PRELIMINARY ACCOUNT OF THE ANTIQUITIES OF THE REGION BETWEEN THE MANCOS AND LA PLATA RIVERS IN SOUTHWESTERN COLORADO BY EARL H. MORRIS 155 CONTENTS Preface KJl Description of the region 163 I. The cliff-ruins of Johnson Canyon 164 Buildings 165 1. In Johnson Canyon 165 Ruins at mouth of Spring Canyon 165 Ruin No. 3 165 2. In Lion Canyon 166 Eagle Nest House 166 Ruin No. 5 171 Ruin No. 6 171 Ruin No. 7 173 Ruin No. 8 174 Artifacts 174 Pottery 174 Structure 174 Coil ware 175 Plain smooth ware 175 Decorated smooth ware 175 Red pottery 175 Pottery mending 176 Pottery designs 176 Stone implements 176 Grinding stones 176 Axes 177 Potlids and griddles 177 Bone implements 177 Wooden objects 177 Fire sticks 178 Quiver 178 Hairbrush 178 Pot rests 178 Matting 179 Feather cloth 179 Cloth 179 Baskets 179 Sandals 179 Miscellaneous articles 180 Summary 180 157 158 CONTENTS Page II. Ruins oQ the mesas 182 Buildings and burial mounds 182 1. Ruins near Mancos Spring 182 Ruin No. 9 182 Ruins at Site No. 10 186 Ruins at Site No. 11 187 Ruin No. 12 188 Ruin No. 13-. 188 2. Ruins on the divide between Salt and Grass Canyons 188 Shrine at Site No. 14 188 Ruin No. 15 189 Ruin '.No. 16 189 Ruin No. 17 190 3. Ruins south of Red Horse Gulch 192 Ruin No. 18 192 4. Ruins at Red Mesa 192 Ruins at Site No. 19 192 Stone ruin at Site No. 20 192 Round tower at Site No. 21 194 5. Ruins below the mouth of Long Hollow 194 Ruins at Site No. 22 1 194 Ruins at Site No. 23 194 Artifacts 195 Pottery 195 Structure 195 Form 196 Food bowls 196 Globular bowls 196 Bowl with perforated ears 196 Bowl with double flare 196 Globular vessels with wide mouths 196 Undecorated water jars 197 Gourd-shaped bottles 197 Ladles 198 Spoon or paddle 198 Bird-form vase 198 Ring-bottomed vase , 198 Lamp (?) 198 Mountain-sheep effigy 198 Cloud blower 198 Color 198 Decoration 199 Stone implements 200 Grinding stones 200 Slabs of undecided function 201 Axes and hammers 201 Pottery smoother 201 Pounding stones 201 Chipped artifacts -...,, 201 Bone implements 202 Summary 202 Conclusions .' 204 Bibliography 206 ILLUSTRATIONS PLATES ^ Page 31. Map indicating sites mentioned in the text 164 32. Eagle Nest House from mouth of Lion Canyon '. 166 33. Eagle Nest House from foot of cliff at west end 166 34. a, Stone ax with handle of skunk bush, b, Section of wall from Ruin No. 2. 0, Incised tracings on wall of kiva in Eagle Nest House 168 35. a, Kiva in Eagle Nest House showing pottery in situ, b, Kiva in Eagle Nest House 1G8 36. Ruin No. 6 172 37. Ruin No. 7 172 38. Coil-ware ollas 172 39. Coil-ware ollas 172 40. a, Plate, b. Water bottle, c, d, e, Coil-ware ollas 176 41. Black-and-white vases 176 42. a, Black-and-red bowl. 6, c, Black-and-white bowls 176 43. Black-and-white water jar 176 44. Wooden objects 178 45. Bone implements 178 46. a, Rush matting b, Reed quiver 178 47. a, Fire sticks and tinder, b, Hairbrush of pine needles, c, Fragment of a basket 178 48. Jar rests 180 43. a, Feather cloth, b, Matting 180 50. Sandals 180 51 . Sandals 180 52. Miscellaneous objects 180 53. Miscellaneous objects 180 54. a, Ball of yucca, b, A twist of yucca, c, Chain of yucca, d, e, Twists of yucca. /, Bundle of herbs 180 55. a, Portion of plaited band in two colors, b, Twist of yucca, c, Chain of yucca, d, Loop of split willow tied with yucca, e, Fragment of coarse rush mat. /, Corncobs tied together with yucca 180 56. Ruin No. 9 182 57. Grave in Ruin No. 9 184 58. Pit room in Ruin No. 17 190 59. View in Ruin No. 17 190 60. View in Ruin No. 17 190 61. a, Grave below mouth of Long Hollow, Ruin No. 23. b, Grave at head of Salt Canyon, Ruin No. 16 , 190 62. a. Stone ruin at mouth of Cherry Creek, Ruin No. 20. 6, Masonry in stone ruin at Site No. 20 192 63. Large ollas 196 64. Food bowls 196 159 . 160 ILLUSTRATIONS Page 65. Food bowls 196 66. Undecorated vases 196 67. Undecorated water jars 196 68. Cooking vessels with banded necks 196 69. a, Bowl mended with yucca ties. 6, c, d, e, /, Pottery 196 70. Pottery and skulls 196 71 . Pottery 196 72. Pottery 196 73. Stone objects 200 74. Chipped implements 200 75. Chipped implements and polished ornaments 200 TKXT FIGITRES 1 Ground plan of Eagle Nest House 167 2. Burial mound at Ruin No. 9 184 3. Ground plan of Ruin No. 17 190 4. Ground plan of Ruin No. 20 192 5. Outlines of gourd-shaped vessels 197 6. Metate and mano 200 7. Design on bowl from mouth of Long Hollow 200 8. Design on bowl from mouth of Long Hollow 201 9. Design on bowl from mouth of Long Hollow 201 10. Design on bowl from mouth of Long Hollow 201 11. Design on bowl from mouth of Long Hollow 202 PEEFACE In the spring of 1913, at the suggestion of Dr. Edgar L. Hewett, director of the School of American Archaeology, the Board of Re- gents of the University of Colorado placed in my hands the means with which to conduct excavations among the ruins in the region be- tween the Mancos and La Plata Rivers. The permit from the Sec- retary of the Interior was obtained through the School of American Archaeology, hence the work during the summer of 1913 is officially recorded as having been done in collaboration with that institution.^ As a result of the first season's explorations, I was sent back to the same field, where I conducted excavations during part of the summer of 1914. In this research the School of American Archaeol- ogy did not collaborate. Because of limited means, the explorations were not so thorough nor so extended as it woidd be desirable to have made them. Time could not be spared to draw plans of all the ruins visited, and those which are given are compiled from measurements taken with a tape- line. In many places it has been necessary to use the terms " about," "roughly," and "approximately" where exact determinations could have been made only by the expenditure of considerable time and money. Wliatever of worth was accomplished depended largely upon those who assisted me, and I wish here to express my thanks to William E. Ross, E. K. Hill, and J. H. Lavery, all of Farmington, New Mexico, for their faithfulness to the work in hand under all cir- cumstances. Mr. Ralph Linton, of Swarthmore, Pennsylvania, con- tributed his services during part of the summer of 1913. Mr. A. B. Hardin, of Denver, Colorado, directed me to several of the most important ruins and furnished valuable information as to the loca- tion of springs and trails. ^Bulletin of the Archwological Institute of America, Vol. IV, Nos. II and III, p. 41. 74936° —19—33 eth 11 161 162 ANTIQUITIES OF SOUTHWESTERN COLORADO [eth. ann. 33 I am especiiilly indebted to Prof. Junius Hendei-son, curator of the Museum of the University of Cohu-ado, for tlie use of liis office and photogi-aphic equipment while preparing this report, as well as for many other serAices which he has rendered. Since the excavations had to do with two very different types of ruins, I have treated each separately. By describing each type of building and the artifacts therefrom as a unit, a much better com- parison of the culture of the ruins in the cliffs with that of the i-uins on the mesas can be made than would otherwise have been possible. E. H. M. PRELIMINARY ACCOUNT OF THE ANTIQUITIES OF THE REGION BI^TWEEN THE MANCOS AND LA PLATA RIVERS IN SOUTHWESTERN COLORADO By Earl H. Morris DESCRIPTION OF THE REGION The region here dealt with consists of a triangular plateau bounded on the west by the Mancos Canyon, on the east by the La Plata River, and on the south by the Colorado-New Mexico line. Its ele- vation varies from 6,000 to 7,000 feet above the sea. It is traversed from northeast to southwest by a low divide composed of a series of broken hills. The canyons which drain to the Mancos are exceed- ingly deep and rough, rivaling those on the opposite side of the river. The arroyos running to the La Plata are less precipitous and much of the country on that side of the divide is a rolling tableland. An unusually dense forest of pifion and cedar covers much of the region, and the parts not covered by forests are overgrown with sagebrush. Along the watercourses are cottonwoods and willows, and in the canyonsdrainingto the Mancos quaking aspens, wild goose- berries, and chokecherries are of common occurrence. A few rock pines stand at the heads of the canyons, and along the foot of cliffs and in the deep coves are numerous spruce trees, some of them of large size. It appears that the pines, spruces, and aspens, together with the other plants common to the associations in which these are predominant, are lieing slowly crowded out by more xerophytic forms, a condition indicating that there is a less abundant rainfall than there was in times past. Until the coming of the whites, deer, elk, bear, and mountain lions, as well as smaller mammals, were plentiful, and even at present they are occasionally encountered in the fastnesses of the canyons. The sagebrush glades interspersed through the heavy timber fur- nished the aboriginal inhabitants with abundant and fertile land for cultivation. In tlie summer of 1914 corn could have been grown successfully without irrigation upon these mesas. Thus it appears that the region offered all the conditions indispensable to primitive culture. To-day it is uninhabited except for a few " dry farmers," who are endeavoring to reclaim the lands west of Cherry Creek. 