 42, middle row, left), cut from the shell of Cardium elatum (identification by Dr. E. W. Haury, director, Arizona State Museum), is undoubtedly a trade piece, but the other three pendants could have been made locally.

Beads.—The collection from the ruin contains three types of beads: tubular bone; figure-8-shaped white shell; and disk-shaped red shell, white shell and gray shale. The shell beads are considered imports, but the bone and shale beads are probably of local manufacture. The types of beads are shown on plates 42, 43, and 44.

Shell necklaces.—The graduated and tapered figure-8-shaped white shell and disk-shaped red shell beads (pl. 43) were found together 6 inches below the surface in the fill of Room 4, Unit Pueblo No. I. There was little to suggest the manner in which they were strung.

Figure-8-shaped beads are not too common but examples have been found in ruins throughout the Southwest and down into Mexico. However, this is the first known instance of their having been found on the Mesa Verde.

Four fine necklaces of figure-8 beads have come from sites in the Pueblo area. The one from Site 16 is very similar to the necklace from a Pueblo III ruin at Kiatuthlanna, Eastern Arizona, pictured and discussed by Dr. Frank Roberts (Roberts, 1931, pp. 161-162, and pl. 45). Another, found by Dr. Roberts in a ruin near Allantown, Ariz., and now at the Laboratory of Anthropology, Santa Fe, N. Mex., is composed of hundreds of tiny beads, so small that one has to look closely to see that they are actually figure-8-shaped (Roberts, 1939, p. 202; 1940, pp. 131-132). The fourth necklace is in the possession of Dr. Jesse L. Nusbaum, senior archeologist, National Park Service, Santa Fe. This necklace, a magnificent specimen in perfect condition, reportedly was found in a cliff dwelling in southeastern Utah. It is exceptionally long and the beads are quite large; in fact, the smallest bead in Dr. Nusbaum's necklace is larger than the largest bead in the necklace from Site 16.

Dr. Frank Roberts was at first inclined to consider the presence of figure-8-shaped beads in the prehistoric Pueblo area a Classic Pueblo III trait (Roberts, 1931, pp. 161-162). However, on later finding the same type beads in a Pueblo II site he revised his opinion, stating "Their presence in the second unit (a Developmental Pueblo in the Whitewater District, 3-1/2 miles south of Allantown, Ariz.) would tend to indicate a somewhat earlier horizon and to date them in the latter part of the Developmental period" (Roberts, 1940, p. 132). The necklace from Site 16 is further evidence that figure-8-shaped beads were known to the Pueblo Indians prior to Classic times.

Beads from the necklace were submitted to Dr. E. W. Haury, director of the Arizona State Museum, for identification. Dr. Haury was not able to identify the species of shells but suggested that since the common thick-walled shell used in the Southwest was Glycimeris, the figure-8-shaped beads might have been cut from that species. He also reported shells recently collected by divers in the Gulf of California have characteristics in common with the red shell beads. These shells are very thick walled and, when the outer crust is removed, a purplish to orange-colored shell, usually heavily worm eaten, is exposed. This exactly describes the appearance of the red shell beads in the necklace. Dr. Haury has not yet received a specimen identification of the Gulf of Lower California shells (information received in a personal communication from Dr. Haury, Feb. 2, 1952).

Shale and shell necklaces.—Two hundred and ten small, disk-shaped, graduated gray shale beads and six disk-shaped white shell beads were found in fill just south of Kiva 1. Possibly this fill is trash, or, on the other hand, it may be the fill of a burned room. The beads were screened from mixed ash and dirt. The shale beads vary in color from pale gray to dull, blackish gray and are well-made. When strung, the beads make a necklace 17-1/2 inches long (pl. 44).

Miscellaneous items.—Three objects pictured on plate 45, Miscellaneous Clay Objects, might be classified as ornaments. The end of an effigy handle of a Mancos Black-on-white ladle, bored for stringing, may have been worn as a pendant (pl. 45, lower row, right). The worked black-on-red sherd is obviously a pendant blank. The fired clay pellet with a broken loop, resembling a button, possibly was used for that purpose, or it may be a clay bead from a necklace (pl. 45, lower row, left).

