CONTENTS. Page. Introduction Chapter I.—Traditionary history of Tusayan 16 Explimatory Summavy of traditions List of traditionary gentes Snpplemeut.ary legend Chapter II.—Ruins and inh.abited villages of Tusayan 42 Physical features of the province Methods of survey Plans and description of ruins '*'' Walpi ruins ^^ Old Mashongnavi Shitaimuvi ^^ Awatubi Horn House ^ Small ruin near Horn House '-^^ Bat House "^ Mishiptonga "^ Moen-kopi Ruins on the Oraibi -nash •^'^ Kwaituki "*' Tebugkihu, or Fire House ^'^ Chukubi f^ Payupki f Plans and descriptions of inhabited villages oi Hano ^^ Sicliumovi Walpi "^ Mashongnavi Shupaulovi Shumopa 71VI '^'' Oraibi Moen-kopi 77 Chapter III.—Ruins and inhabited villages of Cibola «0 Physical features of the province *'" Plans and i 56 XXIV. Dance rock and kiva, Wiiljii 58 XXV. Foot trail to Walpi 60 XXVI. Mashongnavi, plan 62 XXVII. Mashongnavi with Shnpanlovi in distance. 64 XXVIII. Back wall of a Mashongnavi hon.se-row 66 XXIX. West side of a principal row in Mashongnavi 68 XXX. Plan of Shnpanlovi 70 XXXI. View of Shnpanlovi 72 XXXII. A covereil passageway of Shnpanlovi 74 XXXIII. The chief kiva of Shnpanlovi 76 XXXIV. Plan of Shnmopavi 78 XXXV. View of Slinmop.ivi 80 XXXVI. Oraibi, plan In pocket. XXXVII. Key to tlie Oraibi plan, also showing localization of gentes 82 XXXVIII. Aconrt of Oraibi 84 XXXIX. Masonry terraces of Oraibi 86 XL. Oraibi honse row, sliowing court side 88 XLI. Back of Oraibi honse row 90 XLII. Tlie site of Moen-kopi 92 XLIII. Plan of Moen-kopi 94 XLIV. Moeu-kojii 96 7 . 8 PUEBLO ARCHITECTURE. Page. Plate XLV. The Mormon mill at Moen-kopi '. 98 XLVI. Hawikuli, plan 100 XLVII. Hawikuh, view 102 XLVIII. Adobe church at Hawikuh 104 XLIX. Kctchipauiin, plan 106 L. Kctchipaiian 108 LI. Stone church at Ketchipauan 110 LII. K'iakima, plan 112 LIII. Site of K'iakima, at base of Taaaiyalana *. 114 LI V. Recent wall at K'iakima 116 LV. Matsaki, plan 118 LVI. Standing wall at Pinawa 120 LVII. Haloua excavations as seen from Zuni 122 LVIII. Fragments of Halona wall 124 LIX. The mesa of Tilaaiyalana, from Zutii 126 LX. Taaaiyalana, plan 128 LXI. Standing walls of Taaaiyalana ruins 130 LXII. Remains of a reservoir on Taaaiyalana 132 LXIII. Kin-tiel, plan (also showing excavations) 134 LXIV. North wall of Kin-tiel 136 LXV. Standing walls of Kin-tiel 138 LXVI. Kinua-Zinde 140 LXVII. Nutria, plan 142 LXVIII. Nutria, view 144 LXIX. Pescado, plan 146 LXX. Court view of Pescado, showing corrals 148 LXXI. Pescado houses 150 LXXII. Fragments of ancient masonry in Pescado 152 LXXIII. Ojo Caliente, plan In pocket. LXXIV. General view of Ojo Caliente 154 LXXV. House at Ojo Caliente 156 LXXVI. Zuni, plan In pocket. LXXVII. Outline plan of Zuni, showing distribution of oblique openings. 158 LXXVIII. General inside view of Zuni, looking west 160 LXXIX. Zuni terraces 162 LXXX. Old adobe church of Zutii 164 LXXXI. Eastern rows of Zuni 166 LXXXII. A Zuni court 168 LXXXIII. A Zuni small house 170 LXXXIV. A hou.se-building at Oraibi 172 LXXXV. A Tusayau interior 174 LXXXVI. A Zuni interior 1T6 LXXXVII. A ki va hatchway of Tnsayan .• 178 LXXXVIII. North kivas of Shumopavi, from the northeast 180 LXXXIX. Masonry in the north wing of Kin-tiel 182 XC. Adobe garden walls near Zuni 184 XCI. A group of stone corrals near Oraibi 186 XCII. An inclosing wall of upright stones at Ojo Caliente 188 XCIII. Upright lilocks of sandstone built into an ancient pueblo wall. . 190 XCIV. Ancient wall of upright rocks in southwestern Colorado 192 XCV. Ancient Hoor-l)eams at Kin-tiel 194 XCVI. Adobe walls in Zuni 196 XCVII. Wall coping and oven at Zuni 198 XCVIII. Cross-pieces on Zuni ladders 200 XCIX. Outside steps at Pescado 202 1 ILLUSTRATIONS. 9 Piige. Plate C. An excavated room at Kin-tiel 204 CI. Masoury cUimueys of ZuiSi 206 CII. Eeiuaius of a g.ateway in Awatubi 208 cm. Ancient gateway, Kin-tiel 210 CIV. A covered passageway in Masliongnavi 212 CV. Small sqnare openings in Pueblo Bonito 21 CVI. Sealed openings in a detached house of Nutria 216 CVII. Partial filliug-iu of a large opening in Oraibi, converting it into a door way 218 CVIII. Large openings reduced to small windows, Oraibi 220 CIX. Stone corrals and kiva of Mashongnavi 222 ex. Portion of a corral in Pescado 224 CXI. Zuui eagle-cage 226 Fig. 1. View of the First Mesa 43 2. Ruins, Old Walpi mound 47 3. Ruin between Bat House and Horn House 51 4. Ruin near Moen-kopi, plan 53 5. Euiu 7 miles north of Oraibi 55 6. Ruin 14 miles north of Oraibi (Kwaituki) 56 7. Oval fire-house ruin, plan. (Tebngkihu) 58 8. Topography of the site of Walpi 64 9. Mashongnavi and Shupaulovi from Shumopavi 06 10. Diagram showing growth of Mashongnavi 67 11. Diagram showing growth of Mashongnavi 68 12. Diagram showing growth of Mashongnavi 69 13. Topography of the site of Shujiaulovi 71 14. Court kiva of Shumopavi 75 15. Hampassawan, plan 84 16. Pinawa, plan 87 17. Nutria, plan, small diagram, old wall 94 18. Pescado, plan, old wall diagram 95 19. A Tusayau wood-rack 103 20. Interior ground plan of a Tusayan room 108 21. North kivas of Shumopavi from the southwest 114 22. Ground plan of the chief-kiva of Shupaulovi 122 23. Ceiling-plan of the chief-kiva of Shupaulovi 123 24. Interior view of a Tusayan kiva 124 25. Ground-plan of a Shupaulovi kiva 125 26. Ceiling-plan of a Shupaulovi kiva 125 27. Ground-plan of the chief-kiva of Mashongnavi 126 28. Interior view of a kiva hatchway in Tusayan 127 29. Mat used in closing the entrance of Tusayan kivas 128 30. Rectangular sipapuh in a Mashongnavi kiva 131 31. Loom-post in kiva floor at Tusayan 132 32. A Zuni chimney showing pottery fragments embedded in its adobe base 139 33. A Zuni oven with pottery scales embedded in its surface 139 34. Stone wedges of Zuni masonry exjiosed in a rain-washed wall 141 35. An uuplastered house wall in Ojo Caliente 142 36. Wall decorations in Mashongnavi, executed in pink on a white ground 146 37. Diagram of Zuni roof construction 149 38. Showing abutment of smaller roof-beams over round girders 151 39. Single stone roof-drains ' 153 40. Trough roof-drains of stone 153 10 PUEBLO ARCHITECTURE. Page. Fig. 41. Wooden roof-drains 154 42. Curved root-drains of stone in Tusayan 154 43. Tusayan roof-drains ; a discarded metate and a gourd 155 44. Zuni roof-drain, with sjilash-stones on roof below 156 45. A modern notehed ladder in Oraibi 157 46. Tusayan notclied ladders from Mashongnavi 157 47. Aboriginal American forms of ladder 158 48. Stone stejis at Oraibi with jdatform at corner 161 49. Stone steps, with platform at chimney, in Oraibi 161 50. Stone steps in Shumopavi 162 51. A series of cooking pits in Mashongnavi 163 52. Pi-gnmmi ovens of Mashongnavi 163 53. Cross sections of pi-gummi ovens of Mashongnavi 163 54. Diagrams .showing foundation stones of a Znni oven 164 55. Dome-shajjed oven on a plinth of masonry 165 56. Oven in Pescado exposing stones of masonry 166 57. Oven in Pescsido exposing stones of masonry 166 58. Shrines in Mashongnavi 167 59. A poultry liou.'ie in Sichumovi resembling an oven 167 60. Ground-plan of an excavated room in Kin-tiel 168 61. A corner chimney-hood with two supporting poles, Tusayan 170 62. A curved chimney-hood of Mashongnavi 170 63. A JIashonguavi chiuuiey-hood and walled-up fireplace 171 64. A chimney-hood of Shupaulovi 172 65. A semi-detached square chimney-hood of Zuni 172 66. Uuplastered Zuni chimney-hoods, illustrating construction 173 67. A fireplace and mantel in Sichumovi 174 68. A second-story tirei)lace in Mashongnavi 174 69. Piki stone and chiuuiey-hood in Sichumovi 175 70. Piki stone and primitive andiron in Shumopavi 176 71. A terrace firej)lacc ami chimney of Shumopavi 177 72. A terrace cooking-])it and chimney of Walpi 177 73. A ground cooking-pit of Shumopavi covered with a chimney 178 74. Tnsiiyan chimneys 179 75. A barred Zuni door 183 76. Wooden pivot hinges of a Zuni door 184 77. Paneled wooilen doors in Hano 185 76. Framing of a Zuni door panel 186 79. Rude transoms over Tusayan openings 188 80. A large Tusayan doorway, with small transom openings 189 81. A doorway and double transom in Walpi 189 82. An ancient doorway in a Canyon de Chelly clifl:' ruin 190 83. A symmetrical notched doorway in Mashongnavi 190 84. A Tusayan notched doorway 191 85. A large Tusayan doorway with one notehed jamb 192 86. An ancient circular doorway, or " stone-close," in Kin-tiel 193 87. Diagram illustrating symmetrical arrangement of small openings in Pueblo Bonito 195 88. Incised decoration on a rude wiudow-sash in Zuni 196 89. Sloping selcnite window at base of Zuiii wall on upper terrace 197 90. A Zuni window glazed with selcnite 197 91. Small openings in the back wall of a Zuni house cluster 198 92. Sealed openings in Tusayan 199 93. A Zuni doorway converted into a window 201 94. Zuni roof-openings 202 1 ILLUSTRATIONS. 1 Page. Fig. 95. A Zufii roof-openinR witli r.iisetl coping 203 96. Zufii roof-oiieiiings with oue raised end 203 97. A Ziuii root-hole with cover 204 98. Kiva trap-door in Zuui 20,5 99. Halved and pinned trap-door frame of a Zuiii kiva 206 100. Typical sections of Zufii oblique openings 208 101. Arrangement of mealing stones in a Tusayan house 209 102. A Tusayan grain bin 210 103. A Zniii pinnie-bcix 210 104. A Zuni phmic-box 210 105. A Tusayan mealing trough 211 106. An auc icnt pueblo form of metate 211 107. Zuni stools 213 108. A Zuni chair 213 109. Construction of a Zuiii corral 215 110. Gardens of Zuni 216 111. " Kishoni," or uncovered shade, of Tusayan 218 1 12. A Tusayan field shelter, from Southwest 219 113. A Tusayan field shelter, from northeast 219 114. Diagram showing ideal section of terraces, with Twsayau names 223 BUREAU OF ETHNOLOGY % ru^^AYArvI Puce ' ^^ X^^----o^«i?o (h-ofui General Map OF THE Pueblo Region OF Arizona and New Mexico showing relative position of the provinces OF TUSAYAN AND CIBOLA. BY Victor Mindeleff Shaded arai represents the J^avinces of TuscLvart and. Ctboltt-. EIGHTH ANNUAL REPOIT PL. I BUREAU OF ETHNOLOOy 108 107° # ^^ oa.\ -V"i>!f f^' •... V "" ...-. •'>.• " "•'•'.JU.;jl >. General Map 33- OF THE -»S Pueblo Region OF JYatls. ------- V Iload.s, '"^^.^^^ Arizona and New Mexico, Mbiicrii Pueblo.i, -- • SHOWING relative: position of the provinces Ancient " - -^ OF RiUrvs - Tmvers. CUJf-honsrs rfc.a TUSAYAN AND CIBOLA. Burial Places - - - a BY Victor Minde.leff Scale Shaded area represents the Provinces of Dniwii I'll O \VCri>niri Tusavart anci CiboUv. i07' WO" A STUDY OF PUEBLO ARCHITECTURE IN TUSAYAN AND CIBOLA. By Victor Mindeleff. INTRODUCTION, The remains of pueblo architecture are found scattered over thousands of square miles of tlie arid region of the southwestern plateaus. This vast area includes the drainage of the liio Pecos on the east and that of the Colorado on the west, and extends from central Utah on the north beyond the limits of the United States southward, in which direction its boundaries are still undefined. The descendants of those who at various times built these stone vil- lages are few iu number and inhabit about thirty pueblos distributed irregularly over parts of the region formerly occupied. Of these the greater number are scattered along the upper course of the Kio Grande and its tributaries in New Mexico; a few of them, comprised mthin the ancient provinces of Cibola and Tusayan, are located within the drainage of the Little Colorado. From the time of the earliest Spanisli expedi- tions into the country to the present day, a period covering more than three centuries, the former province has been often visited by whites, but the remoteness of Tusayan and the arid and forbidding character of its surroundings have caused its more complete isolation. The archi- tecture of this district exhibits a close adherence to aboriginal practices, still bears the marked impress of its development under the; exacting conditions of an arid environment, and is but slowly yielding to the in- fluence of foreign ideas. The present study of the architecture of Tusayan and Cibola embraces all of the inhabited pueblos of those provinces, and includes a number of the ruins traditionally connected with them. It will be observed by reference to the map that the area embraced in these provinces comprises but a small itortion of the vast region over which pueblo culture ouce extended. This study is designed to l)e followed by a similar study of two typical groups of ruins, viz, that of Canyon de Chelly, in northeastern Arizona, and that of the Chaco Canyon, of New Mexico; but it has been necessary for the -wi-iter to make occasional reference to these ruins iu the present 13 14 PUEBLO ARCHITECTURE. paper, both in tlu- disfussioii of gtMieral arraugemeiit and cliaiaftoristic ground plaus, embodied in Chapters ii and iii and in tlit' comparison bycou.structional details treated in Chapter IV, in order to define elearly the rchitions of tlie various features of pueblo architecture. They belong to the same pueblo system illustrated by the villages of Tusayan and Cibola, and with the Canyon de Chelly group there is even some trace of traditional connection, as is set forth by Mr. Stephen in Chapter i. The more detailed studies of these ruins, to be publislied later, together with the material emljodied in the present paper, will, it is thought, furnish a record of the principal characteristics of an important type of primitive architecture, whicli, under the influence of the aiid environment of the southwestern plateaus, has developed from the rude lodge into the many- storied house of rectangular I'ooms. Indications of some of the steps of this development are traceable even in the architecture of the present day. The pueblo of Zuni was surveyed by the writer in the autuum of 1881 with a view to procuring tlie necessary datai for the construction of a large-scale model of this pueblo. For this reason the work afforded a record of external features only. The modern pueblos of Tusayan were similarly surveyed in the fol- lowing season (18S2-'8o), the plans being sujiiilemented by photographs, from which many of the illustrations accompanying this paper have been drawn. The ruin of Awatubi was also included in the work of this season. In the autumn of 1885 many of the ruined pueblos of Tusayan were surveyed and examined. It was during this season's work that the details of the kiva construction, embodied in the last cliapter of this paper, were studied, together with interioi- details ot the dwellings. It was in the latter part of this season that the farming pueblos of Cibola were surveyed and photographed. The Tusayan farming ])uel)lo of Moen-kopi and a number of the ruins in tlll^ province were surveyed and stmlied in the early part of the season of 1887-88, the latter portion of which season Avas principally devoted to an examination of the Chaco ruins in New Mexico. In the prosecution of the Held work above outlined the author has been greatly indebted to the efficient assistance and hearty cooperation of j\Ir. Cosmos Mindeleff, by whom nearly all the pueblos illustrated, with the exception of Zuhi, have been surveyed and ])latte(l. The plans obtained have involved much careful work with surveying instruments, and ha\e all been so platted as faithfully to record the nunute variations from geometric forms which are so characteristic of the ]>ueblo work, but which have usually been ignored in the hastily prepared sketch plans that have at times appeared. In conseiiuence of the necessary omission ofjust such iuformatiou in hastily drawn i)lans, erroneous im])ressions have becTi given regaidiiig the degree of skill to which the pueblo peox>les had attained in the planning and building of BUREAU OF ETHNOLOGV EIGHTH ANNUAL REPORT PL. II OLD MASHONGNAVI. MINUELEFF.l INTRODUCTION. 15 their villages. In the general distribution of the honses, and m the ahgnment and arrangement of their walls, as indicated in tbe plans shown in Ohapters li and iii, an absenc^e of high architectural att;iin- ment is found, which is entirely in keeping with the lack of skill ap- parent in many of the constructional devices shown in Chapter iv. In ]ir(^paring this pa])er for i)ublication Mr. Cosmos Mindeleff has rendered much assistance in the revision of manuscript, and in the pre- paration of some of the final drawings of ground plans ; on Mm has also fallen tlu> conii)ilation and arrangement of Mr. A. M. Stei>hen's tradi- tionary material from Tusayau, embraced in the first chapter of the paper. This latter material is of sjiecial interest in a study of the i>ueblos as indicating some of the conditions under which this architectural type was developed, and it appropriately introduces the more purely archi- tectural study l)y the auth(n'. Such traditions must be used as history with the utmost caution, and only for events that are very recent. Time relations are often hope- lessly confused and the narratives are greatly incumbered with mytbo- logic details. But while so barren in definite information, these tradi- tions are of the greatest value, often tlirough their merely incidental allusions, in presenting to our minds a picture of the conditions under wliich the repeated migrations of the pueblo builders took place. The development of architecture among the Pueblo Indians was com- paratively rapid and is largely attributable to frequent changes, migra- tions, and movements of the people as described in Mr. Stephen's account. These changes were due to a variety of causes, such as dis- ease, death, the frequent warfare carried on between different tribes and branches of the builders, and the hostility of outside tribes; but a most potent factor was certainly the inhospitable character of their environ- ment. The disappearance of some venerated spring during an unusually dry season would be taken as a sign of the disfavor of the gods, and, in spite of the massive character of the l)uildings, would lead to the migra- tion of the people to a more favorable spot. The traditions of tlu; Zuiiis, as well as those of the Tusayau, ft-e(iueutly refer to such nugrations. At times triljes split up and separate, and again phratries or distant groups meet and l)and together. It is remarkable that the substantial character of the architecture should i>ersist through such long series of compulsory removals, but while the builders were held togi^ther by the necessity for defense against their wilder neighbors or against each other, this strong defensive niotivti would |)er])etuate the lal)orious type of con- struction. Such conditions would contribute to the rapid development of the buikliug art. CHAPTER T. TRADITIONAL HISTORY OF TUSAYAN. EXPLANATORY. Ill tliis chapter ^ is presented a summary of tlio traditions of the Tusiiyan, a mumiIht of which were collected from old iiumi, from Wal|)i on the east to Moen-kopi on the west. A tradition varies much with tlie tribe and the individual; an authoritative statement of the current tradition on any point could be made only with a complete knowledge of all traditions extant. Such knowledge is not possessed by any one man, and the material included in this chapter is presented simply as a summary of the traditions secured. The material was collected by Mr. A. M. Stephen, of Ream's Canyon, Arizona, who has enjoyed unusual facilities for the work, having lived for a number of years pant in Tusayan and possessed the confidence of the principal ])riests—a very necessary condition in work of this char- acter. Though far from (u)mplete, this summary is a more comprehen- sive presentation of the traditionary history of these people than has heretofore been published. SUMMARY OF TRADITIONS. The creation myths of the Tnsayau differ widely, but none of them designate the region now occupied as the place of their genesis. These people are socially divided into family groups called wi'ngwu, the de- scendants of sisters, and groups of wi'ngwu tracing descent from the same female ancestor, and having a common totem called my'unui. Each of these totemic groups preserves a creation myth, carrying in its details special reference to themselves; but all of them claim a common origin in the interior of the earth, altliough the place of emergence to the surface is set in widely separated localities. They all agree in main- taining this to be the fourth plane on which mankind has existed. In the beginning all men lived together in the lowest depths, in a region of darkness and moisture; their bodies were misshaped and liorrible, and they suffered great misery, moaning and bewailing continually. Through the intervention of Myuiiigwa (a vague conception known as the god of the interior) and of Baholikonga (a crested serpent of enor- mous size, the genius of water), the "old men" obtained a seed from which sprang a magic growth of cane. It penetrated through a crevice ' Tills chapter is compiled by Cosmos Miuileleff from material collected by A.M. Steplieu. 