163 I. THE CLIFF-RUINS OF JOHNSON CANYON Johnson Canyon is probably the largest of the eastern tributaries of the Mancos Canyon. It begins as a draw at the divide which foiins the boundary between La Plata and Montezuma Counties, and 2 miles farther west drops down between perpendicular clitfs. From this point the bottom is a V-lil^e gorge, often rendered impass- able by great blocks of stone which have broken away from the rim rock and crashed into the watercourse below. Where such is the case the dim trail ascends the steep talus slope, winds along piecarious ledges, and, as soon as there is an opportunity, descends to the canyon floor. In describing the cliff-dwellings of Mancos Canyon neither Jack- son ^ nor Holmes ^ mentions the ruins in this canyon. Nordenskiiild speaks of them as follows ^: The system of eauons southeast of this river [the Mancos] also contains numerous cliff-ihvellinffs of consideralile size. I did not carry out any exca- vations there liut only photographed a iuuiil)er of tlie most important ruins, namely, those in Johnson Cailon. Prudden does not refer directly to the Johnson Canyon ruins but locates several of them on his map of the prehistoric ruins of the San Juan watershed.* Possibly three-quarters of a mile from the beginning of the box canyon the first fork of any considerable size runs off to the north. In it is located Mancos Spring. We found no other permanent water supply between the La Plata and the Mancos which is acces- sible, and in ccmseqvience this spring served as a base for all our operations in the vicinity. There is a large spring some 3 miles down the canyon, but its water is green and unpleasant to the taste. There are also numerous small drips at the base of the rim rock, which doubtless were used by the aborigines. > [Eighth] Ann. Kept, of the Hayden Surv. for 1S74, p. 369, 1876. « Tenth .\nn. Kept, of the Ilayden Surv. for 1876, p. 393, 1878. •The Cliff-dwellers of the Mesa Verde, p. 69. < Tlie Prehistoric Ruins of the San Juan Watershed, pi. xvi. 164 BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 31 Aztec S/3rt7iff • M is 9 a /'' V 6' r d e '\,'' MAP INDICATING SITES MENTIONED IN THE TEXT morris] the cliff-ruins of johnson canyon 165 Buildings 1. in johnson canyon A short distance below the head of the box canyon an ancient trail scales the north wall. It consists of a number of steps or toe holds cut into the rock, which greatly facilitate the ascent of the sloping surface. Under the first arch of considerable size, also on the northern side of the canyon, are two depressions, with a capacity of about 3 gallons each, i^ecked into the rock floor of the cave. They are arranged to catch the drip from the cliff, and a very inconsequential rain is sufficient to fill them. Ruins at mouth of Spring Canyon.—At this place are the ruins marked Nos. 1 and 2 on the map (pi. 31). Both are in a poor state of preservation. The one on the southern side of Johnson Canyon consists of six rooms built in a crevice which can be entered from the east end. The walls are poorly constructed. In one place they are built entirel}^ of mud into which have been thrust many small frag- ments of stone (pi. 34, &), and in another they contain no stone what- ever, but are thickly chinked with broken pottery. They stand upon the edge of the cliff and reach to the roof of the cave. A passage runs the length of the crevice behind the apartments. Below the mouth of Spring Canyon practically every available site contains the remains of a small building. Few of these could have been used as dwellings, the majority probably having served as stor- age places for the crops raised on the mesas. The finding in one of them of several bushels of corncobs strengthens this conclusion. In the first 5 miles below Spring Canyon the party counted 15 of these ledge houses, and it is probable that there are many more hidden by the line of spruces which skirts the rim rock and concealed in the numerous ramifications which branch off from the main gorge on both sides. Euin No. 3.—Under a high arch on the north side of the canyon are the remains of the first building of noteworthy size (No. 3 on the map, pi. 31). An ascent of 300 feet brings one to the level of the cave in which it stands. The debris and ruined walls extend along the cliff for 150 feet. Four kivas form the most conspicuous feature, three of them at the western end of the cave, the other well toward its eastern extremity. In the central part of the rear of the cave a crack was walled up, and the five rooms thus formed are in- tact. Upon a detached bowlder at the front and near the western end perches a tower 7 feet square and 6 feet in height. The features of the rest of the building can not be determined, since even the bases of the walls have been disturbed. 166 ANTIQUITIES OF SOUTHWESTEEN COLORADO [eth. ann. 33 Several sandals, jar rests, and pieces of matting, besides the frag- ments of two pottery bowls (pi. 42, a, c) were gathered up among the fallen stones, a condition indicating that had there been previous visitors to the cave, they were not in search of relics. The red bowl (pi. 42, a) is of particular interest because it so closely resembles the one found by Xordenskiold in Spring House." At least four burials had been made beneath the shelving rocks which litter the floor of the cave.- These had been pawed out by animals, and whatever oU'erings had been placed with them were scattered and destroyed. In one was found the front of a feather- cloth jacket, part of which is shown in plate 49, a. In the kiva, at the eastern end of the building, were the fi-agments of a strangely shaped vessel (pi. 41, 6, c) and a small water bottle (pi. 40, b), as well as several Ijone implements. In a rat's nest, under a great slab of stone which had fallen from the clitf into the northern side of the kiva, were sections of rush matting evidently taken from a large mat cut to pieces by the rodents. (PI. 49, b.) The easterimiost of the three kivas, at the western end of the cave, had been dismantled and used as a dumping place. The floor was covered to a depth of 18 inches with house sweepings, turkey drop- pings, innumerable bits of string, knotted strips of yucca leaves, feathers, and fragments of pottery. In one of the banquettes were a few fragments of the red bowl mentioned aliove. In the next kiva a beautiful bowl was found (pi. 42, i), but seepage had deslroyed any perishable objects which the room may have contained. Because of dampness the fourth kiva was not disturbed. The kivas present no unusual features, so I sliall not describe them, letting the one in Eagle Nest House stand as a type for all those in Johnson Canyon. 2. IN LION CANYON Eagle Nest House.—About three-quarters of a mile below Ruin No. ;5 Jolinson Canyon is joined from the north by a short and very rugged tributary known locallj' as Lion Canyon. At the junction the canyons are 500 feet deep. Where the west wall of Lion Can- yon rounds off and merges into the north wall of Johnson Canyon the rim rock forms a high arch, which shelters a cave of consider- able proportions. Some GO feet from the bottom a shelf crosses the rear wall of the cave. It is 20 feet wide at the east end, becoming graduallv narrower toward the west until it runs out against the perpendicular cliff. Upon the shelf stands Eagle Nest House. No ruin in the Mesa Verde presents a more picturesque and majestic 1 The Cliff-dwellers of the Mesa Verde, pi. xxxiil and p. 84. ' Nordonskiuld mentions such burials (op. cit., pp. 46, 47). BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 32 EA^jLh Wtbi HuUot FROM MOUTH OF LION CANiiji MORRIS] THE CLIFF-RUINS OF JOHKSON CANYON 167 appearance than does this building, when on i-ounding the bold promontory, at the fork of the canyons, it bui-sts upon the view, perched like the nest of a bird upon the precarious ledge. (PI. 32.) Nordenskiold shows this structure, to which he refers thus: "A figure of one of them is given here (fig. 40) as an example of an inaccessible, or at least almost inaccessible, cliff dwelling."^ So much was I impressed with the nestlike appearance of the ruin that I named it Eagle Xest House, and so refer to it in all my notes. I have found no mention of it except that made by Norden- skiold, and I do not believe any name had been previously applied to it. A hard but not dangerous climb of 400 feet brings one to the base of the cliff below the ruin. Here the observer is impressed with the force of Nordenskiold's statement, for the ruin seems indeed inaccessible (pi. 33). The cliff overhangs above and below the shelf which supports it, and as the distance is too gi-eat to permit Fig. 1.—Ground plan of Eagle Nest House. the casting of a rope over one of the protruding beams, direct access is impossible. However, from the east end of the ledge a crevice continues along the cliff for some distance. Near its end the wall below drops back to perpendicular. Here two large poles had been leaned against the cliff and fastened to the stump of a cedar which had grown conveniently at the bottom. I climbed to the end of these, pushing a. pole ahead of me until only 3 feet of it overlapped the top of the first pair; after lashing this to them and binding another pole beside it I clambered up these and repeated the process. The top of the fourth pair of poles reached to the ledge. Even after they had been securely fastened at the top it was not until the next day that my workmen could be prevailed upon to attempt the ascent. The i-uin contains 12 rooms and a kiva (fig. 1). At the east end the outside wall of the house widens into a stout' pillar built from >Op. clt, p. 69. 