Results of the analysis of the pottery from Site 16 parallel in detail the architectural evidence of the occupational span. Briefly summarized, the pottery evidence is as follows: (1.) Lino Gray-like ware and La Plata Black-on-white indicate use of the site during the Basketmaker III period. (2.) These 2 wares, together with La Plata and/or Bluff Black-on-red point to continued occupation in Pueblo I. (3.) Mancos Black-on-white and Variable or Exuberant Corrugated, the 2 wares comprising 90 percent of the pottery, signify prolonged use of the site in Pueblo II. (4.) The 2 diagnostic wares not present at Site 16, Mesa Verde Black-on-white and Pueblo III corrugated, offer conclusive evidence the site was abandoned prior to the beginning of Pueblo III and never reoccupied.

The pottery types and wares listed below are not described as to manufacturing, firing or decorating techniques, paint, paste, surface treatment, decoration or shape, as all are well known wares. Excellent technical discussions are to be found in Brew (Brew, 1946); Colton (Colton and Hargrave, 1937); Martin (Martin; 1936, 1938, 1939); Morris (Morris, 1939; and Appendix A, by Shepard); Reed (Reed, 1943); and others, and description here would be repetitious.

Basketmaker III—Pueblo I—Ten percent of the sherds from Site 16 are assignable to this category, which includes the Gila Pueblo "Four Corners" and "Chapin Mesa" phases (O'Bryan, 1950, pp. 104-107) There may be some question as to the advisability of grouping the wares of these two periods and considering them as one lot. However, such a classification is the only feasible one for the purposes of this report. No excavations were conducted at this time in Basketmaker III or Pueblo I horizons, though it is felt certain of the basis of findings in Tests 1 and 2 (pl. 19), that structures assignable to both periods are present at Site 16.

In the Mesa Verde area it is almost impossible to distinguish a Basketmaker III from a Pueblo I occupation on the basis of pottery analysis alone, for there is an almost unbroken sequence of what are considered to be typical Basketmaker III wares through the Pueblo I period. This continued use of early wares has been established and demonstrated through excavations conducted by Dr. J. O. Brew in southeastern Utah (Brew, 1946); Dr. Paul S. Martin in the Ackmen-Lowry area of southwestern Colorado (Martin, 1936, 1938, 1939); Earl H. Morris in the La Plata, Johnson, and Red Mesa areas adjacent to the Mesa Verde (Morris. 1939); Dr. Erik K. Reed in the Mancos Canyon (Reed, 1943); and Dr. Deric O'Bryan in Mesa Verde National Park (O'Bryan, 1950).

Lino-like plain wares—Eight and one-half percent of the pottery from Site 16 is a Lino-like plain ware. The ware is not broken down into the usual categories of Lino Gray and Kana-a Gray for the following reasons: Lino Gray, the standard utility ware of Basketmaker III times continued to be manufactured until sometime in Pueblo II when allover corrugated finally superseded other unpainted pottery to become the standard Pueblo utility ware. At Site 102 (Twin Trees Pueblo) for example, Dr. O'Bryan found that, during A. D 950, Lino Gray totaled half of all pottery found; Corrugated but 10 percent. By A. D. 1025, at the Site 1 pueblo (400 yards south of Twin Trees), the situation was exactly reversed (O'Bryan, 1950, p. 93).

The supposedly diagnostic unpainted ware of Pueblo I times, Kana-a Gray, is indistinguishable from Lino Gray except for the characteristic banded-necks of Kana-a jars. This diagnostic trait of Kana-a Gray appears to be of little use in the Mesa Verde region, however, for as Dr. Brew points out in his report on the Archeology of Alkali Ridge, ". . . local potters seldom left the bands showing on the necks of Pueblo I jars" (Brew, 1946, p. 291). Dr. O'Bryan concluded from his excavation in Mesa Verde that while "Kana-a Gray and Lino Gray were locally concurrent, the former was never as popular as the latter and older ware" (O'Bryan, 1950, p. 93). In analyzing the pottery from his sites, Dr. O'Bryan made an arbitrary breakdown of plain wares into Lino and Kana-a on the basis of the percentage of rim sherds of each ware (O'Bryan, 1950, p. 92). This procedure was impossible for Site 16, as not a single rim sherd was found which could be classified as Kana-a Gray.