16 : MINDELEFF.l CREATION MYTHS. 17 in the roof overhead auil inaiikiiul climbed to a higher phiue. A dim light appeared in this stage and vegetation was produced. Another magic growth of cane aftVn-ded the means of rising to a still higher plane on which the light was brighter; vegetation was reproduced and the animal kingdom was created. The final ascent to this present, or fourth plane, was effected by similar magic growths and was led by mythic twins, according to some of the myths, by climbing a great pine tree, in others by climbing the caue, PhragmUes communis, the alternate leaves of which afforded steps as of a ladder, and in still others it is .said to have been a rush, through the interior of which the people passed up to the surface. The twins sang as they pulled the people out, and when their song was ended no more were allowed to come; and hence, many more were left below than were permitted to come above; but the outlet through which manlcind came has never been closed, and Myu'ingwa sends through it the germs of all living things. It is still symbolized by the peculiar coustruction of the hatchway of the kiva and iu the designs on the sand altars in these underground chambers, by the unconnected circle painted on pottery and by de\ices on basketry and other textile fabrics. All the people that were i)ermitted to come to the surface were col- lected and the different families of men were arranged together. This was done under the direction of twins, who are called Pekonghoya, the younger one being distinguished by the term Balingahoya, the Echo. They were assisted by their grandmother, Kohkyang wiiliti, the Spider woman, and these appear in varying guises in many of the myths and legends. They instructed the peojile in divers modes of life to dwell on mountain or on plain, to build lodges, or huts, or windbreaks. They distributed appropriate gifts among tlienr and assigned each a pathway, and so the various families of mankind were dispersed over the earth's surface. The Hopituh,' after being taught to build stone houses, were also divided, and the different divisions took separate paths. The legends inilicate a long period of extensive migrations in separate communities; the gToups came to Tusayan at different times and from different di- rections, but the people of all the villages concur in designating the Snake people as the tirst occupants of the region. The eldest member of that nyunui tells a curious legend of their migration fi'om which the following is quoted At the general dispersal my people lived in snake skins, each family occupying a separate snake skiu bag, and all were liung on the end of a rainbow, which swung around until the enlaius were seen, over which roauied the deer, the antelope, and the BUREAU OF ETHNOLOGV EIGHTH ANNUAL REPORT PL. IV ^. -^i fc-: ^3i»-u.m."'^ T- "... % »*-'*-*^ --tv -*-;. •;:-..;ti';.;*> '" t' -"^, •".' ;'.*''.''•';'•'• *"''\ ^' f^-::^;V;;; .;. -.HiiiAl../;, •'- :.4r"'^,'r'i!-5v-..'> ffe?V*^(y---VvVt$VJ^. m- "?-^- : .^Mr^'.^^-y:m:^^i'r._^mi|imr:: i ..- y Scale. Feet. I I I I I I I ] ] I I I I I I I I I H AWATUBI (TALLA-HOGAN). : MiNDELEFF.] MYTHS AND LEGENDS. 19 bisou. feeding ou never-failiiifj; grasses. Twiuiug through these plains ivere streams of bright water, beautiful to look upon. A place where none but those who were of our people ever gained access. This description siTggests some region of the headwaters of the Rio Grande. Lilve tlie Sualie people, they tell of a protracted migration, not of continnous travel, for they remained for many seasons in one jdace, where they would plant and bnild permanent houses. One of these halting places is described as a canyon with high, steep walls, in which was a flowing stream ; this, it is said, was the Ts^gi (the Navajo name for Canyon de Chelly). Here they built a large honse in a cavern- ons recess, high up in the canyon wall. They tell of devoting two years i to ladder making and cutting and pecking shallow holes up the steep rocky side by which to monnt to the cavern, and three years more were employed in building the honse. While this work was in progress part of the men were planting gardens, and the women and children were gathering stones. But no adequate reason is given for thus toiling to fit this impracticable site for occupation; the footprints of Miisaiiwn, which they were following, led them there. The legend goes on to tell that after they had lived there for a long time a stranger happened to stray in their vicinity, who proved to be a Hopitnh, and said that he lived in the sonth. After some stay he left and was accompanied by a iiarty of the " Horn," who were to visit the land occupied by their kindred Hopituh and return with an accol^nt of them; bnt they never came back. After waiting a long time another band was sent, who returned and said that the first emissaries had found wives and had built houses on the brink of a beautiful canyon, not far from the other Hopituh dwelUngs. After this many of the Horns grew dissatisfied with their cavern home, dissensions arose, they left their home, and finally they reached Tusayan. They lived at first in one of the canyons east of the villages, in the vicinity of Ream's Canyon, and some of the numerous ruins on its brink mark the sites of their early houses. There seems to be no legend distinctly attaching any particular ruin to the Horn people, although there is little doubt that the Snake and the Horn were the two first peoples who came to the neighborhood of the present villages. The Bear people were the next, but they arrived as separate branches, and fi'om opposite directions, although of the same Hopituh stock. It has been impossible to obtain directly the legend of the Bears from the west. The story of the Bears from the east tells of encountering the Fire people, then living about 25 miles east from Walpi ; but these are now extinct, and nearly all that is known of them is told in the Bear legend, the gist of which is as follows The Bears originally lived among the mountains of the east, not far distant from the Horns. Continual quarrels with neighboring villages 'The t«rm yaauna. trau8latointed out as the remains of one of the mission buildings. Other groups followed—the Mole, the Spider, and the " Wiksrun." These latter took their name ft-om a curious ornament worn by the men. A piece of the leg-bone of a l)ear, from which the marrow had been extracted and a stopper fixed in one end, was attached to the fillet bind- ing the hair, and hung down in front of the forehead. This gens and the Mole are now extinct. Shumopavi received no further accession of population, but lost to some extent by a portion of the Bear people moving across to Walpi. No important event seems to have occurred among them for a long period after the destruction of Silcyiltki, in which they bore some part, and only cursory mention is made of the ingress of " enemies from the north ;" but their callage, apparently, was not assailed. The Oraibi traditions tend to <'(mflrm those of Shumopavi, and tell that the first houses there were built by Bears, who came from the latter place. The following is ft-om a curious legend of the early settlement 28 PUEBLO ARCHITECTURE. The Boar people had two chiefs, who were brothers; the ehler was called Ywen-ti-s6-mo, aiul the younger Ma-tci-to. They had a desperate quarrel at Shuniopavi, and tlieir people divided into two factions, accord- ing as they inclined to one or otlior of the contestants. After a long period of contention Ma-tci-to and his followers withdrew to the mesa where Oraibi now stands, about 8 miles northwest from Shniiiopa\i, ind built houses a little to the southwest of the limits of the present town. These houses were afterwards destroyed by "enemies from the north," and the older portion of the existing town, the southwest 'ends of the house rows, were built with stones fi'om the demolished houses. Frag- ments of these early walls are still occasionally unearthed. After Matcf-to and his people were established there, whenever any of the Shuuu)pavi i)eople became dissatisfied with that place they built at Oraibi, Ma-tci-to placed a little stone monument about halfway between these two villages to mark the boundary of the laud. Vweu- ti-so'-mo objected to this, but it wasultimately accepted with the proviso that the vdlage growing the fastest should have the privilege of moving it toward the other village. The monument still stands, and is on the direct Oraibi trail from Shuniopavi, .3 miles ftom the latter. It is a well dressed, rectangular block of sandstone, projecting two feet above the ground, and measure^ Si by 7 inches. On the end is carved the rude semblance of a human head, or mask, the eyes and mouth being merely round shallow holes, with a black line painted around them. The stone is pecked on the side, but the head and front are rubbed quite smooth, and the block, tapering slightly to the base, suggests the ancient Roman Termini. There are Eagle people living at Oraibi, Mashongnavi, and Walpi, and it woirld seem as if they had journeyed for some time with the later Snake people and others ft-om the northwest. Vague traditions attach them to several of the ruins north of the Moeii-kopi, although most ot these are regarded as the remains of Snake dwellings. The legend of the Eagle people introduces them ftom the west, com- ing in by way of the Moen-kopi water course. They found many peo- ple living in Tusayan, at Oraibi, the Middle ]\Iesa, and near the East Mesa, but the Snake village was yet in the valley. Some of the Eagles remained at Oraibi, but the main body moved to a large mound just east of Mashongnavi, on the summit of which they built a village and called it Shi-tai-mu. Numerous traces of small-roomed houses can be seen on this moimd and on some of the lower surroundings. The uneven sumnut is about .300 by 200 feet, and the village seems to have been built in the form of an irregular ellipse, but the gTound xdan is very obscure. While the Eagles were living at Shi-tdi-mu, they sent "Yellow Foot" to the moirntain in the east (at the headwaters of the Eio Grande) to obtain a dog. After many jierilous adventures in caverns guarded by bear, mountain lion, and rattlesnake, he got two dogs and returned. BUREAU Of ETHNOLOGY EIGHTH ANNUAL REPORT PL. IX 3H). ElCMTM ANNUAL REPORT FL. W MiNLELEFK.) TRADITIONS SHUPAULOVI. 29 Tliey were wanted to keep the coyotes out of the corn and the gardens. The (logti grew numerous, and would go to Mashonguavi in search of food, and also to some of the people of that \^lage, which led to serious quarrels between them and the Eagle people. TJltimately the Shi-tdi-mu chief proclaimed a feast, and told the people to prepare to leave the village forever. On the feast day the women arranged the food basins on the ground in a long line leading out of the village. The people passed along this line, tasting a mouthful here or there, but without stopping, and when they reached the last basin they were beyond the limits of the village. Without turning around they continued on down into the valley until they were halted by the Snake people. An arrange- ment was eftected with the latter, and the Eagles built their houses in the Snake village. A few of the Eagle families who had become attached to Mashongnavi chose to go to that tillage, where their descendants still reside, and are yet held as close relatives by the Eagles of Walpi. The land around the East Mesa was then portioned out, the Snakes, Horus, Bears, and Eagles each receiving separate lands, and these old allotments are still approximately maintained. According to the Eagle traditions the early occupants of Tusayan came in the following succession: Snake, Horn, Bear, Middle Mesa, Oraibi, and Eagle, and finally from the south came the Water families. This sequence is also recognized iu the general tenor of the legends of the otlier groups. Shupaulovi, a small village quite close to Mashongnavi, would seem to have been established just before the coming of the Water people. Nor does there seem to have been any very long interval between the arrival of the earliest occupants of the Middle Mesa and this latest colony. These were the Sun people, and like the Squash folk, claim to have come ft'om Paldtkwabi, the Red Land, in the south. On their northward migration, when they came to the valley of the Colorado Chiquito, they found the Water people there, with whom they lived for some time. This combined village was built upon Homolobi, a round terraced mound near Sunset Crossing, where fragmentary ruins cover- ing a wide area can yet be traced. Incoming people from the east had built the large village of Awatubi, high rock, upon a steep mesa about nine miles southeast from Walpi. When the Sun people came into Tusayan they halted at that village and a few of them remained there permanently, but the others continued west to the Middle Mesa. At that time also they say Cliukubi, Shi- taimu, Mashongnavi, and the Squash village on the terrace were all occupied, and they built on the terrace close to the Squash village also. The Sun people were then very numerous and soon spread their dwell- ings over the summit where the ruin now stands, and many indistinct lines of house walls around this dilajjidated village attest its former size. Like the neighboring village, it takes its name from a rock near by, 30 • PUEBLO ARCHITECTURE. which is used as a place for the deposit of votive offerings, but the etymology of the term can not be traced. Some of the Bear people aLso took up their abode at Shnpaulovi, and later a nj^imu of the Water family called Batni, moisture, Ijuilt with them; and the diminished families of the existing village are stUl com posed entirely of these three nyumu. The next arrivals seem to have been the Asanyumn, who in early days lived in the region of the Chama, in New Mexico, at a village called Ka^kibi, near the place now known as Abicpiiu. When they left that region they moved slowly westward to a place called Tuwii (Santo Domingo), where some of them are said to still reside. The next halt was at Kaiw^ika (Laguna) where it is said some families still remain, and they staid ;ilso a short time at A'ikoka (Acoma); but none of them remained at that place. From the latter i)lace they went to Sioki (Zuiai), where they remained a long time and left a number of their people there, who are now called Aiyahokwi by the Zufii. They finally reached Tu- sayan by way of Awatubi. Tliey had been preceded from the same part of New Mexico by the Honaii nyumu (the Badger people), whom they found living at the last-named village. The Magpie, the Putc Kohu (Boomerang-shaped hunting stick), and the Field-mouse families of the Asa remained and built beside the Badger, but the rest of its grcmps con- tinued across to the Walpi Mesa. They were not at first permitted to come up to Walpi, which then occupied its present site, but were allotted a place to build at Coyote Water, a small spring on the east siile of the mesa, just under the gap. They had not lived there very long, however, when for some valuable ser\'ices in defeating at one time a raid of the Ute (who used to be called the Tcingawiiptuh) and of the Navajo at another, they were given for planting grounds all the space on the mesa summit from the gap to where Sichumovi now stands, and the samewidth, extending ai'ross the valley to the east. On the mesa summit tliey buUt the early jjortion of the house mass on the north side of the village, now known as Hano. But soon after this came a succession of dry seasons, which caused a great scarcity of food almost amounting to a famine, and many moved away to distant streams. The Asa jieople went to Tiip- kabi (Deep Canyon, the de Chelly), about 70 miles northeast from Walpi, where the Navajo received them kindly and sui)i)lied them with food. The Asa had preserved some seeds of the ])each, which they planted in the canyon nooks, and numerous little orchards still flourish there. They also brought the Navajo new varieties of food plants, and their relations grew very cordial. Tliey built houses along the base of the canyon walls, and dwelt there for two or three generations, during which time many of the Asa women were given to the Navajo, and the desceiulants of these now constitute a numerous clan among the Navajo, known as the KiAini, the High-house people. The Navajo and the Asa eventually qimrreled and the latter returned to Walpi, but this was after the arrival of the Hano, by whom they : MiNDELEFF.] TRADITIONS THE WATER PEOPLE. 