168 ANTIQUITIES OF SOUTHWESTERN COLORADO [eth. ANN. »a the ledge to the rock above. Behind the piUar, iiich^sed by the outer wall of the house on the left and by the front wall of room 11 on the right, is a passage or entry which ends in a series of steps leading up to what' w-as the level of the kiva roof. This and the space which is dotted in the plan constituted a plaza quite large and commodious in view of the small pi-oportions of the building. It is probable that the roof of room 11 was part of the plaza also. The open side of the court is flanked by a parapet 2i feet high. From the plaza a T-shaped doorway leads into room 1, which, being in as perfect a state of preservation as any room in the Mesa Verde, is worthy of description. Its inner dimensions are, parallel to the cliff, 5 feet G inches by 6 feet G inches in the opposite direc- tion. The height to the ceiling is 5 feet 7 inches. The walls bear successive coats of brown plaster, a new coat having been added, seemingly, when the one beneath became covered with soot and dirt. The roof is supported by two comparatively heavy beams, which run t'he long way of the room and are set into the walls. Upon these at right angles rest four smaller poles, which are covered by a layer of closely placed split sticks, and above them is a layer of indurated mud. In the southeast corner is a fire pit 18 inches in diameter. There is a smoke hole in the roof immediately above it, and the walls in that comer are black with smoke. Upon the roof is a flat slab, which was used to close the opening when there was no fire on the hearth. In the south wall 1 foot 9 inches from the west wall and 2 feet 10 inches above the floor is a neatly j)lastered niche 3^ inches in di- ameter and 4 inches deep. In the southwest' corner near the top of the south wall is a somewhat larger niche, and tliere is still an- other in the north wall 1 foot 7 inches from the northeast corner and 1 foot 6 inches up from the floor. In the northeast corner a small osier eyelet protruded from the wall through which was looped a long strand of yucca cord. Upon the floor were two bone needles. Between the rear wall of I'oom 7 and the cliff were the remains of a burial, which had been disturbed by some agency. A few frag- ments of matting were with the bones. Rooms 8 and 9 contained grinding stones, fragments of pottery, bits of string, and a f(^w bone implements. Room 11 seems to have been the kitchen. Upon the floor were three sets of millstones, and against the west wall were the remains of at least five coil-ware cooking pots, one of which is shown restored in plate 38, b. In the rubbish were the fragments of a baking slab. BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 34 a. STONE AX WITH HANDLE OF SKUNK BUSH b. SECTION OF WALL FROM RUIN NO. 2 c. INCISED TRACINGS ON WALL OF KIVA IN EAGLE NEST HOUSE BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 35 a. KIVA IN EAGLE NEST HOUSE SHOWING POTTERY IN SITU b. KIVA IN EAGLE NEST HOUSE MOBKIS] THE CLIFF-EUINS OF JOHNSON CANYON 169 From the fire pit were taken a hairbrush (pL iT, b) and three sandals. Behind the bench which crosses the east end of the chamber and beneath the southeast corner of the wall was a stone ax (pi. 3-i, a), with its skujik-bush {Schiiialtzia trilohata) handle still attached. The small room at the east end was a storehouse. The walls extend to the rock above, and so little light enters through the small door in the east end that the interior is always dark. It appears that much labor was expended to retain the subter- ranean character of the kiva. As the presence of the ledge made excavation impossible, the space from the foot of the steps to the west wall of room 1 and back to the cliff was filled with loose rock and debris in order that the roof of the kiva might be on a level with the floors of the surrounding rooms. This does not apply to room 11, but doubtless there was a limit beyond which economy of space would not allow the builders to go, even though in conse- quence custom had to be somewhat violated. The kiva was constructed as follows: Except on the north, where the cliff interfered, two walls were built, one within the other. The outer wall was carried up to the desii-ed level of the plaza, while the other was brought up only 2^ feet. Upon it were erected the ped- estals which separate the banquettes and serve to support the roof. The outer wall forms a back for the banquettes and functions as a brace for the pedestals. The roof had fallen, but the beams were sufficiently in place to show that it had been constructed in the same manner as the one figured by Dr. Fewkes,^ so I shall not de- scribe it here. Otherwise the kiva was in an almost perfect state of preservation. In removing the debris three coil-ware jars (pis. 38, a; 40, c, d) were found against the west wall (pi. 35, a). The largest of these was in fragments, but the others were unbroken. With them were parts of two other large pots and toward the center of the room were two small dipper bowls. The measurements of the kiva are: Height, 8 feet 3 inches; diam- eter, 12 feet 9 inches; height of floor to banquettes, 2 feet 6 inches; height to top of pedestals, 4 feet 8 inches ; width of banquettes above liorizontal passage, 4 feet 5 inches; width of other banquettes, 3 feet; depth of banquettes, 11| inches; width of pedestals. 1 foot 8 inches; distance of deflector from wall, 2 feet 2 inches; height of deflector, 2 feet; length of deflector, 2 feet; thickness of deflector, 8 inches; distance of fire pit from inside of deflector, 2 feet; diameter of fire pit, 1 foot 10 inches ; first sipapu, 9 inches from pit ; second, 9 inches from first; height of horizontal passage, 1 foot 5 inches; width, 1 1 Bull., it. Bur. Amcr. Ethii., pi. 15. 170 ANTIQUITIES OF SOUTHWESTEBST COLORADO [ETn. ann. 33 foot 2 inches; bottom, 4 inches above floor; length of horizontal passage, 2 feet 3 inches; depth of ventilator shaft, 8 feet 3 inches. Two sticks crossed at right angles are set into the masonry just below the top of the air shaft. Resting upon these was a block of stone which closed the opening and came almost flush with the level of the i^laza. In the east wall a few inches above the floor is a niche or " cubby- hole" large enough to contain a fair-sized jar (pi. 35, h). An unusual feature is the presence of a small niche in the fireward side of the deflector. I have found no mention of a nithe similarly placed in any kiva in the Mesa Yerde. The presence of the two sipapu seems to render the kiva rather unusual, as only one other instance of the kind is on record.^ Somewhat more than a foot to the east of the first sipapu a mano was tightly plastered into the floor. The floor and the first 17 inches of the walls are plastered with brown clay. Higher up the walls are white and show few evidences of smoke. At the junction of the two zones is a dado like the one figured by Dr. Fewkes from the third story of the square tower in Cliff Palace.^ (See pi. 35.) Beneath each banquette three clay- colored triangles extend up into tlie white, and lietween the series of lai-ge triangles are 29 to 34 smaller figures, such as could be made by a single dab of a brush. Nordenskiold shows practically the same decoration from a kiva in a ruin in Cliff Canyon and mentions having observed it also in two other ruiiig.^ There are numerous incised tracings in the white plaster of the upper walls. Those in the surface of a pedestal at the west side are shown in plate 34, c. In order to photograph these I traced them with charcoal, taking care not to add anything to the original. The masonry of Eagle Nest House is in places good, in others mediocre. Some of the walls toward the western end give evidence of hasty or careless construction. However, room 1 is as well built as are the better parts of Cliff Palace. The T-shaped doorway in the east end excites one's admiration. The sides are so smooth and the angles so true that they might well be the work of a modern mason with his chisels and squai-e. It appears that the stones were rubbed smooth after they were put in place. " It is doubtful whether there can be found in any of the subdivi- sions of Mancos Canyon a better example of a " unit-type " cliff- dwelling than is present by Eagle Nest House.* The alignment 1 FewkPK Bull. !il. Bur. Amer. Etlin., p. 18. -Bull. ')1, Bur. Amer. Ethn., pi. 13, a. 'Op. clt., p. 16. * .\ flrfinition and explanation of this torni may be found in Pruddon, Preliistoric Ruins of the San Juan Watershed, p. 234. MORRIS] THE CLIFF-RUINS OF JOHNSON CANYON 171 ol' dwelling rooms, kiva, and refuse heap found in buildings in the open is not i^reserved here, since by force of necessity the builders were compelled to conform their plans to the site upon which they built. The ruin is a '' unit-type " dwelling adapted to a sijecial site.