La Plata Black-on-white.—A few sherds of La Plata Black-on-white, about 1 percent of the total, were recovered from the ruin. (See plate 46.)

This ware is the local, characteristic Basketmaker III decorated pottery. It continued to be popular in Pueblo I times and so far as is known at present, seems to be the typical black-on-white ware of both periods in the Mesa Verde proper. Other Pueblo I Black-on-white types, Piedra, Kana-a, Rosa, etc., do not appear, on the basis of present research, to have gained much popularity.

La Plata and/or Bluff Black-on-red—Only a few sherds of red ware, 0.3 percent of the total, were found. These have been lumped under the broad classification of La Plata and/or Bluff Black-on-red. The sherds are too few, too small, and, for the most part, too badly eroded to justify a breakdown. Actually, there seems to be some question as to whether La Plata and Bluff Black-on-reds are distinct wares, and some archeologists find it impossible to distinguish one from the other. Dr. Brew has pointed out "Hargrave's Bluff Black-on-red seems to be identical with the pottery known taxonomically in the Mesa Verde area as La Plata Black-on-red . . ." (Brew, 1946, p. 296). Probably the typical black-on-reds of the Mesa Verde should be classified simply as La Plata.

The red wares which originated elsewhere in Basketmaker III times and came into increasing vogue in Pueblo I would seem never to have been too popular with the Mesa Verde potters. Plate 46, row 1 pictures sherds of red wares.

Variations of the above wares—Two types of sherds were found at Site 16 which are occasionally considered to be distinct, or at least given separate classification or recognition. It is strongly felt that these types are merely variations, both intentional and unintentional, of Lino-like plain wares and La Plata Black-on-white. The two types are as follows:

(a) Fugitive red.—A fugitive red wash was frequently applied to the exteriors of Lino-Gray-like jars and La Plata Black-on-white bowls. Since the wash is not permanent, it is impossible to say what percentage of the wares may have been so decorated originally. A few sherds from Site 16 exhibited a fugitive red exterior wash, and in the case of some, this wash, of a distinctive pinkish cast, disappeared, or almost disappeared when the sherds were scrubbed in clear water. Perhaps many more sherds had this wash than was realized.

(b) Smudged and/or burnished—A very few sherds of the Lino-like ware were either smudged or burnished, or both. This apparently resulted from secondary firing of greasy sherds. A test conducted here demonstrated that plain scraped or roughly smoothed sherds of Lino-like pottery, when burned in a dense fire of pinon-juniper, also will come from the fire as perfect examples of smudged and/or burnished pottery. It is not felt that the few sherds exhibiting these characteristics warrant classification as Twin Trees Plain (O'Bryan, 1950, p. 91).

Pueblo II—Ninety percent of the pottery from Site 16 is assignable to the Pueblo II period, or the Mancos Mesa phase. Two new and distinctive wares characterize the period: Mancos Black-on-white, the diagnostic Pueblo II decorated ware of the Mesa Verde region, and allover corrugated, which replaced the Lino-like plain pottery as the standard utility ware.

Mancos Black-on-white—Forty-four percent of the pottery from Site 16 is Mancos Black-on-white, and of this total, 67 percent of the sherds are decorated. Plates 47 through 53, inclusive, picture sherds and vessels bearing typical Mancos designs.

The all-inclusive, iron-paint pottery known as Mancos was manufactured over a long period of time. Manufacture started sometime around A. D. 900 and continued until at least A. D. 1100, possibly later, when Mesa Verde Black-on-white took its place. While Mancos Black-on-white is variable in all respects and reflects the experimentation so characteristic of the formative pueblo years, it is, nevertheless, an easily recognizable ware.

Possibly a breakdown of Mancos Black-on-white could be made on the basis of design trends and styles, surface treatment and finish, etc., if one were given a site with a good refuse mound and undisturbed occupational levels. Such a breakdown has been forecast by the work of Brew (Brew, 1946, pp. 275-279, and accompanying plates), but inasmuch as no one seems to be in doubt as to the validity of the ware, a breakdown does not seem to be indicated. Furthermore, in view of the present chaotic state of affairs as regards Southwestern pottery classification, subdividing an easily recognizable ware would seem unwise.