31 fouiid their old houses occupied. The Asa were taken into the village of Walpi, being given a vacant strip on the east edge of the mesa, just where the main trail comes up to the village. The Navajo, Ute, and Apache had frequently gained entrance to the village by this trail, and to guard it the Asa l)uilt a house group along the edge of the cliff at that point, immediately overlooking the trail, where some of the people still live; and the kiva there, now used by the Snake order, belongs to them. There was a crevice in the rock, with a smooth bottom extending to the edge of the cliff and deep enough for a ki'koli. A wall was built to close the outer edge and it was at first intended to build a dwelling house there, but it was afterward excavated to its present size and made into a kiva, still called tlie wikwalhobi, the kiva of the Watchers of the High Place. The Walpi site becoming crowded, some of the Bear and Lizard people moved out and built houses on the site of the present Sichumovi; several Asa families followed them, and after them came some of the Badger people. The village grew to an extent considerably beyond its present size, when it was abandoned on account of a ma- lignant plague. After the plague, and within the present generation, the village was rebuilt—the old houses being torn down to make the new ones. After the Asa came the next group to arrive was the Water family. Their chief begins the story of their migration in tliis way In the long ago the Snake, Horn, and Eagle people lived here (in Tusayan), bnt their corn grew only a span high, and when they sang for rain the clond god sent only a thin mist. My people then lived in the distant Pa-l^t Kwa-bi in the South. There was a very bad old mm there, who, when he met any one, would spit in his tace, blow his nose upon him, and rub ordure upon him. He ravished the girls and did all manner of evil. Baholikonga got angry at this and turned the world upside down, and water spouted up through the kivas and through the fireplaces in the houses. The e.arth was rent in great chasms, and water covered everything except one narrow ridge of mud; and across this the serpent deity toM all the people to travel. As they journeyed across, the feet of the bad slipped and they fell into the dark water, but the good, after many days, reached dry land. While the water wa,s rising around the village the old people got on the tops of the houses, for they thought they could not struggle across with the younger people; but Baholikonga clothed them with the skins of turkeys, and they spread their wings out and floated in the air just above the surface of the water, and in this way they got across. There were saved of our people Water, Corn, Lizard. Horned Toad, .Sand, two families of Rabbit, and Tobacco. The turkey tail dragged in the water—hence the white on the turkey tail now. Wearing these turkey-skins is the reason why old people have dewlaps under the chin like a turkey; it is also the reason why old i)eoi>le use turkey-feathers at the religious ceremonies. In the story of the wandering of the Water people, many vague ref- erences are made to various villages in the South, which they constructed or dwelt in, and to rocks where they carved their totems at temporary halting places. They dwelt for a long time at Homolobi, where the Sun people joined them ; and probably not long after tlie latter left the Water people followed on after them. The largest number of this family seem : 32 PUEBLO ARCHITECTURE. to liiive made tbeir dwelliugs tiist at Maslioiiguavi and Sbupaulovi; but like the Sun people tliey soon spread to all the villages. The narrative of part of this Journey is thus given by the chief before quoted It occupied 4 years to cross tlip ilisrupteil couutrv. The kwakwauti (:i warrior order) weut ahead of the people aud carried seed of coru, beaus, melous, sqiiaslies, and cottou. They would ])laut coru iu the mud at early moruiug and liy noon it was ripe and thus the people were fed. When they reached solid {;rounermauent houses at Oraibi. There is no sjiecial tra- dition of this movement; it is only indicated by this circumstance, that in addition to the Water families common to every village, there are still in Oraibi several families of that people which have no representa- tives in any of the other villages. At a quite early day Oraibi became a place of importance, and they tell of being sufficiently populous to establish many outlying settlements. They still identify these with ruins on the detached mesas in the valley to the south and along the Moen-kopi ("i>lace of flowing water") and other intermittent streams in the west. These sites were occupied for the purpose of utilizing culti- vable tracts of land in their vicinity, and the remotest settlement, about 45 miles west, was especially devoted to the culti\'ation of cotton, thei place being still called by the Navajo and other neighboring tribes, the "cotton planting ground." It is also said that several of the larger ruins along the course of the Moen-kopi wore occupied by groups of the Snake, the Coj'ote, and the Eagle who dwelt in that region for a long period before they joined the people in Tusayan. The incursions of foreign bands from the north may have hastened that movement, and the Oraibi say they were compelled to withdraw all their outlying col- onies. An episode is related of an attack upon the main village when a number of young girls were carried off, and 2 or .3 years afterward the same marauders returned and treated with the Oraibi, who paid a ransom in corn and received all their girls back again. After a quiet Interval the pillaging bands renewed their attacks and the settlements on the Moen-kopi were vacated. They were again occupied after an- other peace was established, and this condition f>f alternate occui)ancy and abandonment seems to have existed until within quite recent time. While the Asa were still .sojourning in Canyon de Chelly, and before the arrival of the Hano, another bloody scene had been enacted in Tusayan. Since the time of the Antelope Canyon feuds there had been enmity between Awatubi and some of the other villages, especially Wali)i, and some of the Sikyatki refugees had transmitt(>d their feudal wrongs to their descendants who dwelt in Awatubi. Tiu-y had long been perpetrating all manner of offenses ; they had intercepted hunting 8ETH 3 34 PUEBLO ARCHITKCrrRE. parties from the utlior village(s, seized tlieir game, and .sometimes killed the himters ; tliey had fallen upon men in outlying corn fields, maltreating and sometimes slaying them, and threatened still more serious outrage. Awatubi was too strong for Walpi to attack single-handed, so the as- sistance of the other villages was sought, and it was determined to destroy Awatubi at the close of a feast soon to occur. This was the annual "feast of the kwakwanti," which is still maintained and is held during the month of November by each village, when the youths who have been (qualified by certain ordeals are admitted to the councils. The ceremonies last several days, and on the concluding night sjiecial rites are held in the kivas. At these ceremonies every man must be in the kiva to which he belongs, and after the close of the rites they all sleep there, no one being permitted to leave the kiva until after sunrise on the following day. There was stiU some little intercourse between Awatubi and AValjti, and it was easily ascertained when this feast was to be held. On the day of its close, the Walpi sent word to their allies "to prepare the war arrow and come," and in the evening the fighting bands from the other villages assembled at Walpi, as the foray was to be led by the chief of that village. By the time night had ftillen something like 150 marauders had met, all armed, of course; and of still more ominous import than their weapons were the firebrands they carried—shredded cedar bark loosely bound in rolls, resinous splinters of i>inon, dry greasewood (a furze very easily ignited), and pouches fidl of pulverizeil red iieppers. Secure iu the darkness from observation, the bands followed the Walpi chief across the valley, every man with his weapons in hand and a bundle of inflamnuibles on his back. Keaching the Awatubi mesa they cautiously crept up the steep, winding trail to the summit, and then stole round the village to the passages leading to the different courts holding the kivas, near which they hid themselves. They waited till just before the gray daylight came, then the Walpi chief shouted his war cry and the yelling bands ruslied to the kivas. Selecting their positions, they were at them iu a moment, and ([uickly snatching up the ladders through the hatchways, the only means of exit, tlie doomed occupants were left as helpless as rats in a trap. Fire was at liand in the numerous little cookiiig jtits, containing the jars of food iirepared for the celebrants, the inflammable bundles were lit and tossed into the kivas, and the piles of firewood on the terraced roofs were thrown down upon the blaze, and soon each kiva became a ftirnace. The red pepper was then cast upon the fire to add its choking tortures, while round the hatchways the assailants stood showering their arrows into the mass of struggling wretches. The fires were maintained until the roofs fell in and buried and charred the bones of the victims. It is said that every male of Awatubi who had passed infancy perished iu the slaughter, not one escaping. Such of the women and children as were s])ared were taken out, and all the houses were destroyed, after which the captives were divided among the dift'erent villages. BUREAU OF ETHNOLOGV EIGHTH ANNUAL REPORT PL. XII euRCAU Of tTMNOLOOV f'WH ,,,„,, „^^^ ^^ •kubi. : MISDELEFF. TRADITIONS—HANO. 35 The date of this last feudal atrocity can be miule out with some degree of exactness, because in 1692, Don Diego Vargas with a military force visited Tusayan and mentions Awatubi as a populous village at which he made some halt. The Hano (Tewa) claim that they have lived in Tusayan for five or six generations, and that when they arrived there was no Awatal)i iu existence; hence it must have been destroyed not long after the close of the seventeenth century. Since the destruction of Awatubi only one other serious affray has occurred betweeu the vdlages; that was betweeu Oraibi and Walpi. it appears that after the Oraibi withdrew their colonies from the south and west they took possession of all the unoccupied planting gxouuds to the east of the village, and kept reaching eastward till they encroached upon some land claimed by the Walpi. This gave rise to intermittent warfare in the outlying rtelds, and whenever the contending villagers met a broil ensued, until the strife culminated in an attack ui^on Walpi. The Oraibi chose a day when the Walpi men were all in the field on the east side of the mesa, but the Walpi say that their women and dogs held the Oraibi at bay until the men came to the rescue. A severe bat- tle was fought at the foot of the mesa, in which the Oraibi were routed and pursued across the Middle Mesa, where an Oraibi chief turned and implored the Walpi to desist. A conciliation was effected there, and harmonious relations have ever since existed between them. Until \vithin a few years ago the spot where they stayed pursuit was marked by a stone, on which a shield and a dog were dei)icted, but it was a source of irritation to the Oraibi and it was removed by some of the Walpi. In the early part of the eighteenth century the Ute ft'om the north, and the Apache from the south made most disastrous inroads upon the villages, in which Walpi especially suffered. The Navajo, who then lived upon their eastern border, also suttered severely from the same bands, but the Navajo and the Tusayan were not on the best terms and never made any alliance for a common defense against these invaders. Hano was peopled by a different linguistic stock from that of the other villages—a stock which belongs to the Rio Grande group. According to Polaka, the sou of the principal chief, and himself an enterprising trader who has made many journeys to distant localities—and to others, the Hano once lived in seven villages on the Eio Grande, and the village iu which his forefathers lived was called Tceewage. This, it is said, is the same as the present Mexican village of Peiia Blanca. The Hano claim that they came to Tusayan only after repeated solici- tation by the Walpi, at a time when the latter were much harassed by the Ute and Apache. The story, as told by Kwalakwai, who lives in Hano, but is not himself a Hano, begins as follows "Long ago the Hopi'tuh were few and were continually harassed by the Yiitamo (Ute), Yuittceiuo (Apache), and Dacibimo (Navajo). The chiefs of the Tcuin nyu- mu (Snake people) and the H6niu nyuravi (Bear people) met together and made the ba'ho (sacred plume stick) and sent it with a man from each of these people to the house of the Tewa, called Tceewadigi, whiih was far off on the Mliina (river) near Alavia (Saute F(). 36 PUEBLO ARCHITECTURE. The messengers did not succeed in persuading the Tewa to come and the embassy was sent three times more. On the fourth visit the Tewa consented to come, as the Walpi had oifered to dividt^ their laud and their waters with them, and set out for Tusayan, led by their own chief, the village being left in the care of liis son. This first band is said to have consisted of l-tG women, and it was afterwards followed by another and i^erhaps others. Before the Hano arrived there had been a cessation of hostile inroads, and the Walpi received them churlishly and revoked theii' promises re- garding the division of land and waters with them. They were shown where they could build houses for themselves on a yellow sand mound on the east side of the mesa just below the gap. They built there, but they were compelled to go for then- fo(jd up to Walpi. They could get no vessels to carry their food in, and when they held out their hands for some the Walpi women mockingly ijoured out hot porridge and scalded the fingers of the Hano. After a time the Ute came down the valley on the west side of the mesa, doing great harm again, and drove off the Walpi flocks. Then the Hano got ready for war; they tied buckskins around their loins, whitened their legs with clay, and stained their body and arms with dark red earth (ocher). They overtook the Ute near Wipho (about 3 rnUes north from Hano), but the Ute had driven the flocks up the steep mesa side, and when they saw the Tewa coming they killed all the sheep and ijiled the carcasses up for a defense, behind which they lay down. They had a few firearms also, while the Hano had only clubs and Ijows and arrows ; but after some fighting the Ute were driven out and the Tewa followed after them. The first Ute was killed a short distance beyond, and a stone heap still ( '?) marks the spot. Similar heaps marked the places where other Ute were killed as they fled before the Hano, but not far from the Sau Juan the last one was killed. Upon the return of the Hano from this successful expedition they were received gratefidly and allowed to come up on the mesa to live—the old houses built by the Asa, in the present village of Hano, being assigned to them. The land was then divided, an imaginary line between Hano and Sicliumovi, extending eastward entirely across the valley, marked the southern boundary, and fi-om this line as far north as the spot where the last Utah was killed was assigned to the Hano as their possession. When the Hano first came the Walpi saiil to them, "let us spit in your mouths, iincl you will learn our tongue," anil to this the Hano consented. When the Hano came up and bviilt ou the mesa they said to the Walpi, " let us spit iu your mouths and you will learn our timgue," but the AValpi would not listen to this, saying it would make them vomit. This is the reason why all the Hano can talk Hopi. and none of the Hopituh can talk Hano. The Asa and the Hano were close friends while they dwelt in New Mex- ico, and when they came to this region both of them were called Ilanoniuh by the other people of Tusayan. This term signifies the mode in which the women of these people wear their hair, cut off in front on a line with BUREAU OF ETHNOLOGY EIGHTH ANNUAL REPORT PL. XIII BUHEAU OF ETHNOLOO' MiNDKLEFF.] TRADITIONS—HANO AND WALPI. 