^ It presents all the essential features: A kiva subterranean in signifi- cance if not in fact, and a series of chambers, part of them living rooms, and the rest used for storage purposes. It seems that the ratio here presented is: Living rooms, 11; storeroom, 1; kiva, 1. However, certain of the 11 rooms may have been, and probably ^were, used as storerooms. From the broken pottery strewn down the slope below, it appears that the refuse was cast over the cliff. It is impossible to say what disposition was made of the dead. Ruin No. 5.—There are four other ruins in Lion Canyon woi-thy of mention. Following the base of the rim rock 10 minutes' walk from Eagle Nest House one arrives at the site of Ruin No. 5. This stands under a high but shallow arch, which does not protect all parts of it from the elements.^ Four rooms exhibiting very good masonry stand at the foot of the cliff, and the presence of large (juantities of worked stone, as well as of roof beams and floor beams, scattered down the slope indicates that these rooms represent but a small part of the original building. The one kiva visible is at the northern end of the cave. Rains have beaten in upon it until the walls are denuded of plaster and mortar, and it is more than half full of debris from the walls and roofs of neighboring rooms. The parts which extend above the wreckage indicate that this kiva varies in no jDarticular from the one just described. Some 20 feet above the lower ruin a ledge extends around the entire arc of the cave. At the south end, where this is slightly- broader than at any other part of its length, stands a cluster of 10 or 11 rooms. From these a rough, mortarless wall continues to the north end of the crevice. It is probable that the inhabitants of the lower dwelling intended to add to the house Ijegun at the south end and hoisted the rack of loose stone to the ledge for that purpose. Ridn- Xo. 6.—This ruin (pi. 36), the largest cliff dwelling in John- son Canyon or any of its tributaries, is on the same side of the canyon, a few hundred yards above Ruin No. 5. The loose and unstable condition of the detritus upon which it is built and the easy approach to the ruin account for its deplorable condition. It extends along the cliff for more than 200 feet and contains 6 traceable kivas and 31 rectangular rooms. The floor of the cave is very uneven and the walls have been built around and upon detached masses of stone, in many cases on sloping surfaces, with gi'eat care and considerable skill. In jDlaces they rise to a height of three stories, and marks on 1 Fevvki'S, Bull. 1,1, Amer. Ethn., p. 8. 172 ANTIQUITIES OF SOUTHWESTERN COLORADO Ieth. ann. 33 the cliff above show that originally they were surmounted by a fourth story. The great piles of fallen masonry indicate that the entire building was two or more stories in height and probably contained as many as 80 rooms. Because of the great quantity of accumulated debris, the determination of the features of the building and the relation of its parts was too great a task for the expedition to under- take because of its limited funds. The deflectors in two of the six kivas examined are constructed of poles 1 to 2 inches in diameter set into the floor and bound to- gether with willows. These are heavily coated with plaster. Nor- denskiold writes as follows:^ As far as I coukl ascertain by a hurried investigation, the ruins in .Johnson Canon differ in no essential respect from tlie other cliff dwellings on the Mesa Verde. Estufas are present In all the larger ruins and preserve in all respects the ordinary type. I observed one single exception wliich affected only an un- important detail. In one estufa the low wall . . . consisted not of stone, as is usually the case, but of thick stakes driven into the ground close to each other and fastened at the top with osiers. On the side nearest to the hearth this wooden screen was covered with a thick layer of mortar, probably to protect the timber from the heat. It is probable that Nordenskiold refers to one of the kivas in this ruin. The deflector in Kiva K, Cliff Palace, is constructed in the same manner.^ The ruin had been thoroughly ransacked by relic hunters many years before it was visited by the author. Although practically every nook and cranny had been pried into, a few good finds were made. At the southern end a kiva is built in between large bowlders, which have broken away from the cliff above. On top of one of the pilasters and scattered over the debris beneath were many fragments of a large water jar. The floor was cleared in an effort to find enough sherds to make possible a restoration (pi. 41, a). "When tapped with a shovel handle the south half of the floor sounded hollow. The plaster when broken through was found to be resting upon a mass of di\v grass and twigs. Evidently refuse had been thrown into the south side of the room to bring the floor up to the le\el necessitated by the presence of a shelving rock on the north. From the trash were recovered six sandals, a quiver, several jar rests, a wooden hoop with a netlike attachment, some fragments of a most excellent basket, and about 2 quarts of corn, the germs of which had not been destroyed by mice or weevils. A square room was perched on the top of a large V)owlder west of the kiva. Hidden beneath the floor in the northwest corner were two large coil-ware ollas (pi. 39). Over the tops of both were thin stone slabs and across the neck of one corncobs had been placed, the ends I Op. cit.. p. 70. = Fewkes, Bull, il, Bur. Amer. EDin., p. 57. BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 37 RUIN NO. 7 BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 38 BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 39 COIL-WARE OLLAS MORRIS] THE CLIFF-RUINS OF JOHNSON CANYON 173 resting against the flare of the neck. The space above these was filled with clay. Within was about a quart of fine dust not derived from any organic material : hence the reason for sealing the jar is difficult to imagine. The transportation of these large pots down the pre- cipitous cliff and back to camp at Mancos Spring was no small under- taking, as a slight blow would have reduced them to fragments. One was tied in a gunny sack and the other in a shirt, and after much labor they were deposited safe at camp. A'uhi lYo. 7.—Ruin No. 7 is in a deep pocketlike cavern less than a ciuarter of a mile up the canyon from the ruin just described. The building consists of four groups of rooms somewhat separated from one another. The first to be reached on approaching the ruin from the south contains six rooms, which have been formed by walling up and partitioning off a deep crevice. The walls, which are intact, reach up to the rock. Eight)' feet farther north is the central and most important part of the ruin. In this are seven rectangular rooms and two kivas. As may be seen in plate 37, &, one room is in the second story, the walls reaching to the top of the cave. The floor dividing the stories has fallen. A short distance below the top of the walls four stout beams are set into the masonry, forming a square slightly smaller than the room itself. Some object seems to have been susi^ended from these beams, but there is nothing to indicate what this may have been. The kiva which appears in the foreground is nearly filled with debris; this was not excavated. The inclosure between the kiva and the two-storied part of the ruin is of exceptional interest, as it is a rectangidar room which in many features resemljles a kiva. The comers were filled to a height of about 3 feet with masonry, giving the room an oval instead of a rectangidar form. Against the outside of the east wall a buttress of masonry was constructed, into which the horizontal opening extends and through which the ventilator shaft rises. The deflector, a slab of stone, had been broken down, but the fire pit was in the usual position. Xo sipapu was observed, but as the floor was much broken, it may once have been present. There is no trace of banquettes or pilasters, unless the tops of the triangles of masonry in the corners served as l)anquettes. The entire south wall and considerable sections of those on the east and west had fallen, so it was impossible to determine all the features of this singular apartment. This is the only instance obsers^ed in any of the ruins in Johnson Canyon in which a kiva differed from the one in Eagle Nest House in any but minor details. The third section of the ruin is about 100 feet farther along the cliff, where the latter has swung eastward toward the main canyon (pi. 37. a). It consists of a two-story tower, the cliff forming the rear wall, and a series of three rooms extending eastward in line ^ — 174 ANTIQUITIES OF SOUTHWESTERN COLORADO [kth. ANN. 33 with the upper half of the tower. The floor between the stories has been burned away, and the floor of the lower room has been dis- turbed by relic hunters. The fourth jrroup of rooms is situated in a large crevice high above sections 1 and 2. Just north of the firet cluster of rooms is a considerable space almost closed in front by a huge block of stone. In the dust and refuse which partially fill it several burials were nuide. Previous visitors had looted the graves, but part of one skeleton remained in the walled pit in which it had been interred, and bones of others were scattered aliout. It would appear that the first despoilers found many specimens, for large fragments of beau- tiful pottery, parts of a basket, some bits of feather cloth, and part of a split-willow burial mat were picked up among the trash. In the northwest corner of the oval kiva was the greater part of a splendid water jar. a restoration of which is shown in plate 43. Upon a sloping rock in front of the first group of chambers a human hand and a few other pictographs are pecked into the smooth surface. These are figured by Nordenskicild.