No stratigraphic changes could be established in the Mancos Black-on-white from the three Pueblo II levels. Some changes in design styles are apparent but it is felt such changes are attributable to the whims of individual potters and do not reflect any basic trend. This feeling is strengthened when a sherd by sherd comparison is made of the pottery from the three levels. For example, the Mancos Black-on-white from the floors of Towers A, B, C, and Kiva 1 have an allover appearance of being earlier than the Mancos sherds from the floors of Kivas 2 and 3, and the floors of the rooms in Unit Pueblo No. I. One would expect the reverse to be true.

It is not felt the Mancos ware from Site 16 reflects the gradual change to Mesa Verde Black-on-white, as did the pottery from Unit 1, Site 13, Alkali Ridge, excavated by Dr. Brew (Brew, 1946, p. 279). At Unit 1, where occupation continued from Pueblo II into Pueblo III, Dr. Brew was able to trace the shift in fashion which resulted in Mesa Verde Black-on-white, the diagnostic ware of Pueblo III. Dr. Brew emphasizes the change was gradual, so possibly the trend was overlooked at Site 16. Lacking any restorable pieces other than the two pictured on plate 53, the picture presented by fragmentary sherd patterns may be misleading. Had a number of vessels been available for study, the feeling about the designs, style trends and shapes might be different.

Pueblo II Corrugated—Over 45 percent of the sherds from this ruin are corrugated and of the 7 restorable vessels found, 5 are wide-mouthed, narrow-rimmed corrugated jars. These 5 jars are illustrated on plate 55.

Like its companion ware, Mancos Black-on-white, the corrugated utility ware is exceedingly variable. In fact the descriptive term, "Variable," (Reed, 1943, p. 144) is applied to the ware, while another popular designation is "Exhuberant." (Roberts, 1935, p. 13). Either name is well-chosen. It seems the Pueblo II potters, once they mastered allover coiling and indenting, went to extremes to achieve the most elaborate decoration possible with their new techniques. It is difficult to explain this sudden vogue for decorating culinary vessels, nor is it understood why such a technique, requiring considerable skill, was employed.

Plate 54 illustrates a few of the effects achieved by manipulation of the coils.

Tables 4, 5, 6, 7, and 8, pages 83, 85, and 86, tabulate the results of the analysis of the sherds from Site 16.

The Pueblo Indians were farmers, dependent primarily upon the crops they raised—corn, beans, and squash. However, it is known they took advantage of many native plants yielding edible greens, roots, nuts, seeds, and berries. Wild mammals and birds were hunted and trapped. The turkey and dog were domesticated, but it has not been satisfactorily demonstrated to date whether or not the people of the Mesa Verde ever raised domestic animals for food.

As is the case in most shallow surface sites, little evidence was encountered of actual remains of food plants. One burned bean was found; no remains of native plants were recovered. A few animal bones were recovered from the fill of rooms and kivas. These were deer bones for the most part. The trash mound no doubt contains bones of several species of birds and mammals and, had its excavation been feasible, a fair list of animals utilized for food might have been compiled.

One interesting cache of bird, mammal and reptile bones was found in the deep banquette cist on the east side of Kiva 3. The bird and mammal bones, remains of former feasts, were undoubtedly tossed in the cist by the people using the kiva. It is felt the reptile bones are of more recent origin. Probably they are the skeletons (there are vertebrae of two specimens) of snakes that died in a rodent burrow which penetrated the soft fill of the cist.

The finding of bird and mammal bones in the cist brings to mind a present-day Hopi custom. When a feast is held in the kiva, any bird or mammal bones, as well as fruit pits and melon seeds, are placed in a special jar, never thrown in the fire and burned. When the jar is full it is taken from the kiva and the bones and seeds carefully deposited in a special place. This not only serves as an offering to the gods, but shows proper respect for the animals and plants and appreciation of their use as food, hence propitiating their spirits and assuring good hunting and crops in the future.