37 the iiiouth and carelessly parted or hanging over the fa«e, the back hair rolled np in a compact quene at the nape of the neck. This uncomely fashion prevails with both matron and maid, while among the other Tusayan the matron parts her hair evenly down the head and wears it hanging in a straight queue on either side, the maidens wearing theirs in a curious discoid arrangement over each temx^le. Although the Asa and the Hano women have the same pecuhar fash- ion of wearing the hair, still there is no affinity of blood claimed between them. The Asa speak the same language as the other Tusayan, but the Tewa (Hano) have a (piite distinct language which belongs to the Tanoau stock. They claim that the occupants of the following pueblos, in the same region of the Eio Grande, are of their people and speak the same tongue. K6tite : 38 PUEBLO ARCHITKCTUKE. of the village. Their kiva is named after this circuinstance as that of "the Watchers of the Pligh Place." Some of the Bear and Lizard families being crowded for building space, moved from Walpi and built the first honses on the site of the present village of Sichnmovi, which is named from the .Sivwapsi, a shrnb which formerly grew there on some monnds (cluuiio). This was after the Asa had been in Walpi for some time probably ; about 125 years ago. Some of the Asa, and the Badger, the latter descendants of Avomen saved fioni the Awatnlji catastrophe, also moved to Sichnmovi, but a plague of smallxjox caused the village to be aban doned shortly afterward. This jiestilence is said to have greatly re- duced the number of the Tusayau, and after it disappeared tliere were many vacant houses in every village. Sichnmovi was again occupied by a few Asa families, but the first houses were torn down and new ones constructed from them. LIST OF TRADITIONARY GENTES. In the following table the early phratries (nyu-mu) are arranged in the order of their arrival, and the direction fi'oin which each came is given, excei)t in the case of the Bear people. There are very few represent- atives of this phratry existing now, and very little tradition extant con- cerning its early history. Tlie table does not sliow the condition of these organizations in the present community but as they appear in the tra- ditional accounts of their coming to Tusayau, although representatives of most of them can still be found in the various villages. There are, moreover, in addition to these, many other gentes and sub-geutes of more recent origin. The subdi\ision, or rather the miiltiplication of gentes may be said to be a continuous process; as, for exami)le, in "corn" can be found families claiming to be of the root, stem, leaf, ear, blossom, etc., all belonging to corn; but there may be several families of each of these components constituting district sub-gentes. At jiresent there are really but four phratries recognized among the Hopituli, the Snake, Horn, Eagle, and Eaiu, which is indifferently designated as Water or Corn 1. Ho'-nau—Bear. 2. Tcu'-a—Rattlesnake—from the west Ho'-nau Bear. and north—Continued. Ko'-kyafl-a Spider. U'-ae Cactus, candela- Tco'-zir Jay. bra, or branch- Hc'k-pa Fir. ing stemmed species. 2. Ttu'-a— Rattlesnake— from the west He'-wi Dove. .and north. Pi-vwa'ni Marmot. Tcu'-a Rattlesnake. Pi'h-tca Skunk. Yu'n-ya Cactus—opunti a. Ka-la'-ci-au-u Raccoon.i Pii'u-e Cactus, the spe- 3. A'-la—Horn—from the east. cies that grows So'-win-wa Deer. i n dome - like Tc'ib-io Antelo])c. masses. Pa'n-wa Mountain sheep. #-1 ^m f :# .ii^ m z m > < $ o t "^i*^- ,£ if" ,^,,^. -s» 'r ^.#t.ri ,^^^f^l^ Mr, 117 ' - " .V >i . : - MINDELEFF.] LIST OF TRADITIONARY GENTES. 39 4. Kwa'-hii—Eagle—from the west aud 6. A'sa—a plant (linknowu)—from the south. Chama—Continued. Road Kwa'-hii Eagle. Hoc'-bo-a runner, or Kwa'-yo Hawk. chaparral kwa'-yo cock.Mas-si' . . Chlckeu hawk. Po-si'-o Tda'-wa Sun. Magpie. Kwi'nobi Ka-ha'-bi Willow. Oak. Te'-bi Greasewood. 7. Ho-na'-ni—Badger—from the east. 5. Ka-tci'-na— Sacred dancer—from the Ho-ua'-ui Badger. east. Miin-ya'u-wn Porcupine. Wu-so'-ko Ka-tci'-na Sacred dancer. Vulture. Bu'-li Gya'-zro Parroquet. Butterfly. Un-wu'-si Eaveu. Bu-li'-so Evening prim- Si-kya'-tci Yellow bird. rose. Si-he'-bi Cottonwood. Na'-hii Medicine of all Spruce. kindsSa-la'-bi generic.; 6. A'sa—a plant (unknown) — fi'om the 8. Yo'-ki—Rain—from the south. Chama. Yo'-ki Rain. A'sa O'-mau Cloud. Tca'-kwai-na Black earth Kat- Ka'i-e Corn. cina. Mu'r-zi-bu-si Bean. Pu'tc-ko-hu Boomerang Ka-wa'i-ba-tuu-a .Watermelon. hunting stick. Si-vwa'-pi Bigelovia g r a- Pi'-ca Field mouse. veoleus. The foregoing is tlie Water or Rain phratry proper, but allied to them are the two following phratries, who also came to this region with the Water phratry. L1Z.\RD RABBIT. Kii'-kii-tci So'-wl Jackass rabbit. Ba-tci'p-kwa-si Species of liz- Tda'-bo Cottontail r a b Na'-nan-a-wi . ards . bit. Mo'-mo-bi Pi'-ba Tobacco. Pi'-sa White sand. Tcofi-o Pipe. Tdu'-wa Red sand. Tcu'-kai Mud. Polaka gives the following data Te'-wa geutes and phratries. Tewu Hopi'tiik Navajo. Ko"'-lo ) Ka'-ai Nata'" Corn. Cii i Pi'-ba Na'-to Tobacco. Ke Ho'-uau Cac Bear. Tce'-li Ca'-la-bi Ts'-co Spruce. Ke'gi Ki'-hu Ki-a'-ni House. Tun Tda'-wu Tjon-a-ai' Sun. O'-ku-wun O'-mau Kus Cloud. Nuf) TcTi'-kai Huc-klic« Mud. The gentes bracketed are said to " belong together," but do not seem to have distinctive names—as phratries. 40 PUEBLO ARCHITECTURE. SUPPLEMENTARY LEGEND. All interesting: ruin which occurs on a mesa point a short distance north of Mashongnavi is Ivuown to the Tusayan under the name of Payupki. There are traditions and legends concerning- it among the Tusayan, but the only version that could be obtained is not regarded by the writ«r as being up to the standard of those incoriiorated in the "Summary" and it is therefore given separately, as it has some sug- gestive value. It was obtained through Dr. Jeremiah Sullivan, then resident in Tusayan. The ijeople of Payupki spoke the same language as those on the first mesa (Walpi). Long ago they lived in the north, on the San Juan, but they were compelled to abandon that region and came to a place about 20 miles northwest from Oraibi. Being compelled to leave there, tliey went to Canyon de CheUy, where a band of Indians from the southeast joined them, with whom they formed an alliance. Together the two tribes moved eastward toward the .Temez Mountains, wlience they drifted into the valley of the Eio Grande. There they became converts to the fire-worship then prevailing, but retained their old customs and language. At the time of the great insurrection (of 1680) they sheltered the native priests that were driven from some of the Rio Grande villages, and this action created such distrust and hatred among the people that the Payupki were forced to leave their settlement. Their first stop was at Old Laguna (12 miles east of the modern village) and they had with them then some 3.5 or 40 of the priests. After leav- ing Laguna they came to Bear Spring (Fort Wingate) and had a fight there with the Apache, whom they defeated. They remained at Bear Spring for several years, until the Zuni compelled them to move. They then attem])ted to reach the San Juan, but were deceived in the trail, turned to the west and came to where Pueblo Colorado is now (the present post-office of Gauado, between Fort Defiance and Keam's Canyon). They remained there a long time, and through their success in farming became so favorably known that they were urged to come farther west. They refused, in consequence of which some Tusayan attacked them. They were captured and brought to Walpi (then on the point) and afterwards they were distributed among the villages. Previous to this capture the priests had been guiding them by feathers, smoke, and signs seen in the fire. AVhen the priest's omens and oracles had proved false the people were disposed to kill them, but the priests persuaded them to let it depend on a test case—offering to kill them- selves in the event of failure. So they had a great feast at Awatubi. The priests had long, hollow reeds inclosing various substances—feath- ers, flour, corn-pollen. Sacred water, native tobacco (piba), corn, beans, melon seeds, etc., and they formed in a circle at sunrise on the plaza and had their incantations and prayers. As tlie sun rose a priest step))cd forth before the people and blew through his reed, desirous of blowing MLNDELEFF] SUPPLEMENTARY LEGEND. 41 that which was therein away from him, to scatter it abroad. But the wind would not blow and the contents of the reed fell to the ground. The priests were divided into <;ronps, according to what they carried, lu the evening aU but two groups had blown. Then the elder of the twain turned his back eastward, and the reed toward the setting sun, and he blew, and the wind caught the feather and carried it to the west. This was accepted as a sign and the next day the Tusayan freed the slaves, giving each a blanket with corn in it. They went to the mesa where the ruin now stands and built the houses there. They asked for planting grounds, and fields were given them; but their crops did not thrive, and they stole corn fiom the Mashonguavi. Then, fear- ftil lest they should be sur^jrised at night, they built a waU as high as a man's head about the top of their mesa, and they had big doorways, which they closed and fastened at night. When they were compelled to plant corn for themselves they jilanted it on the ledges of the mesa, but it grew only as high as a man's knees; the leaves were very small and the grains grew only on one side of it. After a time they became friendly with the Mashongnavi again, and a boy from that village con- ceived a passion for a Payupki girl. The latter tribe objected to a mar- riage but the Mashongnavi were very desirous for it and some warriors of that village proposed if the boy could persuade the girl to fly with him, to aid and protect him. On an appointed day, about sundown, the girl came down from the mesa into the valley, but she was discovered by some old women who were baking pottery, who gave the alarm. Hearing the noise a party of the Mashongnavi, who were lying in wait, came up, but they encountered a party of the Pajnipki who had come out and a fight ensued. During the fight the young man was killed ; and this caused so much bitterness of feeling that the Pajaipki were frightened, and remained quietly in their pueblo for several days. One morning, however, an old woman came over to Mashongnavi to borrow some tobacco, saying that they were going to have a dance in her village in five days. The next day the Payupki quietly departed. Seeing no smoke from the village the Mashongnavi at first thought that the Payupki were preparing for their dance, but on the third day a band of warriors was sent over to inquire and they found the village abandoned. The estufas and the houses of the priests were pulled down. The narrator adds that the Payupki returned to San Felipe whence they came. CHAPTER II. RUINS AND INHABITED VILLAGES OF TUSAYAN. PHYSICAL PEATUKES OF THE PROVINCE. That portion of the southwestern plateau country comprised in the Province of Tusayan lias usually been approached from the east, so that the easternmost of the series of mesas upon which the villages are sit- uated is called the " First Mesa." The road for 30 or 40 miles before reaching this point ti-averses the eastern portion of the great plateau whose broken margin, farther west, furnishes the abrujit mesa-tongues upon which the villages are built. The sandstone measures of this plateau are distinguished from many others of the southwest by their neutral colors. The vegetation consisting of a scattered growth of stunted pinon and cedar, interspersed with occasional stretches of dull- gray sage, imparts an effect of exti'erae monotony to the landscape. The efl'ect is in marked contrast to the warmth and play of color frequently seen elsewhere in the pLateau country. The plateaus of Tusayan are generally diversified by canyons and buttes, whose precipitous sides break down into long ranges of rocky- talus and sandy foothills. The arid character of this district is espe- cially pronounced about the margin of the plateau. In the immediate vicinity of the villages there are large areas that do not support a blade of grass, where barren rocks outcrop through drifts of sand or lie piled in confusion at the bases of the cliffs. The canyons that break through the margins of these mesas often have a remarkable similarity of appear- ance, and the consequent monotony is extremely embarrassing to the traveler, the absence of running water and clearly defined drainage con- fusing his sense of direction. The occasional springs which furnish scanty water supply to the in- habitants of this region are found generally at gTcat distances apart, and there are usually but few natural indications of their location. They often occur in obscure nooks in the canyons, reached by tortuous trails winding through the talus and foothills, or as small seeps at the foot of some mesa. The convergence of numerous Navajo trails, however, ftirnishes some guide to these rare water sources. The series of promontories upon which the Tusayan \-illages are built are exceptionally rich in these seeps and springs. About the base of 42 BUREAU OF ETHNOLOGY EIGHTH ANNUAL REPORT PL. XVI BUREAU OF ETHNOLOGY EIOHTH ANNUAL REPOflT PL. MINDELEFF.] THE FIRST MESA. 43 the " First Mesa" (Fig. 1), withiu a distance of 4 or 5 miles from tlie vil- lages located upon it, there are at least five i^laces where water can be obtamed. One of these is a mere surface reservoir, but the others ap- pear to be permanent springs. The quantity of water, however, is so small that it produces no impression on the arid and sterile effect of the surroundings, except in its immediate vicinity. Here small patches of green, standing out in strong relief against their sandy back-grounds, mark the position of clusters of low, stunted peach trees that have ob- tained a foothold on the steep sand dunes. .-^ Fig. 1. View of the First Mesa. In the open jylains surrounding the mesa rim ((5,000 feet above the sea), are seen broad stretches of dusty sage brush and prickly grease- wood. Where the j)lain rises toward the base of the mesa a scattered growth of scrub cedar and piiion begins to appear. But little of this latter growth is seen in the immediate vicinity of the villages ; it is, however, the characteristic vegetation of the mesas, while, in still higher altitudes, toward the San Juan, open forests of timber are met with. This latter country seems scarcely to have come within the ancient builder's province; possibly on account of its coldness in winter and for the reason that it is open to the incursions of warlike hunting tribes. Sage brush and greasewood grow abundantly near the villages, and these curious gnarled and twisted shrubs fiumish the principal fuel of the Tusayan. Occasionally grassy levels are seen that for a few weeks in early sum- mer are richly carpeted mth multitudes of delicate wild flowers. The beauty of these patches of gleaming color is enhanced by contrast with the forbidding and rugged cliaracter of the surroundings; but in a very short time these blossoms disappear from the arid and parched desert 44 PUEBLO ARCHITECTURE. that they have temporarily beautified. These beds of bloom are uot seen iu the immediate vicinity of the present villages, but are unex- pectedly met with in portions of the neighboring mesas and canyons. After crossing the or 7 miles of comijaratively level country that intervenes between the mouth of Keam's Canyon and the first of the occupied mesas, the toilsome ascent begins; at first through slopes and dunes and then over masses of broken talus, as the summit of the mesa is gradually approached. Near the top the road is flanked on one side by a very abrupt descent of broken slopes, and on the other by a pre- cipitous rocky wall that rises 30 or 40 feet above. The road reaches the brink of the promontory by a sharp rise at a point close to the village of Hano. METHODS OF SURVEY. Before entering upon a description of the villages and nuns, a few words as to the preparation of the plans accompanying this paper will not be amiss. The methods pursued in making the surveys of the in- habited pueblos were essentially the same throughout. The outer wall of each separate cluster was run with a comi^ass and a tape measure, the lines being closed and checked upon the corner from which the beginning was made, so that the plan of each group stands alone, and no accumulation of error is possible. The stretched tapeline afi'orded a basis for estimating any deviations from a straight line which the wall presented, and as each sight was plotted on the si)ot these deviations are all recorded on the x>lan, and afford an indication of the degree of accuracy with which the building was carried out. U])()n the basis thus obtained, the outlines of the second stories were drawn by the aid of measurements from the numerous jogs and angles; the same process being repeated for each of the succeeding stories. The plan at this stage recorded all the stories in outline. The various houses and clusters were connected by compass sights and by measurements. A tracing of the outline plan was then made, on which the stories were distin- guished by lines of different colors, and upon this tracing were recorded all the vertical measurements. These were generally taken at every corner, although in a long wall it was customary to make additional measurements at intervening points. TTpon the original outline were then drawn all such details as coping stones, chimneys, trapdoors, etc., the tapeline being used where neces- sary to establish positions. The forms of the chimneys as well as their position and size were also indicated on this drawing, which was finally tinted to distinguish the ditterent terraces. Upon this colored sheet were located all oiienings. These were numbered, and at the same time described in a notebook, in which were also recorded the necessary vertical measurements, such as their height and elevation abovt die ground. In the same notebook the ojienings were also ftilly described. The ladders were located upon the same sheet, and were consecutively MINDtLEFF.l METHODS OF SURVEY. 