^ Although there are in Johnson Canyon rock surfaces which offered excellent oppor- tunities for the execution of pictographs, these are the only ones observed. In many places there are grooves and depressions caused by the grinding of axes and awls, but pictographs are notably few. Ruin No. 8.—In a deep cove close-grown with majestic spruces, almost directly across the canyon from Ruin No. 5, Ruin No. 8 is situated. It is small and pi'escnts only one feature worthy of men- tion. The walls of one room are built of poles set upright, bound together with osiers, and thickly coated with adobe plaster. This is a very unusual method of construction in cliff-dwellings of the Mesa Verde, but in northeastern Arizona it is common.'' It is of particular interest here, since, as I shall show later, the walls of the houses on the mesas were built almost entirely in this manner. If there are any ruins of note in the main gorge below the mouth of Lion Canyon, our party failed to find them.* Artifacts POTTERY Structure.—The pottery from Johnson Canyon is of three types coil ware, plain smooth ware, and decorated smooth ware. It ' One of the ruins in tills canyon was the site of the phenomenal find made by the Wetherllls and described by Nordenskiold, op. clt., pp. 46-47. = Ibid., pi. XX. 2. 'Fewkes, Bull. SO, Bur. Anwi: Ethu., p. 14. ' In September. 1915, Mr. N. C. Nelson and the writer found a ruin containing over 40 rooms and 3 kivas at the head of a long but shallow canyon parallel to and west of Lion Canyon. MORRIS] THE CLIFF-RUINS OF JOHNSON CANYON 175 appears that all types were constructed by the- coiling process, the resulting undulations having been obliterated, except upon the exte- riors of vessels of the first type. Coil ^Yare.—The seven coil-ware jars shown in plates 38, 39, and 40, varying in height from 6 to 15 inches, constitute an excellent series. The typical shape is marked by a globular base tapering toward the top and surmounted by a recurved lip upon which the coils have been erased. It is interesting to note that the coil-ware vases never have the concave bottoms found almost without excep- tion in the large black-and-white vessels of the Mesa Verde area. Although decorations other than the crenulations due to structure are seldom found, coiled fillets of clay applied over the ridges appear in plates 38, h, and 39, h. Plain Smooth Ware.—The plain smooth ware is illustrated by plate 40, a. I was at a loss to know what to call this vessel. It is a thick- walled, friable, shallow bowl, upon the interior of which is a layer of indurated ashes growing thicker from the rim to the bottom of the dish. It calls to mind baskets coated with clay which were used by some southwestern tribes as roasters. The material to be parched was placed in the dish together with live coals, after which the re- ceptacle was rotated and the ashes blown out with the breath. In the ruins of the Pajarito Plateau are found similar objects, which served as molds for the bases of large oUas. Decorated Smooth Ware.—Decorated smooth ware is the dominant type of pottery and offers the greatest variety of shapes. In many cases a wash of light-colored earth was applied over the darker paste of the vessel. B}' rubbing with a smooth stone or like object an extremely fine, often glossy, surface was produced. Upon it designs were traced, which were made permanent by firing. Bowls comprise the most typical form, of which those appearing in plate 42, b, c, are characteristic examples. The rims are not titpering or recurved. The large asymmetrical vase shown in plate 41, 6, c, is a unique specimen. The mouth is oval instead of round and the base is deeply concave. Just beneath the rim (pi. 41, &) the coils are still apparent. The surface is not covered with a slip. In plate 43 is shown a water jar with pinkish-yellow and very triable paste. The slip is as white as chalk and superbly polished. The base of this vessel, as well as that of the other large water jar (pi. 41, a), is concave. It would be difficult to find a more beautiful example of the ceramic art of the Mesa Verde. Red Pottery.—Red pottery is extremely rare in the cliff-dwellings of the Mesa Verde. From a few fragments recovered from Ruin ^ 176 ANTIQUITIES OF SOUTHWESTERN COLORADO [eth. ann. 33 No. 3, most of which fortunately fitted together, I restored the bowl shown in plate 42, a. The j^aste is slate-gray in the center, be- coming yellow toward the surface. The slip is a dark brilliant red. The design, traced in black, is a combination of the rain-cloud and bird jDatterns, or at least of the symbols which are so interpreted on pottery from ruins known to be closely connected with recent Pueblo cultui-e. These deep-red bowls with incurving sides and slighth' flaring rims seem to be of a type widely distributed over the Southwest. Hough ^ figures one from Blue River, Arizona, identical in shape, and bearing a design resembling that upon the one here shown. Nordenskiold ^ recovered the fragments of another from the debris in Spring House, and the author found a segment of one in a refuse heap near Farmington, New Mexico. It is obvious that red vessels were highly prized, and it is probable that they were used for cere- monial purposes, a fact which would tend to m;ike them still more precious. For such reasons they would be carried in trade far beyond the boundaries of the ceramic area to which they rightfully pertain. Pottery Meiullng.—The high regard in which the ancients of Johnson Canyon held their pottery is shown by the fact that several of the vessels are carefully mended. The olla figured in plate 39, a, has a long crack across its bottom. Along this opposite sets of holes were drilled and yucca thongs were inserted to bind the seam to- gether, some of these still being in place. In the bottom of the pot shown in plate 40, hra,ymites. Interlaced strips of yucca hold it together, and the bottom is closed with a wad of corn husks. Although it is known that most of the quivers used by the clitf-dwellers were made of skin, it is difficult to assign any other function to the object in question. HAIRBRUSH A hairbrush made of the needles of the rock pine (P/'n-us scopu- lorum) was found in Eagle Nest House (pi. 47, 7/). It is bound about the middle with a three-strand cord of twisted hunuin hair, and still contains a liberal quantity of black and dark-brown combings. POT RESTS In plate 48 are shown five hoops which were used as rests for the bases of large jars. They are made of willow, cedar bark, com husks, some of shredded yucca leaves, bound together with strips of yucca. >Op. cit., pi. 20. BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 45 f^m-- >, l€i BONE IMPLEMENTS BUREAU OF AMERICAN ETHNOLOGY TH I RTY-TH I RD AN NUAL REPORT PLATE 45 1 BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 47 ihlMlB'i'M*' a. __i%vKt<^ is a bundle of corn husks of unknown use. Such bundles are A-ery common. Plates 54, 55. Plate 54, a, is a small ball of finely divided yucca with a minute quantity of yellow earth in the center; 54, h, is a twist of yucca ; 54, c, and 55, f, are chains of yucca ; 54, d, e, and 55, h, are twists of the same material; 54, /, is a bundle of herbs which thus far I have not been able to identify; 55, a, is a portion of a plaited band done in two colors; 55, d, is a loop of split willow tied across with yucca ; 55, e, is a fragment of a coarse rush mat ; 55, /, is a number of corncobs tied together with yucca. The use of none of these is known. Summary From the foregoing discussion it appears that there existed in Johnson Canyon a typical example of the rather resti'icted culture » Bull. 51, Bur. Amer. Ethn., pp. 72-73. BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 48 ^r ' BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 49 a. FEATHER CLOTH 6. MATTING BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 50 ^ m.mm̂m <:^ •m&>^. BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 51 4 BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 54 u. BALL OF YUCCA, b. A TWIST OF YUCCA, c. CHAIN OF YUCCA, il, i. TWISTS OF YUCCA. /. BUNDLE OF HERBS BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 55 r. PORTION OF PLAITED BAND IN TWO COLORS, b. TWIST OF YUCCA, f. CHAIN OF YUCCA. d. LOOP OF SPLIT WILLOW TIED WITH YUCCA. (. FRAGMENT OF COARSE RUSH MAT. /. CORNCOBS TIED TOGETHER WITH YUCCA. UOBEIS] THE CUFF-RUINS OF JOHNSON CANYON 181 characteristic of the Mesa Verde region. The materials for build- ing, weaving, and pottery making were procured in the immediate vicinity, and the fruits of wild trees and plants, as well as the cultivated crops, came with few exceptions from the near-by can- yons and