Dating of the three Pueblo II ruins at Site 16 is approximate and comparative. Architectural developments in Pueblo II were rapid and constant changes were taking place. For this reason it was possible for a village under construction, in which the latest architectural styles were being incorporated, to be out of date as concerns one or more of its features before building was complete. While it is possible to pin down the time of introduction of certain features with a fair degree of accuracy as more and more ruins are excavated, the time of incorporation of these features within individual structures or pueblos is often problematical, and the presence or absence of one or more distinctive features cannot be used to date as early or late any given village within the period. The sum total of all evidence must be considered, then a conservative estimate made. The easiest of all dating is that of superimposed structures, so while an error may be made in the time of actual construction, there can be no doubt about the order of construction. Every effort has been made to be as conservative as possible in the dating of the three superimposed Pueblo II villages at Site 16.

Because of the several reoccupations and subsequent levelings of the site, little wood was obtained suitable for tree-ring dating. The post and adobe village was the only structure destroyed by fire. However, the charred remains were cleared away preparatory to the construction of Unit Pueblo No. I, and the ends of the upright posts remaining in the ground were so rotted they disintegrated on exposure to air. The roof was torn from the four-post kiva, so no timbers were available from that structure. Ends of upright posts in the kiva banquette, ends of others found in the front wall of Unit Pueblo No. I, and remains of roof beams lying over the pilasters and on the banquette of Kiva 3, were all too decayed to be of use. Unit Pueblo No. II and Kiva 1 were robbed of their construction materials. However, when the timbers were wrenched from the roof of Kiva 1, the materials used to fill the spaces between beams, such as slender poles, twigs, etc., slid onto the banquette and pilasters. This material, which later burned, furnished the only datable charcoal from the site.

When Dr. O'Bryan tested the kiva in 1941 he struck the banquette on the northeast side between two pilasters, as the 1950 excavation disclosed the outline of his trench. Under the terms of his permit, Dr. O'Bryan furnished the Mesa Verde National Park Museum with a list of dates he obtained, and this list gives dates for 18 of the 67 specimens taken from Kiva 1. However, the dates are tentitive, and Dr. O'Bryan did not consider their correlation to be satisfactory, so he did not include them in his recently published list of "Tree-Ring Dates for Mesa Verde Ruins." (O'Bryan, 1950, Appendix A, pp. 112-115). This charcoal from Site 16, along with other wood specimens obtained under secretarial permit from Mesa Verde National Park and adjacent Federal lands, has not yet been turned over to the Arizona State Museum and Tree-Ring Laboratory with the rest of the Gila Pueblo collections, but remains for the present in the possession of Harold S. Gladwin. For this reason, it has not been possible to have the charcoal obtained by Dr. O'Bryan studied and checked by Dr. Schulman and his assistants. As the tentative dates are both unpublished and unchecked, they are not included in this report.

Less than a dozen specimens were obtained from Kiva 1 during the 1950 excavation of the structure. The charcoal, all small pieces, was submitted to Dr. Schulman at the Tree-Ring Laboratory, and it was gratifying to have two of the specimens yield dates entirely in keeping with the architectural evidence. Both specimens, with bark attached, gave dates of A. D. 1074 (Schulman, 1951, pp. 28-29; Smiley, 1951, p. 23, No. 88z). Dr. Schulman states that specimen MV—494, is "in extraordinarily fine agreement with the master chronology for some 45 rings or so preceding the bark date of A. D. 1074 . . . and specimen MV—495, with excellent bark attached, is a very short, open sequence, but exactly parallels the outer portions of 494." (Statement from a personal letter from Dr. Edmund Schulman to Park Archeologist Don Watson, dated April 30, 1951, Tucson, Ariz.) All factors considered, these dates come close to being actual construction dates for Kiva 1.

Lacking other dates, the time of construction of the three villages must be estimated. The first village, comprising post and adobe rooms associated with a 4-post kiva, was built very early in Pueblo II, as present evidence points to post and adobe construction having replaced slab, post and adobe construction about A. D. 900, and at this time 4-post kivas replaced deep pitrooms. These architectural changes mark the beginning of the Pueblo II period in the Mesa Verde area. The village is not as advanced as the pueblo at the Twin Trees Site, where crude masonry rooms are associated with a four-pilastered kiva. As Dr. O'Bryan has assigned a date of approximately A. D. 950 to the Twin Trees pueblo (O'Bryan, 1950, pp. 28-35, 107, and Appendix A), the post and adobe village at Site 16 probably was built just about or shortly after A. D. 900.