45 lettered and describt'd in t\w uotebuok. This description furuishes a record of the ladder, its prqjectiou above the coping, if any, the differ- ence iu the length of its poles, the character of the tiepiece, etc. Altogether these notebooks furnish a mass of statistical data which has been of great service? iu the elaboration of this report and iu the preparation of models. Finally, a level was carried over the whole vil- lage, and the lieight of each corner and jog above an assumed base was determined. A reduced tracing was then made of the plan as a basis for sketching in such details of topography, etc., as it was thought ad- visable to preserve. These plans were primarily intended to be used in the construction of large scale models, and consequently recorded an amount of informa- tion that could not be reproduced upon the published drawings without causing great confusion. The methods followed in surveying the ruins underwent some changes from time to time as the work progressed. In the earlier work the lines of the walls, so far as they could be determined, were run with a com- pass and tapeliue and gone over with a level. Later it was found more convenient to select a- number of stations aiid connect them by cross- sights and measurements. These points were then platted, and the walls and lines of debris were carefully drawn in over the framework of lines thus obtained, additional measurements being taken when nec- essary. The heights of standing walls were measured from both sides, and openings were located on the plan and described in a notebook, as was done in tlie survey of the iuhabited villages. The entire site was then leveled, and fi'om the data obtained contour lines were drawn with a 5-foot interval. Irregularities in the directions of walls were noted. In the later plans of ruins a scale of symbols, seven in number, were employed to indicate the amount and distribution of the debris. The plans, as published, indicate the relative amounts of debris as seen upon the ground. Probable lines of wall are shown on the plan by dotted lines drawn through the dots which indicate debris. With this excep- tion, the plans show the ruins as they actually are. Standing walls, as a rule, are drawn in solid black; their heights appear on the field sheets, but could not be shown upon the published plans without con- fusing the drawing. Tlie contour lines represent an interval of 5 feet; the few cases in which the secondary or negative contours are used will not produce confusion, as their altitude is always given in figures. PLANS AND DESCRIPTIONS OF RUINS. The ruins described in this chapter comprise but a few of those found within the province of Tusayan. These were surveyed and recorded on account of their close traditional connection witli the present villages, and for the sake of the light that they migiit throw upon the relation of the modern pueblos to the innunu'rable stone buildings of unknown date so widely distributed over the southwestern plateau country. Such 46 PUEBLO ARCHITECTURE. traditional counectiou witli tlie present peoples could probably be estab- lished for many more of tbe ruins of this country by investigations sim- ilar to those conducted by Mr. Stephen in the Tusayan gi'oup; but this phase of the subject was not included in our work. In the search tor purely architectural evidence among these nuns it must be confessed that the data have proved disappointingly meager. No trace of the numerous constructive details that interest the student of pueblo archi- tecture in the modern villages can be seen in the low mounds of broken down masonry that remain in most of the ancient villages "of Tusayan. But little masonry remains standing in even the best preserved of these ruins, and villages known to have been occupied within two centuries are not distinguishable from the remains to Mhich distinct tradition. (save that they were in the same condition when the first people of the narrators' gens came to this region) no longer clings. Though but little architectural information is to be derived from these ruins beyond such as is conveyed by the condition and character of the masonry and the general distribution of the plan, tbe ])lans and relation to the topography are recorded as forming, in connection with tlie traditions, a more com- plete account than can perhaps be obtained later. In our study of architectural details, when a comparison is suggested between the i)ractice at Tusayan and that of the ancient builders, our illustrations for the latter must often be drawn from other portions of the builders' territory where better preserved remains furnish the neces- sary data. WAXPI RUINS. In the case of the pueblo of Walpi, a portion of whose people seem to have been the first comers in this region, a number of changes of sites have taken place, at least one of which has occurred within the historic period. Of the various sites occupied one is pointed out north of the gap on the first mesa. At the present time this site is only a low mound of sand-covered debris with no standing fragment of wall visible. The present condition of this early Walpi is illustrated in Fig. 2. In the absence of foundation walls or other definite hues, the character of the site is expressed by the contour hues that define its relief. Another of the sites occupied by the Walpi is said to have been in tlie open valley separating the first from the second mesa, but here no trace of the re- mains of a stone village has been discovered. This traditional location is referred to by Mr. Stejihen in his account of Walpi. The last site occupied previous to tlie present one on the mesa summit was on a lower bench of the first mesa promontory* at its southern extremity. Here the houses are said to have been distributed over quite a large area, and occasional ft-agments of masonry are still seen at widely sepa- rated points; but the ground plan can not now be traced. This was the site of a Spanish mission, and some of the Tusayan point out the position formerly occupied by mission buildings, but no architectural evidence of such structures is visible. It seems to be fairly certain, however, that BUREAU OF ETHNOLOGY EIGHTH ANNUAL HEPOBT PL. XVIII at«^»*=^' ElOHTH ANNUAL BEPOBT PU > tunvM or trxmixxty WINDELEFF.] OLD MASHONGNAVI RUINS. 47 this was the site of Walpi at a date well ^\nthin the liistoi-ie period, altliougli uow literally there is not one stone upon another. The de- struction in this instance has probably been more than usually complete on account of the close proximity of the succeeding pueblo, making the older remains a very convenient stone quarry for the construction of the houses on the mesa summit. Of the tlu-ee abandoned sites of Walpi re- ferred to, not one furnishes sufficient data for a suggestion of a ground plan or of the area covered. Fig. 2. Euins. Old Walpi mound. OLD MASHONGNAVI. In the case of Mashongiiavi we have somewhat more abundant ma- terial. It will be desirable to quote a few lines of narrative from the account of a Mashongnavi Indian of the name of Nuvayauma, as in- dicating the causes that led to the occupation of the site illustrated. We turned and came to tbe north, meeting the Apache and " Beaver Indians," with whom we had many battles, and lieing few we were defeated, after which we came 48 PUEIiLO AKCHITECTURE. up to Masliougnavi [the ruin ;it the "Giant's Chair"] aud gave that rock its name [name not known], and built our houses tliere. Tlic Apache came upon us again, witli the Comanche, and then we came to [Old Mash6ngnavi]. We lived there in peace many years, having great success with crops, and our people increased in num- bers, and the Apache came in great numbers and set fii'e to the houses and burned our corn, wliich you will find to-art]y cleared out one of thi'se chambers and used it as a depository for ceremonial idunie-sticks, etc., but the Navajo came and carried off their sacred deposits, tempted probably by their market value as ethnologic specimens. No trace of these kivas was visible at the time the ruins wtn'e surveyed. The Awatubi are said to have had sheep at the time the village was destroyed. Some of the Tusayan point out the remains of a large sheep corral near the spring, which they say was used at that time, but it is quite as likely to have been constructed for that purpose at a much later date. HORN HOUSE. The Horn House is so called because tradition connects this village with some of the people of the Horn phratry of the Hopituh or Tusayan. The ruin is situated on a projecting point of the mesa that forms the western flauk of Jeditoh Valley, not far from where the Holbrook road to Ream's Canyon ascends the brink of the mesa. Tlie village is almost completely demolished, no fragment of standing wall remaining in i)Iace. Its general plan and distribution are quite clearly indicated by the usual low ridges of fallen masimry ])artly covered by drifted sand. There is but little loose stone scattered about, the sand having tilled in all the smaller irregularities. It will be seen from the plan, I'l. vri, that the village has been built close to the edge of the mesa, following to some extent the irregularities of its outline. The mesa ruin at this point, however, is not very high, the more abrupt portion having a height of 20 or 30 feet. Near the north end of the village the ground slopes very sharply toward the east and is rather thickly covered with the small stones of fallen masonry, though but faint vestiges of rooms remain. In plan the ruin is quite elongated, following the direction of the mesa. The houses were quite irregularly disposed, j)articularly in the northern i)ortion of the ruin. But here the indications are too vague to determine whether the houses were originally built about one long court or about two or more smaller ones. The south end of the pueblo, however, still shows a well defined court bounded on all sides by clearly traceable rooms. At the extreme south end of the ruin the houses have very irregular outlines, a result of their adaptation to the topogra|)liy, as may be seen in tlic illustration. The plan shows the position of a small group of Cottonwood trees, just below the edge of the mesa and nearly opposite the center of the BUREAU OF ETHNOLOGY '' '^"^ « >i. 4^-., -"*i^' _E"'"TM ANNUAL REPORT PL. XX fy^/^^.')^\.M^^^ ^ >v>' A.:.,^ 'i-^ ^i. -^^v--, 4j'y/;,/^'^t f^4. « ; . SceJe O Z5 SO 75 TOO I I I I I I I I I ' II I I r I ] I ' I Via. 4. Kiiin near Moen-kopi, plaD. three rows around a small court. This ruin also follows the general northeastern trend which has been noticed both in the ruined and in the occupied pueblos of Tusayan. The rows here were only one room deep and not more than a single story high at any ])oint, as indicated by the very small amount of debris. As the plate shows, nearly the entire plan is clearly detined by fragments of staiuling walls. The walls are built of thin tablets of the dark-colored sandstone which caps the mesa. Where the walls have fallen the debris is comiiaratively free 54 PUEBLO ARCHITECTURE. from earth, indicating that adobe has been sparingly used. The walls, in places standing to a height of 2 or .3 feet, as may be seen in the illus- tration, PI. X, show unusual precision of workmanship and finish, re- sembling ill this respect some of the ancient pueblos farther north. This is to some extent due to the exceptional suitability of the talmlar stones of the mesa summit. The almost entire absence of pottery frag- ments and other objects of art which are such a constant accompani- ment of the ruins throughout this region strongly suggest that it was occupied for a very short time. In Chapter ill it will be shown that a similar order of occupation took place at Ojo Caliente, one of the Zuui farming villages. This ruin is probably of quite recent origin, as is the present village of Moen-kopi, although it may possibly have bek)uged to an earlier colony of which we have no distinct trace. This fertile and well watered valley, a veritable garden spot in the Tusayan deserts, must have been one of the first points occupied. Some small clifif- dwellings, single rooms in niches of a neighboring canyon wall, attest the earlier use of the valley for agricultural purposes, although it is doubtful whether these rude shelters date back of the Spanish invasion of the province. A close scrutiny of the many favorable sites in this vicinity would probably reveal the sand-encumbered remains of some more important settlement than any of those now known. RUINS ON THE ORAIBI WASH. The wagon road from Keam's Canyon to Tuba City crosses the Oraibi wash at a point about 7 miles above the village of Oraibi. As it enters a branch canyon on the west side of the wash it is flanked on each side by rocky mesas and broken ledges. On the left or west side a bold promontory, extending southward, is quite a conspicuous feature of the landscape. The entire fiat mesa summit, and much of the slope of a rocky butte that rises from it, are covered with the remains of a small pueblo, as shown on the plan. Fig. 5. All of this knoll except its east- ern side is lightly covered with scattered d6bris. On the west and north sides there are many large masses of broken rock distributed over the slope. Tliere is no standiug wall visible from below, but on closer approach several interesting specimens of masonry aie seen. On the north side, near the west end, there is a fragment of curved wall wliicli follows the margin of the rock on which it is built. It is about 8 or 10 feet long and 3 feet high on the outer side. The curve is carefully executed iind the workmanshij) of the masonry good. Farther east, and still on the north side, there is a fragment of masonry exhib- iting a reversed curve. This piece of wall spans the space between two adjoining rocks, and the top of the wall is more than 1(1 feet above the rock on wliich it stands. The shape of this wall and its relation to the surroundings are indicated on the plan. Fig. 5. On the south side of the ruin on the mesa surface, and near an outcropping rock, are the re- ..N 5 ,M \. -,.>^'^&(y-- MINDELEFF. I RUINS ON THE ORAIBI WASH. 55 mains of what appears to have been a circular room, perhaps 8 or 10 feet in diameter, though it is too much broken down to determine this accurately. Only a small portion of the south wall can be definitely traced. On the south slope of the mesa are indications of walls, too vaguely defined to admit of the determination of their direction. Similar vestiges of masonry are found on the north and west, but not extend- ing to as great a distance from the knoll as those on the south. Fig. 5. Ruin 7 miles uorth of Oraibi. In that portion of the ruin which lies on top of the knoll, the walls so for as traced conform to the shape of the site. The ground i)lan of the buildings that once occuitied the sloi)es can not be traced, and it is im- possible to determine whether its walls were carried through continu- ously. The masonry exhibited m the few surviving fragments of w^all is of unusually good quality, resembling somewhat that of the Fire House, Fig. 7, and other rains of that class. The stones are of medium size, not dressed, and are rather rougher and less flat than is usual, but the wall has a good finish. The stone, however, is of poor quality. Most of the debris about the ruin consists of small stone fragments and sand, comparatively few stones of the size used in the walls being seen. The material evidently came from the inunediate vicinity of the ruin. Pottery fragments were quite abundant about this ruin, most of the ware represented being of excei)tional (piality and belonging to the older types ; red ware with black lines and black and white ware were especially abundant. 56 PUEBLO ARCHITECTURE. There is (juite an extrusive view from the ruiu, the to]) ' the butte commanding' an outlook down the valley past Oraibi, about ~> miles iiortli. There is also an extended outlook up the . llowed by the wajL;ou road above referred to, and over two branch •, vs, one on the east and another of muoh less extent on the west. "L^ site was well adapted for defense, which nuist have been one of the nincipal motives for its selection. KWAITtKI. The ruin known to the Tusayan as Kwaituki (Fig. 0) is also on the west side of the Oraibi wash, 14 miles above Oraibi, and about 7 miles above the ruin last described. Its general reseniblanc the latter is very striking. The builders have apjiareutly been actuateu t,j' Fig. 6. Rviin 14 miles north of Oraibi (Ewaituki). the same motives in tlieir choice of a site, and their manner of utilizing it corresponds very closely. The crowning feature of the rocky kuoll in this case is a picturesque group of rectangular masses of sandstone, somewhat irregularly distributed. The bare summit of a large block-like mass still retains tlie vestiges of rooms, and probably most of the groups were at oii(> time covered with buildings, fin'niing a ])roniineiit citadel- like group in the midst of the village. To the nortli of this rocky butte a large area seems to have been at one time inclosed by buildings, form- ing a court of unusual dimensions. Along the outer margin of the pueblo ^r^^^r, ^^^^ 'j}lr^ . } v^ MiNDELEFP.] FIKE HOUSE RUIN. 57 occasional fragmeuts of walls defiue former rooms, but the amount aud character of the debris indicate that the inner area was almost completely inclosed with buildings. The i-emains of masonry extend on the south a little beyond the base of the central grou]) of rocks, but here the ves- tiges of stonework are rather faint and scattered. In the nearly level tops of some of the rocks forming tlie central pile are uiaaiy smoothly worn depressions or cavities, which have evidently been used for the grinding and shaping of stone implements. A remarkable feature occurring within this village is a cave or under- ground fissure in the rocks, which evidently had been used by the in- habitants. The mouth or entrance to this cavern, partly obstructed and concealed at the time of our visit, occurs at the point A on the i)lau. On clearing away the rubbish at the mouth and entering it was found so obstructed with broken rock' and fine dust that but little progress could be made in its exploration ; but the main cre\ice in the rock could be seen by artificial light to extend some 10 feet back from the mouth, where it became very shallow. It could be seen that the original cavern had been improved by the pueblo-builders, as some of the timbers that had been placed inside were still in position, and a low wall of masonry on the south side remained intact. Some Navajos stated that they had discovered this small cave a couple of years before and had taken from it a large unbroken water jar of ancient pottery and some other speci- mens. The place was probably used by the ancient occupants simjily for storage. Fragments of pottery of excellent quality were very abundant about this ruin and at the foot of the central rocks the ground was thickly strewn with fragments, often of large size. The defensive character of this site parallels that of the ruin 7 miles farther south in quite a remarkable manner, and the villages were apparently built and occupied at the same time. TEIU'GKIIII', OH FIHE HOUSK. About 15 miles northeast of Ream's Canyon, and about 25 miles from Walpi, is a small ruin called by the Tusayan "Tebugkihu," built by peo- ple of the Fire gens (now extinct). As the plan (Fig. 7) clearly shows, this pueblo is very different from the typical Tusayan villages that have been previously described. The api)arent unity of the plan, and the skillful workmanship somewhat resembling the pueblos of the Chaco are in marked contrast to the irregularity- and careless ciuistruction of most of the Tusayan ruins. Its distance from the center of the province, too, suggests outside relationship; but still the Tusayan traditions un- doubtedly connect the place with sonu^ of the ancestral gentes, as seen in Chai)ter i. The small and compact cluster of rooms is in a remarkable state of preservation, especially the outside wall. This wall was carefully and massively constructed, and stands to the height of several feet around 58 PUE13LO ARCHITECTURE. the entire circumfeieuee of the ruin, except along the brink of the cliff, as the plan shows. This onter wall eontains by far the largest stones yet found incor- porated in pueblo masonry. A fraguieut of this masonry is illustrated in PI. XI. The largest stone shown measures about 5 feet in length, and the one adjoining on the right measures about 4 feet. These dimen- sions are quite remarkable in pueblo masonry, which is distinguished by the use of very small stones. The well defined outer wall of this cluster to the unaided eye appears to be elliptical, but it will be seen liom the plan that the ellipse is somewhat pointed on the side farthest fi-om the cliff. As in other cases of ancient ' pueblos with curved outlines, the outer wall seems to have been built first, and the inner rooms, while kept as rect- angular as possible, were adjusted to this curve. This arrangement often led to a cumulating divergence from radial lines in some of the partitions, which irregularity was taken up in one room, as in this instance, in the space near the gate. The outer waU is uniform in construction so far as pre- served. Many irregularities appear, however, in the construction of the in- ner or partition walls, and some of the rooms show awkward attempts at ad- justment to the curve of the outer wall. The ruin is situated on the very Oval (Fire House) ruin, plau (Tebngkihu). brink of a small canyon, wliich i)rob- ably contained a spring at the foot of the cliff close under the ruin site, as the vegetation there has an unusual appearance of freshness, sug- gesting the close jjroximity of water to the surface. A steep trail evi- dently t-onnected the village with the bottom of the canyon. Some of the rocks of the mesa rim were marked by numerous cup-like cavities similar to those seen at Kwaituki, and used in the polishing and form- ing of stone implements. The type of pueblo here illustrated belonged to a people who relied largely on the architecture for defense, differing in this respect from the spirit of Tusayan architecture generally, where the inaccessible character of the site was the chief dependence. The ruin called Ohukubi by the Tusayan (PI. xii) is situated on the Middle Mesa, about 3 miles northeast of Mashongnavi. It occupies a promontory above the same broad sandstone ledge that forms such a »nNUKLEi.F.i CHlIKUm AND rAVlU'KI UllINS. 