mesas. The general characters of the masonry, the struc- tural features of the kivas and secular rooms, the methods of burial, and the pottery and other artifacts indicate that the cliff-dwellings in this canyon southeast of the Mancos River are culturally and approximately chronologically 'contemporaneous with the large ruins of the Mesa Verde National Park. II. EUIXS ON THE MESAS For a number of years it has been a growing conviction with the author that the faikire to investigate the badly weathered and ap- parently very ancient ruins which dot the mesas of northwestern New Mexico and southwestern Colorado has left unworked one art of kiva 1 were excavated. The north and east walls of room 1 are BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 62 a. STONE RUIN AT MOUTH OF CHERRY CREEK, RUIN NO. 20 b. MASONRY IN STONE RUIN AT SITE NO. 20 MOKKIS] RUINS ON THE MESAS 193 18 inches thick. The wall between room 1 and kiva 1 is 30 inches in thickness and stands to a height of 8 feet. The masonry is ex- cellent (pi. 62, h). The small sandstone blocks are dressed to conform to the curve of the wall and the cracks are chinked with tiny si^alls. Kiva 1 is 15^ feet in diameter. The banquettes, six in number, are unusually shallow, being but 5 inches deep. The ventilator shaft opens to the south, as appears to be the case with the other kivas. Since the exhaustion of funds made it necessary to abandon the excavation of this most interesting ruin, the nature of the other features of the kiva was not determined. The fact that the kivas are above ground impresses one as unusual. 1 have not observed another instance in the San Juan drainage where kivas built in the open were not subterranean. A discovery of the relation between this stone building and the numerous " pole-and-mud " ruins in the vicinity would be of the utmost importance in determining whether or not the ruins in the cliffs and those upon the mesas were built by the same people. The rooms excavated contained no artifacts which would help to settle the question. Slightly northeast of the building is a group of slab-inclosed boxes. These occur singly and in groups all over the mesas, but in order to avoid repetition I have deferred mentioning them until this time. Holmes ' and Jackson ^ describe similar inclosures on the mesas west of the Mesa Verde, and the author has been told that they extend eastward toward the Animas Eiver. Some are nearly round, some square, and others rectangular. Their average size is about 3 by 4 feet. Slabs of stone form the sides and in many cases there is a slab on the bottom. Their use is difficult to determine. They are commonly known as " Indian graves," but there is nothing to prove that such was their function. They occur in and about ruins and in isolated places far from any sign of a building. In many of them there are small quantities of charcoal and ashes and now and then an animal bone. We dug up about 50 of these "graves," but did not find in one of them enough traces of fire to lead to the conclu- sion that a body might have been cremated therein. They may have been fireplaces, but if such were the case it is hardly probable that they would be found so far removed from more or less permanent habitations. It is not to be expected that an Indian would transport heavy slabs of stone a considerable distance in order to construct a cooking place for a temporary camp. '^ Tenth Ann. Rept. of the Ilaydcn Survey, mir,, pp. 3S5-386. 'Ibid., p. 414. 74936° —19—33 eth 13 194 AXTIQUITIES OF SOUTHWESTERN COLORADO [EIH. ANN. 33 Were these " graves " not so numerous, it might be concluded that they were shrines, and it may be that such was the purpose of some of them. Round Tower at Site No. 21.—On the point of a bhiff about a quarter of a mile up the river from the stone ruin there is a small stone tower. The wall is in a poor state of preservation, standing to a height of only 3 feet. From the point there is a good view of the valley and of the broad mesas which sti-etch eastward from the river, hence it seems that the tower served as a lookout station. 5. RUINS BELOW THE JIOL^TH OF LONG HOLLOW Ruins at Site No. 22.—Purely by accident a group of burials was found farther down the La Plata. About three-quarters of a mile below the mouth of Long Hollow the wagon road ascends from the river bottom, runs for a short distance across a point, and drops back to the lower level. Not 20 feet from the road I noticed black earth and fragments of pottery, and the first thrust of a spade brought up human bones. The skeleton of which they were part was flexed with the head to the northwest. South of it were two other bodies, also flexed, one of them that of a child. The point had worn away imtil these skeletons were barely covered with earth, and if any pottery was put away with them it had been broken and the fragments scat- tered. East of them was a skeleton stretched at full length upon its back. By the head was an undecorated vase. North of this burial was a grave which was constructed with more care. In a rectangular pit 2J fe«t deep the flexed body reposed upon its right side with the head to the north. Beneath the skull was a flat stone, and in front of the body were four decorated bowls (pi. 65, c, e, /). In all cases the bodies were covered with black earth and refuse, and it appeared from the ashes, fragments of pottery, and chips of flint that trash must have been dumped on the graves for a long period of time. Southwest of the graves, on the tip of the point, were the remains of a faii'ly large " pole-and-mud " ruin. On a promontory not far from the ruin the Powells, who own the land in the river bottom and on the opposite side of the river, imearthed a skeleton accompanied by a pipe and a string of beads. Several of the beads are turquoise, and one of them, conical in form, is of ivory. It appears to have been made from the canine tooth of a large animal. I did not see the pipe. Ruins at Site No. 23.—There are many ruins on the Powell ranch west of the river. Those chistered along the edge of the second ter- race are all of the " pole-and-mud " type. The bodies of two chil- dren and four adults were found near one of them. One large skele- ton was extended with the head to the east, and the rest were flexed. MORKis] EUIXS ON THE MESAS 195 One grave contained two skeletons interred at different levels. The first (pi. 61, a) la}' upon its left side with the head to the west. Near the skull were three pottery vessels and a worked stone, presumably a pottery smoother, and by the feet were two more vessels. 'Wlien the skull was raised it was found to be resting upon the knees §i another body lying at right angles to the first, with the head to th.t! south. Bj' the skull were three bowls, a lamp (?), and a vase. Not one of the 11 skulls from these two series of graves was flat- tened at the back. Most of these crania were so badly decomposed that they fell to pieces when moved, but three were recovered entire. One appears in plate 70, e. In an examiiuition of more than 200 skulls taken from graves in the valleys of the La Plata, the Animas, and the San Juan Eivers, I have seen no others which fail to show pro- nounced flattening in the occipital region. Upon the first bench above the river begin the cobblestone ruins so numerous farther down the valley. These are, in general, mounds in the form of a semicircle with a kiva between the horns of the half moon, and a burial mound south or southeast of the building. The shape and construction of these buildings, as well as the pottery strewn over them, suggest that they represent a culture differing in many particulars from that characterized by the ruins in the upper valley and on the mesas to the west. Artifacts POTTERY Structube Structurally the pottery from the mesas is inferior to that from the cliffs. In general the paste consists of a fine-grained matrix, through which are scattered many dark-colored granules. This indi- cates an imperfect reduction of the clay or an admixture of a sec- ondary material, possibly crushed potsherds. The color and compo- sition of some of the paste suggest a volcanic rock such as I have not found on the Mesa Verde. The hardness varies greatly. Some vessels are quite friable, while the finer ones are not easilj^ scratched and emit a clear, bell-like tone when struck. I have not been able to demonstrate the presence of a superficial slip on a single specimen. While the surfaces of several are of surprising whiteness, this seems to have resulted from the polishing which brought to the surface the lighter-colored, finer- grained portion of the paste. Although it is probable that all the pottery was constructed by the application of successive coils of clay, from the standpoint of sur- face treatment it may be di^-ided into four classes—(1) smooth ware, (2) smooth ware the examples of which have banded necks. 196 ANTIQUITIES OF SOUTHWESTERN COLORADO [eth ann. 33 (3) smooth iDolished ware without decoration, and (4) smootli polished ware with decoration. True coil ware was not exhumed from uny of the sites, although a few fragments occurred upon the surface. The excellent vessel shown in plate 63, a, although found in the ojjen, was not within the confines of a ruin, and may well have been left at a temporary camp by the cliff people or deposited at a shrine, further evidence of the existence of which has not been preserved. Form It is doubtful whether from any other locality in the Southwest a series of 64: ceramic objects could be chosen at random which would exhibit a greater diversity of form than those shown in the accompanying plates. Food bowls.