The post and adobe village burned, and sometime later the site was leveled and Unit Pueblo No. I was built. This pueblo, as has been pointed out, exhibits considerable architectural advancement when compared to the Twin Trees pueblo. Just south of the Twin Trees pueblo is Site 1, also excavated by Dr. O'Bryan. The unit pueblo at Site 1 is more advanced architecturally than Unit Pueblo No. I at Site 16. Dr. O'Bryan assigns an approximate date of A. D. 1025 to the Site 1 pueblo (O'Bryan, 1950, pp. 50-51, 107, and Appendix A), so it is reasonable to suppose the first unit pueblo at Site 16 was built sometime after A. D. 950 and prior to A. D. 1025. For this reason, approximate dates of A. D. 975 to 1000 are assigned Unit Pueblo No. I.

Unit Pueblo No. II was obviously built late in Pueblo II. The masonry of the village is approaching that used in the construction of compound pueblos early in Pueblo III, and the presence of circular towers, considered a late Pueblo II development, tends to confirm the evidence of the masonry. The best evidence for late construction, however, is Kiva 1. This well-advanced structure could easily pass for an early Pueblo III ceremonial room if the associated pottery had been Mesa Verde Black-on-white instead of Mancos Black-on-white. The tree-ring dates of A. D. 1074 from Kiva 1 may be considered, in view of the architectural evidence, as approximate construction dates for Unit Pueblo No. II.

A terminal date of around A. D. 1100 may be assigned Site 16 at present, subject of course to revision in light of future excavation. The last occupation apparently took place by, or shortly after, A. D. 1074, and it is presumed this occupation lasted for at least a few years. However, the most significant evidence for the time of abandonment is negative: no sherds of Mesa Verde Black-on-white pottery, the diagnostic ware of Pueblo III in this area, were found at the site. Manufacture of this ware is considered to have started by, or shortly after, A. D. 1100. This does not agree with Dr. O'Bryan, who assigns early Mesa Verde Black-on-white, which he calls McElmo Black-on-white, a beginning date of A. D. 1050. (O'Bryan, 1950, p. 109). A revision of Dr. O'Bryan's date for McElmo Phase seems indicated in view of the A. D. 1074 date from Site 16, a ruin centrally located in the once densely populated Chapin Mesa area of the Mesa Verde, which yielded not one sherd of Pueblo III Black-on-white, whether it be called McElmo, Early Classic, or just Mesa Verde Black-on-white. Since the Mancos Black-on-white from Site 16 does not reflect any of the changes which marked the transition to Mesa Verde Black-on-white, it is unlikely that any Pueblo III Black-on-white was being manufactured in the Mesa Verde prior to about A. D. 1100.

Site 16 was occupied, intermittently, over a period of several hundred years, from sometime in Basketmaker III to very late in Pueblo II, and final abandonment took place prior to the beginning of Pueblo III. The 1950 excavation was confined to the Pueblo II occupational level since the purpose was to make available for exhibit and use in the interpretive program a pueblo intermediate in type and in time between the Twin Trees Developmental unit and the Early Classic Sun Point Pueblo.

Three Pueblo II villages were uncovered instead of the 1 expected, and these 3, superimposed as they are, make Site 16 one of the finest comparative exhibits available in the field of Southwestern archeology. The 3 ruins demonstrate progressive architectural development, and since the sherds from all 3 villages are predominantly Mancos Black-on-white, and there is no later ware present at the site, the pottery proves this development took place within the limits of Pueblo II.

With the last occupation occurring late in Pueblo II, the value of Site 16 is its clear-cut demonstration of the extent of architectural development which took place in the northern San Juan area in Pueblo II times. The evidence from the site is proof that Pueblo II saw the introduction of most of the architectural traits which reached their ultimate refinement in Pueblo III, and demonstrates conclusively that the almost complete development of the Mesa Verde kiva took place within the period.