59 conspicuous feature in the vicinity <>l' Masiion^ruivi and Shnpaulovi, and wliicli supjiorts the buttes upon whi(^li these villages are built. Little masonry now remains on this site, but her(^ and tlierti ai Crag- nient aids in dclinin^' the jiencral jthm of thi^ pueblo. In ji'enera! form the village was a large rectangle with a line of buildings across its center, dividing it into two unequal courts, and a, |)r()jectiiig wing on the west side. As may be .seen from the illustration, on(^ end of the ruin forms a clearly delined rc^ctanguhir court, composfMl ol' buildings mostly two rooms dec]). Here, as in other ruins of Tusayan, tins ar- rangement about inclosed courts is in contrast with the ])aiallelisin of rows, so noticeable a featur(^ in the occupied villages. At the c^ast (Mid of the ruin are .several curious excavations. The soft sandstone has been hollowed out to a depth of about 10 incdies, in i)rolonga,tion of the outlines of adjoining rooms. Such excavation to obtain h^vcl floors is quite unusual among tiie pueblo builders; it was i)racti(!ed to a very small extent, and only where it could be done with little trouble. Any serious in('(pmlity of .siirfacu' was nsuaJly incor]iorated in the construc- tion, as will be noticed at Walpi (I'l. xxiii). Vestiges of masoniy in- dicating detached rooms were seen in (mcIi of the courts of the main rectangle. On the slope oi' the hill, just above tlui broad ledge previously de- scribed, there is a flue spring, but no trac(! of a trail connecting it. with the pueblo could be found. This village was advantageously placed for defense, but not to the same degree as Payupki, illustrated in PI. xiii. The ruin called Payupki (PI. xiii) occupies the summit of a bold promontory .south of the trail, from Walpi to Oraibi, and about (i miles northwest from Mashongna.vi. The outer extremity of tliis promontoi'y is separated from the mesa l)y a deirj) notcii. Tiie suiiuuit is riiached from the mesa by way of the neck, as tin' outer point itself is very abrupt, much of the sandstone ledge l)e,ing vertical. A bench, 12 or 15 feet below the summit and in places (piite bioad, (Micirch^s the promontory. This bench also breaks off very abruptly. As may be seen from the i)lan, tlie village is (|uite symmetrically laid out and well arranged for defense. It is placed at the mesa, end of the promontory cap, and for greater security the .second ledge has also been fortified. All along the outer margin of this ledge are the remains of a stone wall, in some places still standing to a height of 1 or li feet. This wall a|)i)ears to have extended originally all along the IcMlge around three sides of the village. The steei)ness of the cliff on the remaining side rendered a wall su])ertluous. On the jilain below this i)romontory, and inunediately under the overhanging cliff, are two corrals, and also 60 PUEBLO AKCHITECTUKE. the remains of a structure that resembh>s a kiva, bat which apjiears to be of recent construction. In the village i)roi)er (PI. xiv) are two distinctly traceable kivas- One of these, situated in the court, is detached and appears to have been partly uudergrcjund. The other, located in the southeast end of the village, has also, like the first, apparently been sunk slightly l)elow the surface. There is a jog in the standing wall of this kiva which corre- sponds to that usually found in the typical Tusayau kivas (see Figs. 22 and 25). On the i»roniontory anil east of the village is a single room of more than average length, with a well formed door in the center of one side. This room has every appearance of being contemporary with the rest of the village, but its occurrence in this entirely isolated position is very unusual. Still farther east there is a mass of debris that may have belonged to a cluster of six or eight rooms, or it may possibly be the remains of temjiorary stone shelters for outlooks over crops, built at a lat(*r date than the i)ueblo. As may be seen fi'om the illustration (PI. XV), the walls are roughly built of large slabs of sandstone of vari- ous sizes. The work is rather better than that of modern Tusayan, but much inferior to that seen in the skillfully laid masonry of the ruins farther north. In many of these walls au occasional sandstone slab of great length is introduced. This peculiarity is probably due to the character of the local material, which is more varied than usual. All of the stone here used is taken from ledges in the immediate vicinity. It is usually light in color and of loose texture, crumbling readily, and subject to rapid decay, particularly when used in walls that are roughly constructed. Much of the pottery scattered about this ruin has a very modern ap- pearance, some of it having the characteristic surface finish and color of the Rio Grande ware. A small amount of ancient pottery also occurs here, some of the fragments of black and white ware displaying intri- cate fret i)atterns. The quantity of these potsherds is quite small, and they occur mainly in the refuse heaps on the mesa edge. This ruin combines a clearly defined defensive plan with utilization of one of the most inaccessible sites in the vicinity, jiroducing alto- gether a combination that would .seem to have been impregnable by any of the ordinary methods of Indian warfare. BUREAU OF ETHNOLOGY EIGHTH ANNUAL REPORT PL. XXV FOOT TRAIL TO WALPI. MiNiiELEFF.) INHABITED VILLAGK8 HAND. 61 PLANS ANU DESCRIPTIONS OF THE INHABITED VILLAGES. HANO. The villii.LfP of Hauo, or Tewa, is intrusive and does not properly be- long to tlie Tasayan stock, as appears from tiieir own traditions. It is somewhat h)osely planned (PI. xvi) and extends nearly across the mesa tongue, whicli is here quite narrow, and in general tliere is no ap- preciable difference between the arrangement here followed and that of the other villages. One portion of the village, however, designated as House No. 5 on the plan, differs som(^wliat ti-om the typical arrange- ment in long irregular lows, and a|)|)roaches the i)yramidal form found among the more eastern pueblos, notably at Taos and in portions of Zutli. As has been seen, tradition tells us that this site was taken uj) by the Tewa at a late date and subsequent to the Spanish conquest; but some houses, formerly belonging to the Asa people, formed a nucleus about which the Tewa village of Hauo was constructed. The pyramidal house occupied by the old governor, is said to have been built over such remains of earlier houses. The largest building in the village appears to have been added to from time to time as necessity for additional space arose, resulting in much the same arrangement as that characterizing most of the Tusayau houses, viz, a long, irregular row, not more than three stories high at any point. The small range marked No. 4 on the plan contains a sec- tion three stories high, as does the long row and also the pyramidal cluster above refen-ed to. (PI. xvii.) The kivas are two in number, one situated within the village and the other occupying a position in the margin of the mesa. These cere- monial chambers, so far as observed, appear to be much like those in the other villages, both in external and internal arrangement. Within the last few years the horse trail that afforded access to Hauo and Sichumovi has been converted into a wagon road, and during the progress of this work, under the supervision of an American, consider- able blasting was done. Among other changes the marginal kiva, which was nearly in line with the proposed improvements, was removed. This was done des])ite the protest of the older men, and their predic- tions of dire calamity sure to follow such sacrilege. A new site was selected close by and the lunvly acquired knowledge of the use of pow- der was utilized in blasting out the excavation for this subterranean chamber. It is altogether prol)able that the sites of all former kivas were largely determined by accident, these rooms being built at points where natural fissures or open spaces in the broken mesa edge fur- nished a suitable depression or i^avity. The builders were not capable of w.trking the stone to any great extent, and their operations were probably limited to trimming out such natural ex(%ivations and in jiart lining them with masonry. There is a very noticeable scarcity of roof-holes, aside from those of the tirst terrace. As a rule the first terrace has no external openings 62 Pl'EliLO AKCHITIX'TUKE. Oil the grouud a,ud is t'litciTil tioiii it.s roof thiouyli large triip-doors, as sliowu on the plans. The lower rooms within this first terrace are uot inhabited, but are nsert as storerooms. At several points ruined walls are seen, remains of abandoned rooms that have fallen into decay. Occasionally a rough, buttress-like projec- tion from a wall is tlie only vestisi'e of a room or a cluster of rooms, all traces on the ground ha\'ing been obliterated. The mesa summit, that forms the site of this village, is nearly level, with very little earth on its surface. A thin accumulation of soil and rubbish lightly covers the inner ccmrt, but outside, along the face of the long row, the bare rock is exposed continuously. Where the rooms have been abandoned and the walls have fallen, the stones have all been utilized in later (constructions, leaving no vestige of the former wall on the rocky site, as the stones of the masonry have always been set upon the surface of the rock, with no excavation or preparation of footings of any kind. SICHUMOVI. According to traditional accounts this village was founded at a more recent date than Walpi. It has, however, undergone many changes since its first establishment. The principal building is a long irregular row, siiniliar to that of Hano (PI. XViii). A portion of an L-shaped cluster west of this row, and a small row near it parallel to the main building, form a rude aiiproximation to the inclosed court arrangement. The terracing here, however, is not always on the court side, whereas in ancient examples such arrangement was an essential defensive feature, as the court furnished the only approach to ujiper terraces. In all of these villages there is a noticeable tendency to face the rows eastward instead of toward the court. The motive of such uniformity of direction in the houses must have been strong, to counteract the tendency to adhere to the ancient arrangement. The two kivas of the village are built side by side, in contact, probably on account of the presence at this point of a favorable fissure or depres- sion in the mesa surface. On the south side of the village are the reonains of two small clusters of rooms that apparently have been abandoned a long time. A portion of a room still bounded by standing walls has been utilized as a corral for burros (PI. xix). At this village are three small detached houses, each composed of but a single room, a feature not at all in keeping with the spirit of pueblo construction. In this instance it is iirobably due to the selection of the village as the residence of whites connected with the agency or school. Of these single-room houses, one, near the south end of the long row, was being built by an American, who was living in another such house near the middle of this row. The third house, although foirly well preserved at the time of the survey, was abandoned and falling into ruin. Adjoining the middle one of these three buildings on : BUREAU OF ETHNOLOGY '% \ % , 1 ^' f -v,. §, 4=^ I %"= -V t \\ft I f %J ^-^i A \ EIGHTH ANNUAL REPORT Pu XXVI ^!t«:.^^aq?{| '%..;^:p3,,.j*^5,^ -T'-^v^ EIGHTH ANNUAL REPORT PI- XXVI iUNDELEFF.) INHABITED VILLAGES WALPL 63 the south side are the outlines of tw^o small compartments, which were evidently built as corrals for burros and are still used for that i)urpose. This village, though limited to two stories in height, has, like the others of the first mesa, a number of roof holes or trapdoors in the upper story, an approach to the Zuni practice. Tliis feature among the Tusayan villages is probably due to intercourse with the more eastern pueblos, for it seems to occur chietiy among those having such communication most frequently. Its presence is probably the result simply of borrow- ing a convenient feature from those who invented it to meet a necessity. The conditions under which the houses were built have hardly been such as to stimulate the Tusayan to the invention of such a device. The uniform height of the second-story roofs seen in this village, con- stituting an almost unbroken level, is a rather exceptional feature in pueblo architecture. Only one depression occurs in the whole length of the main row. Of all the pueldos, occupied or in ruins, within the i)ro\inces of Tusayan and Cibola, Walpi exhibits the widest departure from the typical pueblo arrangement (PL xx). The carelessness characteristic of Tusayan architecture seems to have reached its culmination here. The confused arrangement of the rooms, mainly due to the irregularities of the site, contrasts with the work at some of the other villages, and bears no comparison with much of the ancient work. The rooms seem to have been clustered together with very little regard to symmetry, and right angles are very unusual. (See Fig. 8.) The general plan of the village of to-day (confirms the traditional ac- counts of its foundation. According to these its growth was gradual, be- ginning with a few small clusters, which were added to from time to time as the inhabitants of the lower site upon the spur of the mesa, where the mission was established, moved up and joined the pioneers on the summit. It is probable that some small rooms or clusters were built on this conspicuous promontory soon after the first occupation of this region, on account of its exceptionally favorable position as an * mtlook over the fields (PL XXI). Trough the peculiar conformation of the site on which the village has been built has produced an unusual irregularity of arrangement, yet even here an imperfect example of the tyi^ical inclosed court may be found, at one point containing the principal kiva or ceremonial chamber of the village. It is probable that the a(;cideutal occurrence of a suitable break or depression in the mesa top determined tlie position of this kiva at an early date and that the first buildings clustered about this point. A unique feature in this kiva is its connection with a secoml subter- ranean chamber, reached from the kiva through an ordinary doorway. The depres,sion used for the kiva site must have been either larger than was needed or of such form that it could not be thrown into one rec- 64 PUEBLO AECHITECTURE. tangular chamber. It was impossible to ascertain tbc form of this second room, as the writer was not permitted to apjjroach the connect- ing doorway, which was closed with a slab of cottonwood. This chamber, used as a receptacle for religious i)araphernalia, was said to connect with an upper room within the cluster of dwellings close by, but this could not be verified at the time of our visit. The plan indicates that such an adjoining chaudjcr, if of average size, could easily extend partly under the dwellings on either the west or south side of the court. The rocky mesa .^ummit is quite irregular in this vicinity, with rather an abrupt ascent to the passageway on the south as shown in PI. XXii. Southeast from the kiva there is a large mass of rocks projecting above the general level, which has been incori)orated into a cluster of dwelling rooms. Its character and relation to the architecture may be seen in PI. xxiii. So irregular a site was not likely to be built upon until most of the available level surface had been taken n\). for Qvcn in masonry of much liigher development than can be found in Tusayan the builders, unable to overcome such obstacles as a large mass of protruding rock, have accom- iiiodated their buildings to such irregularities. This is very noticeable in the center cluster of Mummy Cave (in Canyon del Muerto, Arizona), where a large mass of sandstone, fallen ft'om the roof of the rocky niche in which the houses were built, has been inc the house Tnpoj^'aph\ ot the site of Walpi. cluster. Hetween this and an- other kiva to the north the mesa top is nearly level. The latter kiva is MiNUELEFF] THE WALPl PROMONTORY. 