—Of food bowls there are 16 (pis. 64, 65). In general their shape is that of a section of a hollow sphere, although one (pi. 65, c) has an incurving edge and another (pi. 65, e) has a flat bottom. With few exceptions the sides taper to a thin fragile rim. The interiors show much better finish than do the exteriors. Globular Bowls.—Two of the three globular bowls appear in plates 70, a, and 71, d. The one not figured has a heavy handle attached to one side near the opening; the others have i^airs of j^erforations op- posite each other, through which cords might be passed to carry or to susj^end the vessels. Bowl with Perforated Ears.—In jilate 71, /, is shown a small, very deep bowl with perforated ears. Bowl with Double Flare.—The bowl figured in plate 71, c, is a unique specimen. The constriction in the sides allows it to be held conveniently and securely in the hand. Globular Vessels with Wide Mouths.—Plate 66 illustrates vases with approximately globular bases and mouths of large proportion- ate diameter. One has a single handle consisting of a ridge of clay pinched onto one side of the neck. The recurved neck imparts a pleasingly graceful form to the vase appearing in plate 66, d. Ten similar vessels, the necks of which are embellished with broad rather low ribs or bands, are shown in plates 63, b, 68. and 70, c and d. Somewhat similar vessels are figured from northeastern Arizona by Dr. Fewkes^ and from St. George. Utah, by Holmes.- One has a handle (pi. 70, r) and in another instance there are three tiny protuberances symmetrically placed just below the rim (pi. 68, c). Some of these vessels contained charred com, which, together with the fire stains almost invariably apparent, and the adhering 2 Pottery of tho Ancient riicMos, fig. 242. ^Bull. SO, Bur. Amer. Etiin., pi. 18, 6. BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 63 LARGE OLLAS BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 64 FOOD BOWLS BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 65 BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 66 UNDECORATED VASES BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 67 BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 68 e I COOKING VESSELS WITH BANDED NECKS BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 69 c f a. BOWL MENDED WITH YUCCA TIES, h, c, d. e,f. POTTERY BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 70 POTTERY AND SKULLS BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 71 CS MOEEIS] BUINS ON THE MESAS 197 soot, indicate that these vessels perfoi-med a function analogous to that of the coil-ware oUas of the cliffs—that of culinarj' utensils. Vndecorated Water Jars.—There were many large water jars (pi. 67) in Euin No. 17, and fragments of these are abundant in all the refuse heaps. In shape these resemble similar vessels from the cliffs (pis. 41, a, and -13), but they bear no sign of decoration and but few of the bases are concave and none are flat. Two of the jars (pi. 67, () have banded necks, while those of the others are plain. There is an imusual variation in the shape of the handles. On one vessel they are merely solid stubs of clay protruding fi-om the sides of the base (pi. 67, /). Four of them (pi. 67, d^ e) have large open handles like the ones from the cliffs, through which three and sometimes four fingers can be inserted. One (pi. 67, c) has similar handles except that they consist of two ropes of clay in- stead of a single band. The handles on the vessels-shown in plate 67, a, b, are broad, flat protuberances which turn down like the stubby tail of a bird. This type is exceed- ingly common in the refuse heaps. A water bottle closely resembling the above (pi. 70, h) was found in Kuin Xo. 9. The top is very mucli flattened, and the base is concave. Gourd-shaped Bottles.—The four ^ ^ gourd-shaped bottles (pi. 72, a, d, e, fig. 5.—Outunes of goum-shkped and /) are unique among the pottery vessels. of the upper San Juan drainage. Unfortunately, not one of them retains its neck entire, hence the outline of that part of the vessels is problematical. Jackson^ figures a bottle from the pueblo of Zuiii, a drawing of the outline of which appears in figure 5, a. One side of a like orifice appears in the tops of three of the bottles here shown, and it may be that the resemblance to a gourd was heightened by their completion in a similar manner. However, I am inclined to think that the curve particularly apparent in the one shown in plate 72, e, was continued, and that the neck curved back and ended in a point which was contiguous, but not attached, to the incurving slope of the vessel, as is shown in figure 5, i. There are rough spots on the sides of plate 72, d and /', in the proper positions, as if the presence of the end of the handle had not permitted the polishing of the surface at these points. A detached handle of this sort was found in Euin No. 17, and Holmes figures one constructed in this manner.^ 1 Tenth Ann. Rept. of the Hayden Survey, pi. LXVi. ^ Pottery of the Ancient Pueblos, fig. 354. 198 ANTIQUITIES OF SOITTHWESTEEN COLOKADO [ETH. ANN. 33 Ladles.—Two ladles appear in plate 69, e and /. Thc_v are divided into compartments by a partitidn across the middle. While ladles "with long hollow handles are known from the clitf-dwellings and from the Animas Valley, I have not observed other specimens in which the part corresponding to the handle is divided from the bowl by a ridge of clay. Spoon o-r Paddle.—A spoonliko implement, the handle of which is missing, was found in Kuin No. 9 (pi. 69, d). Bird-fonn Vase.—A^Tiile fragments indicate that bird-form vases wei'e fairly nmnerous, there is but one such specimen in the collec- tion (i^l. 71, a). The tail is upturned, and there are two pairs of protuberances, one of which probably represents wings. The bird topogi-aphy is not accurately delineated, but it appears that the vase is rightly named. Bing-bottoTned Vase.—This term, inelegant though it be, I have applied to the vessel appearing in plate 71. h. The base is a hollow ring with a roughly circular neck rising from one side. A cylin- drical handle connects the top of the neck and the opposite side of the base. Lamp (?).—The writer suggests that the vessel shown in plate 71, e, is a lamp, at the sa:me time realizing fully that this suggestion will meet with considerable criticism. The vessel consists of a rough globular body with a small hole in the top, and a hollow cylindrical spout, which rises from the point of greatest diameter at a slight angle to the plane of the base. The end of the spout, which is much blackened, has been affected by some agency to the extent that part of it crumbled to dust when taken from the bowl in which it was found. Long-continued action of flame might thus destroy the cohesive qualities of the clay. Mountain-sheep Epgy.—There was found on the surface of Ruin No. 17 a small animal effigy, which from the shape and proiJortions of the horns is thought to represei't a mountain sheep (pi. 69, c). It is probably a prayer emblem, similar to those used by the Hopi. Cloud Blower.—The collection contains one small pottery cloud blower or pipe (pi. 69, h). Throuch the center there is a hole the diameter of which lessens progressi\ely toward the smaller end of the cylinder. The surface is decorated with diagonal lines of pits. Two rather nonde-script bottles appear in plate 72, h and c COLOK Vessels of gi'ay, varying all the way from black to white, occur in the collection. In addition there are shades of yellow, orange, red, and brown. The vase shown in plate 66, a, is a glossy black through- out, which appears not to be the result of use as a cooking pot. As MOKEISJ RUINS ON THE MESAS 199 opposed to this, one bowl (pi. 64, e) has an interior as white as well-cleaned kaolin. As has been stated above, this does not seem to be dependent on a surface slip. One globular bowl (pi. 70, a) is a deep, beautiful red. The paste is rather coarse and red all the way through, but the jiolishing has accentuated the redness of the surface. The bowl figured in plate 64, c, is of orange color in which the yellow is predominant. This bowl once bore decorations which are now untraceable. There must have been much iron in the clay from which a con- siderable proportion of the pottery was made. Heat of different degrees has produced i^eculiar blotchings on several of the vessels. The neck of one, otherwise gray, is almost lemon yellow {p\. 