The material evidence from the ruin parallels that from other excavated sites in the Mesa Verde region which are assignable to the period. The diagnostic decorated ware is Mancos Black-on-white. This pottery, though variable in the extreme as regards decoration, remains unchanged throughout the occupational span, a distinctive and easily recognizable ware. Except for the greater prevalence of full-grooved axes and hammers, and the indicated use of flat-slab metates, stone tools differ little from those in use in Pueblo I. This is also true of bone tools, which are like those of the preceding period. It is interesting to note that not one artifact manufactured from a turkey bone was recovered, and it is perhaps significant that the Mesa Verde Museum collections do not contain a single turkey bone artifact of Pueblo II origin. This may support the current theory that the turkey was raised only for its feathers prior to Pueblo III times. Figure-8-shaped carved shell beads constitute the only new item recovered from Site 16, new only for Pueblo II in the Mesa Verde area, as they have been found in Developmental sites elsewhere in the Pueblo region.

Of more than passing interest is the evidence encountered in excavation of the thoroughness with which the buildings had been stripped of their constructional materials. Only the bases of the masonry walls of the last 2 pueblos and 3 towers remained, and very few building stones were encountered in the fill. Excavation also revealed that the roofs had been torn from both Kiva 1 and Kiva 2. Robbing abandoned sites of their building stones and roof timbers appears to have been a common practice in prehistoric times. Lack of appreciation of this fact has resulted in considerable misinterpretation of tree-ring data and is largely responsible for the seeming confusion in period dating in the Mesa Verde region.

The trash mound at Site 16 is large, but not particularly deep, the greatest depths encountered in the limited testing varying from 18 to 24 inches. In view of the fact that each village was the home of, at the most, only a few families, the absence of any appreciable depth of trash is not necessarily indicative of short occupations. No burials were excavated, but testing of the thoroughly pot-hunted debris indicated it was used for interment of at least some of the dead. It might be stated in this connection that burials were not encountered beneath the floors of rooms, towers, or in the open court areas.

Site 16 has contributed more to an understanding of the Pueblo II period than any site so far excavated in the Mesa Verde. Because of the time of the occupations, one at the beginning, one in the middle and one toward the end of the 200 year interval, and because there was no subsequent occupation in Pueblo III to confuse the evidence, Sire 16 affords the best limitations to date on what should be included in a definition of the period for the Mesa Verde. Briefly summarized, the evidence gives the following general outline for Pueblo II:

1. The period is characterized by several architectural mediums, post and adobe, single and double-coursed stone masonry, and by extremes in architectural experimentation. Unit type pueblos, with the kiva located outside and south of the house block, are typical.

2. The Mesa Verde kiva made its appearance at the beginning of the period, and by the end of the period was almost completely developed and highly standardized. Kivas in all stages of development, from 4-post structures to masonry-lined, 6-pilastered kivas with southern recess, are characteristic of Pueblo II.

3. A unique and typical Mesa Verde structure, the circular tower, also made its first appearance in Pueblo II, probably late in the period.

4. Two wares characterize the period: diagnostic Mancos Black-on-white and Variable, or Exuberant Corrugated.

5. Pueblo II is marked by very little change, except for pottery, in the nonperishable material culture. No artifacts are found, with the possible exception of flat-slab metates and figure-8-shaped shell beads, which were not present in the preceding Pueblo I period.

Site 16, with its superimposed structures, is the most valuable interpretive exhibit in the park. The site is and will be, with the hoped-for excavation in the future of the Pueblo I and Basketmaker III structures known to be present, an outstanding exhibit of architectural development. However, its greatest interpretive value lies in the story it tells of people, of successive groups of people who chose to live, time and again, where other people had lived before them. The superimposed structures are testimony to the continuing efforts expended by these people to improve their homes and enhance their religious edifices, of their growing interest in their religion and devotion to its causes. The story of Site 16 is a story of human advancement, and as such this insignificant appearing ruin adds its bit to the history of man.

Table 5.—Sherd Percentages by Excavated Unit—Site 16
[Based on the total from each excavated unit]

¹1 This includes the sherds from the post and adobe village for the village. was located solely as a result of stripping the area.
²The higher percentage of Basketmaker III—Pueblo I sherds in the fill of Kiva 2 is to be expected. This kiva was backfilled with the material from the excavation of Kiva 3. As pointed out, Kiva 3 apparently cut a Basketmaker structure and the material would have been thrown into Kiva 2. Undoubtedly debris from the leveling of the burned post village was also thrown into this kiva and the cleanup of the site would naturally include some trash from older levels.

Table 6.—Sherd Percentages by Periods—Site 16
[Based on the total sherds of each ware and the totals assignable to each period]