65 iilso subterraueau aud was built in au accideutal Ijreak iu saudstoue. On the very margin of this fissure stands a curious isolated rock that has survived the general erosion of the mesa. It is near this rock that the celebrated Snake-dunce takes place, althougli the kiva from which the dancers emerge to perform the open air ceremony is not adjacent to this monument (PI. xxiv). A short distance farther toward the north occur a group of three more kivas. These are on the very brink of the mesa, and have been built iu recesses in the crowning ledge of sandstone of such size that they could conveniently be walled up on the outside, the outer sur- face of rude walls being continuous with the precipitous rock face of the mesa. The positions of all these ceremonial chambers seem to corresiiond with exceptionally rough and broken portions of the mesa top, showing that their location in I'elation to the dwelling clusters M^as due largely to accident and does not possess the significance that position does in many ancient pueblos built on Unci and unencumbered sites, where the adjustment was not controlled \>y the character of the surface. The Walpi promontory is so abrupt and diflflcult of access that there is no trail by which horses can be brought to tlie village without i)as- sing through Hano and Sichumovi, traversing the whole length of the mesa tongue, and crossing a rough break or depression in the mesa .summit close to the village. Several foot trails give access to the vil- lage, partly over the nearly perpendicular faces of rock. All of these have required to be ai'tifii'ially improved in order to render them prac- ticable. Plate XXV, from a photograph, illustrates one of these trails, which, a portion of the way, leads ui) between a huge detached slab of sandstone and the fiice of the mesa. It will be seen that the trail at this point consists to a large extent of stone steps that have been built in. At the top of the flight of steps where the trail to the mesa summit turns to the right the solid sandstone has been pecked out so as to furnish a series of footholes, or steps, with no projection or hold of any kind alongside. There are several trails on the west side of the mesa leading down both from Walpi and Sichumovi to a spring below, which are (luite as abrupt as the exami)le illustrated. All the water used in these villages, except such as is caught during showers in the basin- like water pockets of the mesa top, is laboriously brought u]) these trails in large earthenware canteens slung over the backs of the women. Sui)plies of every kind, provisions, harvested crops, fuel, etc., are brought up these steep trails, and often fron^ a distance of several miles, yet these conservative people tenaciously cling to the inconvenient sit- uation selected by their fathers long after the necessity for so doing has passed away. At present no argument of convenience or comfort seems sufticient to induce them to abandon their homes on the rocky heights and build near the water supply and the fields on which they depend for subsistence. 8 ETH 5 (36 PUEBLO ARCHITKCTrKE. One of tbf, tniiLs referred to iu the de.serii)tioii of JJaiio has been eon- verted into a wagon road, as lias been already described. The Indians preferred to ('xi)eiid th(> enormous iunouiit of lalior necessary to convert this bridle patli into a wagon road in order slightly to overcome the inconvenience of transporting every necessary to the mesa upon their own backs or by the assistance of burros. This concession to modern ideas is at best but a poor substitute for the convenience of homes built in the lower valleys. MASHONCNAVI. ^lasliongnavi, situated on tlie summit of a rocky knoll, is a compact though irregular village, and the manner in which it conforms to the ft^f - «- -"^-~-, Fig. 9. Maslumgmivi and Shupaulovi from Sluunopavi. general outline of the available ground is shown ou the plan. Con- venience of access to the fields on the east and to the otlier villages probably promi)ted the first occui)ation of the east end of this rocky butte (PI. XXVI). In Mashongnavi of to-day the eastern portion of the village forms a more decided court than do the other portions. The completeness in itself of this eastern end of the pueblo, in connection with the form of the adjoining rows, strongly suggests that this was the first portion of the pueblo built, although examination of the masonry and construction furnish but imperfect data as to the relativ(> age of diftercnt portions of the village. One uniform gray tint, with oidy slight local variations ill character and finish of masonry, imparts a monotonous effect of antiq- uity to the whole mass of dwellings. Here and there, at rare intervals, is seen a wall that has been newly plastered; but, ordinarily, masonry of 10 years' age looks nearly as old as that built 200 years earlier. Another feature that suggests the greater antiquity of the eastern court of the pueblo is the presence and manner of occurrence here of the kiva. The old builders may Lave been influenced to some extent in their choice of site by the jiresetice of a favorable depression for the construi'- tion of a kiva, though this particular exam])le of the ceremonial room is only partly subterranean. The other kivas are almost or quite below the ground level. Although a favorable depression might readily occur on the summit of the knoll, a deep cavity, suitable for the construction of tlie subti'rranean kiva, would not be likely to occur at such a distance from the margin of th«^ sandstone ledge. The builders evidently ]ire- ferred to adopt such half-way measures with their first kiva in order to ^^^ 'S..l/^'' MINDELEFF.] GROWTH OF MASHONGNAVI. 67 secure its iuclosure witliiu tlie court, tlius couforming to tlie typical pueblo arrangement. The numerous exceptions to this arrangement seen in Tusavaii are due to local causes. The general view of Mashong- ^--''* -- JC' Fig. 10. Diagram showing growth of MashoDgnavi. navi given in PI. xxvii shows that the site of this pueblo, as weU as that of its neighbor, Shupaulovi, was not i)articularly defensible, and that this fact would have weight in securing adherence in the first por- 68 PUEBLO ARCHITECTLKE. tioii of the pueblo Imilt to the defeusive inclosed eourt coutaiuiug the ceremouial chamber. The plan strousily iutlieates that the other courts of the iiuebhi were ad(lelo, without changing its exterior appearance. In order to make clearer this order of growth in Mashongnavi, a series of skeleton diagrams is added in Figs. 10, 11, and 12, giving the outlines of the ]iueblo at various supposed periods in the course of its enlarge- ment. The larger plan of the village (PI. xxvi) serves as a key to these terrace outlines. The first diagram illustrates the supposed original cluster of the east court (Fig. 10), the lines of which can be traced on the larger plan, and it includes the long, nearly straight line that marks the western edge of the third story. This diagram shows also, in dotted Hues, the gen- eral plan that may have guided the first additions to the west. The second diagram (Fig. 11) renders all theabovematerlalinfull tint, again indicating further additions by dotted lines, and so on. (Fig. 12.) The ])ortions of a terrace, which face westward in the newer courts of the pueblo, illustrated in PI. xxix, were pi-obabl\- built after the western row, completing the indosure, and were far enough advanced to indi- cate definitely an inclosed court, upon which the dwelling rooms faced. BUREAU OP ETHNOLOGY EIGHTH ANNUAL REPORT PL. XXX SHUPAULOVI. MLNDELEFF.] SHUPAULOVI. 71 SlUI'Al LOVI. This village, by far the smallest pueblo of tlie Tusayaii group, illus- trates a simple aud direct use of the i^riueiple of the iuclosed court. The plan (PI. xxx) shows that the outer walls ai'e scarcely broken by terraces, and nearly all the dwelling aiJartmcnts open inwards upon the inclosure, in this respect closely following the previously described ancient tyi)e, although widely differing from it in the irregular disposi- tion of the rooms. (PI. xxxi.) A com))arison with the first of the series of diagrams illustrating the growth of iMashonguavi, will show how similar the villages may have been at one stage, and how suitable a nucleus for a large i)ueblo this village would ])rove did space and character of the site pernut. Most of tlic available summit of the rocky knoll has already been covered, as will be seen from the toi)ogra]>hic sketch of the site (Fig. 13). The ])]an sliows also that some efforts at Fig. 13. Topography of the site of Shupanluvi. extension of the pueblo have been made, but the houses outside of the main cluster have been abandoned, and are rapidly going to ruin. Several small rooms occur on the outer faces of the rows, but it can be readily seen that they do not form a part of the original plan but were added to an already complete structure. In the inclosed court of this j)ueblo occurs a small box-like stone inclosure, covered with a large slab, which is used as a sort of shrine or depository for the sacred plume sticks and other ceremonial offerings. 72 PUEBLO ARCHITECTURE. This featuie is found at some of the other villages, notably at Masliong- navi, in tlie cfiitral coiut, and at llano, where it is locatcid at sonic dis- tance outside of tii(! villaj^c, near the main trail to tiie mesa. Tlic i)l;in of this small village shows three covered passageways sim- ilar to tiiosc^ noted in Walpi on the first nuisa, thongh their jiresejice here can not be ascribed t(j the same motives that impelled the Walpi tfj build in tliis way; for thi' densely crowded site (iccui)ied by tiie lat- ter com])elled them to resort to this expedient. One of tiiese is illus- tratcMl in I'l. xxxii. Its presence may be due in this instau<'-e to a deter- mination to n(lii(!re to the ])i()t('ctcil court wliih- sec^iiing to securer con- venient means of access to the inclosed Mica. It is remarkable that this, tlu! smallest of tlie group, should c(nit:iin this feature. Tills village lias but two Ui\iis, one of which is on the rocky suiiiiiiit near the houses and the otlicr on the lower ground near the loot of the trail that leads to the village. The ujiper kiva is nearly subterranean, th(! roof being but a little, above the ground on the side toward the village, hut as the rocky sit(^ slo])es awa.s a ixntioii of side wall is ex- posed. This was ri)uj;hl.\- iiuilt, with no attempt to impart finish to its outer face, either liy careful laying of the masonry or by jilastering. IM. XXX'III illustrates this kiva in connection with the southeastern por- tion of the vi]]ag(\ Tlie plan sliows how tll(^ jirolongation of the side rows of tin? villager forms a suggestion of a second court. Its develop- ment into any such featuie as the secomlary or additional (uiiirts of Mashongnavi was jirohibitcd liy the restricted site. As in other villag<^s of this grou)), the desire to adhere to the subter- ranean form of ceremonial chamber outweighed the inducement to place it within the village, or, in tlie case of the second kiva, even of placing it on the same levcd as the houses, which are 30 feet above it with an abrui>t trail between tliem. It is curious and instructive to see a room, the use of which is so intimately connected with th(Miiiier life of the village, jilaced in such acomparatively remote and inaccessilile position through an intensely conservative adherence to ancient |)ractice re(piii- iiig this <-haml)er to he de])ressed. The general view of the village given in I'l. XXXI strikingly illus- trates the blending of the rectangular foi'ms of the architecture with the angular and sliaiply defined fractuics of tln^ surrounding rock. This close correspondence in Ibrm lietween tlui architecture and its im- mediate surroundings is greatly heightentul by the similarity in color. Mr. Stcqilu^n has called attention to a similar effect on the western side of Waliii and its adjacent mesa edge, wliicli he thought indi(riites a dis- tinct etfort at concealment on the iiart of the builders, by lilendiiig the arcliitccture with tlie surroundings. Tliis similarity of effect is often accidental, and due tles of special and detailed corre- .spoudence as the cases cited, nor could it have any weight in developing a rectangular tyjie of architecture. In the development of primitive arts the advances are slow and laborious, and are prodiu-ed l)y adding snnill increments to current knowh'dge. So vague and undeftned an influence as that exerted by the larger forms of surrounding nature are seldom recognized and ac- knowledged by the artisan; on the contrary, experiments, resulting in improvement, are largely prompted by practical requirements. Par- ticidarly is this the case in the art of house-building. SHUMOPAVI. This village, although not so isolated as Oraibi, has no near neigh- bors and is little visited by whites or Indians. The inhabitants are rarely seen at the trading i>ost to which the others resort, and they seem to be pretty well off and independent as comj>ared with their neighbors of the other villages (PI. xxxiv). The houses and courts are in keeping with the general character of the people and exhibit a de- gree of neatness and thrift that contrasts sharply with the tumble-down appearance of some of the other villages, especially those of the ^Middle Mesa and Oraibi. There is a general air of ne«^H'ss about the place, though it is questionable whether the architecture is more recent than that of the other villages of Tusayan. This effect is partly due to the custom of frequently renewing the coating of mud x)laster. In most of the villages little care is talicn to repair the houses until the owner feels that to postpone sirch action longer would endanger its stability. Many of the illustrations in this chapter indicate the ijroportion of rough masonry usually exjiosed in the walls. At Shumoi)avi (PI. xxxv), however, most of the walls are smoothly plastered. In this resi)ect they resemble Zuni and the eastern pueblos, where but little naked masonry can be seen. Another feature that adds to the effect of neat- ness and fluish in this village is the frequent use of a whitewash of 74 PUEBLO ARCHITECTURE. gypsum ou the outer lace of the walls. This wash is used partly- as an ornament and j^artly as protection against the rain. The material, called by the Mexicans "yeso," is very comiiionly used in the interior of their houses throughout this region, both by Mexicans and Indians. More rarely it is used among the pueblos as an external wash. Here, however, its external use forms quite a distinctive feature of the vil- lage. The Siime custom in several of the cliff houses of Canyon de Clielly attests the comparative antiquity of the practice, though not necessarily its pre-Columbian origin. Shumopavi, compared with the other villages, shows less evidence of having been built on the open court idea, as the i)artial inclosures as- sume such elongated forms in the direction of the long, straight rows of the rooms; yet examination shows that the idea was present to a slight extent. At the southeast corner of the pueblo there is a very marked approach to the oi)en court, though it is quite evident that the easternmost row has its back to the court, and that the few rooms that face the other way are later additions. In fact, the plan of the village and the dis- tribution of the terraces seem to indicate that the first construction consisted only of a single row facing nearly east, and was not an in- closed court, and that a further addition to the pueblo assumed nearly the same form, with its face or terraced side toward the back of the first row only partly adapting itself by the addition of a few small rooms later, to the court arrangement, the same operation being con- tinued, bu^t in a form not so clearly defined, still farther toward the west. The second court is not defined on the west by such a distinct row as the others, and the smaller clusters that to some extent break the long, straight arrangement bring about an approximation to a court, though here again the terraces only partly fai^e it, the eastern side being bounded by the long exterior wall of the middle row, two and three stories high, and almost unbroken throughout its entire length of 400 feet. The broken character of the small western row, in conjunction with the clusters near it, imparts a distinct effect to the plan of this jKir- tion, differentiating it in character from the masses of houses formed by the other two rows. The latter aic connected at their southern end by a short cross row which converts this portion of the village practi- cally into a single large house. Two covered passageways, however, which are designated on tlie plan, give access to the southeast portion of the court. This portion is ]>artly separated from the north half of the inclosure by encroaching groups of rooms. This partial division of the original narrow and long court appears to be of later date. The kivas are fimr in nu7nber, of which l)ut one is within the village. Tlie latter occupies a partly inch)sed i)osition in the southwest i)ortion, and probably owes its place to some local facility for building a kiva on this spot in the nature of a depression in the mesa summit; but even / MINBELKFF. SHUMOPAVI. 75 with such aid the (•ciciiiouial chaiiibcr was built ouly partly uudcr ground, as may be seen in Fig. 14. The remaining three kivas are more distinctly subterrauean, and in order to obtain a suitable site one of these was located at a distance of more than 200 feet from the village, toward the mesa edge on the east. The other two are built very close together, apparently in contact, just beyond the northern extremity of the village. One of these is about 3 feet above tlie surface at one corner, but nearly on a level with the ground at its western side where it adjoins its neighbor. These two kivas are illustrated in PI. Lxxxviii and Fig. 21. Here again we tind that the ceremonial clianiber that forms so iiiipor- ?