63, b), and a red-orange cheek appears on the side of the vessel figured in plate 67, e. The red and orange sherds are of particular interest. They repre- sent most of the shapes present in the collection, besides many others the entire contours of which it is at present impossible to restore. The colored ware of the region will furnish beautiful and instructive material to future excavators. Decoration From the standpoint of decoration the first fact to impress one is that of 62 vessels only 26 bear painted designs. In a typical col- lection from the cliffs, or from the lower La Plata, the proportions would be more than reversed. In fact, in the author's collection of more than 200 spe<;imens collected near the mouth of the La Plata there are only 5, excepting the coil ware, without decoration. In color the decorations merge from, black through brown to red. The brown and red can not be considered to have resulted from a chemical alteration in a dye originally black. These colors are uniform over the entire surface of the dish; the black is permanent, while the red can be removed with a damp cloth. Several of the bowls found near the mouth of Long Hollow (Ruins Nos. 22 and 23) had their entire exteriors painted with a light-red substance, which comes off very readily in the presence of moisture. The decoration is in most cases crude, although fair skill is shown in two instances (pis. 64, a, and 72, e). With one exception the symbolism differs from that previously observed on Mesa Verde pottery. This exceptional symbolism ap- pears on a water bottle (pi. 70, h). It consists of a liollow square with arms extending from the corners. The essential features of 200 ANTIQUITIES OF SOUTHWESTERN COLORADO [eth. ann. 33 the design appear on a bowl from Cliff Palace.^ It is a much-used svniliol on the ]iotterv from the Animas Valley, there being several examples in the au- thor's collection. The star appears twice, once with seven points (pi. C4, a) and once with only five (pi. 64. e). The bowl figured in plate 65, 5, has what appears to be a spotted serpent coiled spirally from the bottom to the rim. Another serpent pattern oc- curs on the bowl given in a of the same plate. — The rims of theFiu. C. Metatu and niano. bowls, when deco- rated at all, bear a solid black line inste ad of the row of dots or zigzag lines found on the pottery from the cliffs. The designs shown in fig- ures 7-11 are from bowls not in the collection. They were found by Mr. Ealph Linton, of Swarthmore. Pa., on the Powell ranch. Until a more complete series of designs shall have been gathered, the writer does not think best to at- tempt an analysis or classi- fication of them. STONE IMPLEMENTS Grinding Stone s.— Fig. 7.—Design on bowl from mouth of Long Metates and manos are most Hollow. conspicuous among the stone implements from the mesa sites. A pair of these is shown in figure 0. The metate is a slab of moderately ' Fewkes, Bull. 51, Bur. Amer. Ethn., pi. 24. BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 73 ^^^^S BUREAU OF AMERICAN ETHNOLOGY THIRTY-THIRD ANNUAL REPORT PLATE 74 CHIPPED IMPLEMENTS BUREAU OF AMERICAN ETHNOLOGY TH I RTY-TH I RD ANNUAL REPORT PLATE 75 CHIPPED IMPLEMENTS AND POLISHED ORNAMENTS MOBBIS] RUINS ON THE MESAS 201 Fig. S. — Design on bowl from mouth of Long Hol- low. 202 ANTIQUITIES OF SOUTHWESTERN COLORADO [eth. ann. S3 classed as arrowpoints, knives, drills, and scrapers. The materials are mainly jaspei', flint, chalcedony, quartzite, and obsidian. The dull gloss or luster, which is particularly apparent on the obsidian imple- ments, shows them to be very old. In length these vary from ^ to 2J inches. The workmanship of many is poor, but some are of unusual beauty. A few ornaments are also figured. Two of these show diag- onal incised lines. It is impossible to say how many of these objects were lost by the inhabit- ants of the cliffs and how Fig. 11.—Design on bowl from mouth of Long jn a n V belonged tO the Hollow. d1wel1l1ers on Jt.hZe mesas. The most notable feature in connection with the stone implements from the mesa sites is the lack in variety and quantity of everything but chipped instruments. BONE IMPLEMENTS Bone implements are not numerous. Besides three bone awls no other artifacts of bone were exhumed in the mesa ruins. As miglit be expected, all articles of wood and other perishable material not consumed by the burning of the buildings had entirely decomposed. Summary' The inhabitants of the mesas were an agricultural people whose domiciles were one-storied aggregations of cell-like chambers, usu- ally grouped to form a rectangle. Generally speaking, the rooms extended down into the earth, and with few exceptions the sections of the walls above gi'ound were constructed of upright poles cov- ered with plaster. Thus far no kiva has been found in, or connected with, a jacaJ dwelling. It is possible, however, that when excavated some of the numerous circular pits will prove to have features linking them unmistakably to the kivas of later time. MOERIS] RUINS ON THE MESAS 203 The pottery from the mesas exhibits a wide range of form and surface treatment, but structurally it is inferior. Less than 42 per cent of it is decorated with painted designs, and true coil ware does not appear. The symbolism is unlike that on the pottery from the neighboring cliffs and shows less conventionalization. Stone implements are few, but such as have been found are of the same general types and exhibit as good workmanship as those of the cliff people. Of the work in bone, wood, and other perishable materials almost nothing is known. ^ CONCLUSIONS It is evident tliat there existed on the mesas between Mancos and La Plata Eivers a culture differing in many respects from that of tlie cliff-dwellers of the Mesa Verde. The meager information at hand seems to connect it with the pre-Puehlo pit-house culture now generally conceded to have existed in the Southwest. The true Pueblo culture has as its diagnostic character compact community villages several stories in height, usually in terrace form. The absence of the terraced form of architecture serves equally well to characterize the pre-Pueblo culture. The pottery especially indicates the greater antiquity of the dwell- ing sites on the mesas. The wide range of form, as well as the un- skilled workmanship displayed, shows the ceramic art still to have been quite plastic, and not bound by the rigid convention which is apparent in the pottery from the cliff'-dwellings. The same may be said of the symbolism. Though crude, the designs are more boldly executed than are the conventionalized decorations on the cliff pottery. The skulls offer another point of difference. Of 33 ci'ania 11, or 33^ per cent, do not show the occipital flattening general among the crania from the cliff-dwellings. Too much importance nnist not be attached to this, however, until further research shall prove whether the variation continues in evidence, or whether the skulls offering a basis for the statements here made comprise only a small aberrant group. The limits of the type of remains here described can be determined only by extended excavations. Whether they are typical of a small area, or whether they continue and connect with other localities in which the pit-house culture is already known, remains to be seen. The discovery in the northern part of their domain of a more ancient culture than that of the cliff-dwellers should be of special interest, since it appears that the region north of the San Juan Eiver is the center from which migration carried the true Pueblo culture to the south, southeast, and perhaps to the west. The direct relationship between the people of the cliffs and those of the mesas can not be established at present. The inhabitants of 204 MOEEis] CONCLUSIONS 205 the caverns may have been directly descended from the builders of the jacal houses, or there may have been a hiatus between the two periods of occupancy. Necessarily these conclusions are but tentative, and are offered as nothing more. Now that a beginning has been made, it is to be hoped that some one will see fit to undertake a work of sufficient amplitude to lead as nearly as possible to the solution of the problems which the brief research here recorded has done little more than to suggest. BIBLIOGRAPHY Fewkes. J. Walter. Antiquities of the Mesa Verile National Park: Spruce- tree House. Bull. 41, Bur. Amer. Ethn., Washington, 1909. Preliminary report on a visit to the Navaho National Monument, Ari- zona. Bull. .50, Bur. Amer. Ethn., Washington, 1911. Antiquities of the Mesa Verde National Park: Clitf Palace. Bull. 51, Bur. Amer. Ethn., Washington, 1911. Holmes, W. H. Pottery of the ancient Pueblos. Fourth Kept. Bur. Amer. Ethn., pp. 2.")7-3G0, Washington, 1886. Report on the ancient ruins of southwestern (Colorado. lOfh Ann. Rcpt. V. S. Oeol. and Geog. Surv. Terr. [Hayden Survey'] for 1876, pp. 383-408, Washington, 1878. Hough, Walter. Culture of the ancient Pueblos of tlie Upper (!ila River region. Bull. 87, V. S. Nat. 3Ius., AVashington, 1914. Jackson. Wm. H. Ancient ruins in southwestern Colorado. lEiyhth] Ann. Rcpt. V. S. Gcol. and Geog. Surv. Terr. [Haydeii Survey] for 1S7-',, pp. 3G9-381, Washington. 1876. Report on the ancient ruins examined in 187.^ and 1877. 10th Ann. Kept. V. S. Geol. and Geog. Surv. Terr. [Hayden Survey] for 1876, pp. 411- 450, Washington, 1878. Nordenskiold, G. The cliff dwellers of the Mesa Verde. Stockholm, 1803. Pbudden, T. Mitchell. Prehistoric ruins of the San Juan watershed in Utah, Arizona, Colorado, and New Mexico. Atn. Anthrop., n. s. vol. v, pp. 224-288, Lancaster, Pa., 1903. Stevenson, Matilda Coxk. The Zuiii Indians. Twenty-third Ann. Kept. Bur. Amer. Ethn., Washington, 1904. 206