>j.awf Fig. 14. Cunrt kiva of Shniiioi)jivi. taut a feature among these people, occupies uo tixed relation to the dwellings, and its location is largely a matter of accident, a site that would admit of the partial excavation or sinking of the chamber below the surface being the main requisite. The northwest court contains another of the small inclosed shrines already described as occurring at Shupaulovi and elsewhere. The stonework of this village also possesses a somewhat distinctive character. Exposed masonry, though comparatively rare in tliis well- plastered pueblo, shows that stones of suitable fracture were selected and that they were more carefully laid than in the other villages. In places the masonry bears a close resemblance to some of the ancient work, where the spaces between the longer tablets of stone were care- fully chinked with small bits of stone, bringing the whole wall to a uniform face, and is much in advance of the ordinary slovenly methods of construction followed in Tusayan. Shumopavi is the successor t)f an older village of that name, one of the cities of the ancient Tusayan visited l)y a detachnumt of Ooronado's ex- pedition in 1540. The ruins of that village still exist, and they formerly contained vestiges of the old church and mission buildings established 76 PUEBLO ARCHITECTl'KE. by the monks. The sciuared beams IVoiii these buildings were cousidered vahiable enough to be incorporated in the construction of ceremonial kivas ill some of the Tusayan viDages. This old site was not visited by till' party. This is one of the largest modern pueblos, and contains nearly half the population of Tusayan yet its great size has not materially attccted ; the arrangement of the dwellings. The general ])lan (see PI. xxxvi) simply shows an unusually large collection of typical Tusayan liouse- rows, with the general tendency to face eastward displayed in the other \illages of the group. There is a remarkable uniformity in the direction of the rows, but there are no indications of the order in which the suc- cessive additions to the village were made, such as were found at Ma- shongnavi. The clusters of rooms do not surpass the average dimensions of those in the smaller villages. In five of the clusters in Oraibi a height of tour stories is reached by a few rooms; a height seen also in Wali)i. At several iioints in Oiaibi. notably on tlie west side of cluster No. 7, maybe seen what appears to lie low terraces faceil with rough iiiasoiiry. The same thing is also seen at Walpi, on the west side of the iiiiitlierii- most cluster. This effect is produced by the gradual tilhng in of aban- doned and broken-down marginal houses, with fiillen masonry and drifted sand. The appearance is that of intentional construction, as may be seen in PL xxxix. The rarity of covereil passageways in this village is noteworthy, an.' KJ -Xes' MINDELEFF.] MOEN-KOPI. 77 the other liaud, it was the necessity for defense that led to the close clustering- of the dwellings and the consequent employment of the cov- ered way. A further contrast between the general plans of Oraibi and Zuni is aft'orded in the different manner in which the roof o])eiiings have been emph)yed in the two cases. The plan of Zuiii, PI. lxxvi, shows great numbers of small openings, nearly all of which are intended exclusively for the admission of light, a few only being provided with ladders. In Oraibi, on the other hand, there are only seventeen roof openings above the first terrace, and of these not more than half are intended for the admission of light. The device is correspondingly rare in other villages of the group, particularly in those west of the first mesa. In Mashong- uavi the restricted use of the roof oi)enings is particularly noticeable; they all are of the same type as those used for access to first terrace rooms. There is but one roof opening in a second story. An examina- tion of the plan, PI. xxx, will show that in Shupaulovi but two such openings occur above the first terrace, and in the large village of Shumopavi, PI. xxxiv, only about eight. None of the smaller villages can be fairly compared with Zufii in the employment of this feature, but in Oraibi we should expect to find its use much more general, wei-e it not for the fact that the defensive site has taken the place of the close clustering of rooms seen in the exposed village of Zuni, and, in consequence, the devices for the admission of light still adhere to the more primitive arrangement (Pis. xl and xli). The highest type of pueblo construction, emljodied in the large com- munal fortress houses of the valleys, could have developed only as the builders learned to rely for protection more upon their architecture and less upon the sites occupied. So long as the sites furnished a large projjortion of the defensive efficiency of a village, the invention of tlie builders was not stimulated to substitute artificial for natural advan- tages. Change of location and consequent development must frecjuently have taken place owing to the extreme inconvenience of defensive sites to the sources of subsistence. The builders of large valley i)ueblos must frequently have been forced to resort hastily to defensive sites on finding that the valley towns were unfitted to withstand attack. This seems to have been tlie case with the Tusayan; but that the Zuni have adhered to their valley pueblo through great difficulties is clearly attested by the internal evidence of the architecture itself, even were other testimony altogether wanting. MOEN-KOPI. About 50 miles west from Oraibi is a small settlement used by a few families from Oraibi during the farming season, known as Moen-kopi. (PL XLiii). The present village is comparatively recent, but, as is the case with many others, it has been built over the remains of an older settlement. It is said to have been founded within the memory of — 78 PUEBLO ARCHITECTURE. SOUK' of the Mormon pioneers at the neighboring town of Tnba tUty, named after an ohl Oraibi chief, recently deceased. Tlie site wonld probably have attracted a ninch larger nnmber of settlers, had it not been so remote from the main pneblos of the pro- vince, as in many respects it far surpasses any of the present village sites. A large area of fertile soil can be conveniently irrigated from copious sjirings in the side of a small branch of the Moen-kopi wash. The vil- lage occupies a low, rounded knoll at the junction of this branch with the main wash, which on the opposite or southern side is quite precipi- tous. The gradual encroachments of the Mormons for the last twenty years hav(^ had some eft'ect in keeping the Tusayan from more fully utilizing the advantages of this site (PI. XLii). Moen-kopi is built in two irregular rows of one-story houses. There are also two detached single nxnns in the village onv of them built for a kiva, though apparently not in use at the time of our survey, and the other a small room with its principal do(n' facing an adjoining row. The arrangement is about the same that prevails in the other villages, the rows having distinct back walls of rude masonry. Rough stone work predominates also in the fronts of the houses, though it is occasionally brought to a fair degree of finish. Some adobe Work is incorporated in the masonry, and at one point a new and still unroofed room was seen liuilt of adobe bricks on a stone foundation about a foot high. There is but little adobe masonry, however, in Tusayan. Its use in this case is probably due to Mormon influence. Moen-kopi was the headquarters of a large business enterprise of the Mormons a inimber of years ago. They attempted to concentrate the product of the Navajo wool trade at this ])oint and to establish here a completely appointed woolen mill. Water was brotight from a series of reservoirs built in a small \alley several miles away, and was conducted to iv point on the Moen-kopi knoll, near the end of the south row of houses, where the ditch terminated in a solidly constructed box of masonry. From this in turn the water was delivered through a large pipe to a turbine wheel, which famished the motive power for the works. The ditch and masonry are shown on the ground i)lan of the village (PL XLiii). This mill was a large stone building, and no expense was spared in fitting it up with the most complete machinery. At the time of oiu' visit the whole establishment had been abandoned for some years and was rapidly going to decay. The frames had been torn from the win- dows, and both the floor of the building and the ground in its vicinity were strewn with fragments of expensive machinery, broken cog-wheels, shafts, etc. This building is shown in PI. XLV, and may serve as an illustration of the contrast between Tusayan masonry and modern stone- mason's work carried out with the same material. The comparison, however, is not entirely fair, as applied to the pueblo builders in gen- eral, as tht^ Tusayan mason is unusually caicless in his work. Many old examples are seen in which the finish of the walls compares very BUREAU OF ETHNOLOGY , - L sal, ^ „, - nnles to the south of Zuili, consisted of irregular groujis of densely clnstered cells, occupying the ])oint of a spur jirojecting from a low rounded hill. The houses are in .such a ruined couditiou that few separate rooms can be traced, and these are nuich obscured Ivy debris. This debris covers the entire area extending down the east slojie of the hill to the site of the church. The large amount of debris and the (•omi(arati\e thinness of such walls as are found suggest that the dwellings had been densely clustered, and carried to the height of several stories. Much of the space between the village on the hill and the site of the Spanish church on the plain at its foot is covered with masonry debris, part of which has slid down from above (PI. XLVI). 80 MINDELEFF.] KETCHIPAUAN. 81 The arriuigemciit sii};cj;vsts ;i laijio iniiicii);il court of iiTcjiular form. Tbe siirrouiidiiij;' clustei-.s arc very irrej^ularly disposed, the directions of the prevailing Hues of walls greatly varying in ditterent groups. There is a suggestion also of several smaller courts, as well as of alley- ways leading to the ])rincipal one. The church, built on the i)laiu helow at a distance of about 200 feet from thenuiin village, seems to have been surrounded by several groujis of rooms and iudosures of various sizes, (littering somewhat in character fi-om those within the village. These groups are sc^attered and open, and the small amount of dtibi'is leads to the conclusion that this ])ortion of the village was not more than a singh; story in height. (I'l. XLVll.) Tbe destruction of the village has been so comi)lete that no vestige of constructional details remains, with the excei)tion of a row of posts in a building near the church. The governor of Zuni stated that these posts were part of a projecting porch similar to those seen in connection with modern houses. (See Pis. lxxi, i.xxv. ) Suggestions of this feature are met with at other points on the plain, but they all occur within the newer portion of the village around the church. Some of the larger inclosures in this portion of the village were very lightly constructed, and cover large areas. They were ])robably used as corrals. Inclosures for this purpose occur at other pueblos traditionally ascribed to the same age. The church in this village was constructed of adobe bricks, without the introduction of any stonework. The bricks appear to have been molded with an unusual degree of care. The massive angles of the northwest, or altar end of the structure, have survived the stonework of the adjoining village and stand to-day 13 feet high. (PL XLViii.) KKTCHII'AUAN. The small village of Ketchipauan appears to have been arranged about two courts of unequal dimensions. It is difficult to determine, however, how nuich of the larger court, containing the stone church, is of later construction. (PI. XLix.) All the northwest portion of the village is now one large inclosure or corral, whose walls have apjiarently been built of the fallen masonry from the surrounding houses, leaving the central si)ace clear. This wall on the northeast side of the large inclosure apparently follows the jogs and angles of the original houses. This may have been the outer line of rooms, as traces of buildings occur for some distance within it. On the opposite side the wall is nearly contiiuious, the jogs being of slight pro- jection. Here some traces of dwellings occur outside of the wall in places to a depth of three rooms. The same thing occurs also at the north corner. The continuation of these lines suggests a rectangular court of (considerable size, bounded symmetrically by groups of com- partments averaging three rooms deep. (PI. L.) Several much smaller inclosures made in the same way occur in the village, but they apparently do not conform to the original courts. 8 ETH G 82 PUEBLO AKrHITKCTI'RE. At the jn-csi'iit tiiiif dwelling,'' rooms are trace;il(lc <>\i-v a jxirtioii of tlu" area sotitli and west of tlie elmreli. As sliowii on tlie plan. ui)rijjlit [)osts oceasionally oeciir. These appear to have heen iiieorjjorated into the original walls, but the latter are so ruined that this can not be stated ])ositively, as such posts have soinetimes been incorjKn'ated in niodein corral walls. In jilaces they suggest the balcony-like feature seen in modern liouses, as in Hawikuh, but in the east portion of tlie pueblo they are irregularly scattered about the rooms. A considerable area on the west side of tlie ruin is covered with loosely .scattered stones, att'oj-ding no suggestions of a ground plan. They do not seem sufficient in amount to be the remains of dwelling looms. The S])anish church in tliis ]iueb]o was built of stone, l)nt tin- walls were mucli more massive than those of the dwellings. The building is well preserved, most of the walls standing 8 or 10 feet higli, and in places 14 feet. This church was apparently built by Indian labor, as the walls everywhere show the '•hinking with small stones characteristic of the native work. In this village also, the massive iSpauish coustnic- tion has siirvived the dwelling houses. The ground ]i1an of the church sliows tliat tlie openings were splayed in the thickness of tlie walls, at an angle of about 45° . In the doorway, in the east end of the building, the greater width of the opening is on the inside, a rather unusual arrangement; in the window, on the north side, tliis arrangeuuMit is reversed, the splay being outwanl. On the south side are indications of a similar opening, but at the present time the wall is so broken out that no Mcll defined jamb can be traced, and it is imjiossible to determine wlietlier the sjilayed ojicning was used or not. The stones of the masonry are laid with extreme care at the an- gles and in the faces of these sjilays, producing a highly finished effect. The position of the beam-lioles (ui the inner face of the wall suggests that the fioor of the church had been raised somewhat above the ground, aiul that there may have been a cellar-like space under it. No beams are now found, however, and no i-emains of wood are seen in the "altar" end of the church. At the iiresent time there are low jiartl tions dividing the inclosed area into six rooms or cells. The Indians state that these were built at a late date to convert the church into a defense against the hostile Ajiache from tlie south. These partitions apparently formed no ])ait of tlie original design, yet it is difficult to see how they could Inn » served as a defense, unless they Avere intended to be roofed over and thus converted into com]iletely inclosed rooms. A stone of somewhat larger size than usual has been built into the soutli wall of the church. Upon its surface some native artist lias en- graved a rudely drawn mask. About 150 yards southeast from the church, and on the edge of the low mesa upon Avhicli the ruin stands, has been constructed a reservoir of large size which furnislied the jiucblo with a reserve water .suiiidy. The ordinary supply was probably deri\'ed from the valley below, where BUREAU OF ETHNOLOGY EIGHTH ANNUAL REPORT PL. XXXVII Parroouet YoungCobn Rabbit COYOTL Sand Badger Bow BurrowingOwl Heed tm tm KEY TO ORAIBI PLAN, ALSO SHOWING LOCALIZATION OF GENTES, AllNl»t,LtFF. CHALOWK HAMPASSAWAN. S3 water is toiiiid at no great distance from the pueblo. Springs may also liiive formerly existed near tlie village, but this reservoir, located where the drainage of a large area discharges, must have materially increased the water sujjply. The basin or depression is about 110 feet iu diame- ter and its present depth in the center is about 4 feet; but it Las un- (loul)tedly been tilled in by sediment since its abandonment. More than half of its circumference was originally walled in, but at the pres- ent time the old masonry is indicated only by an interrupted row of larg(i foundation stones and fallen masonry. Some large stones, appar- ently undisturbed portions of the mesa edge, have been incorporated into the inclosing masonry. The Indians stated tliat originally the bottom of this basin was lined with stones, but these statements could not l)e verified. Without excavation on the ui)per side, the basin faded imperceptibly into the rising gTound of the surrounding drainage. Other examples of these ba.sin reservoirs are met with iu this regiou. have survived until recently, for, although at the |)resent time the room is covered with a. roof of rudely split cedar beams, the remains of tlie old, carefully built roof lie scattered about in the corners of the lOom, under the dirt and debris. The ojieniiigs are very small and seem to have been modified since the original construction, but it is diflicult to distinguish between the older original structure aiid the more recent additions. K'IAKIMA. On the south side of the isolated mesa of Taaaiyalana and occupying a high rounded spur of foothills, is the ruined village of K'iakima (PI. Lii). A long gulch on the west side of the spur contains, for 300 or 400 yards, a small stream which is fed from springs near the ruined village. The entire surface of the hill is covered with scattered debris of fallen walls, which must at one time have formed a village of considerable size. Over most of this area the walls can not be traced ; the few rooms which can be distinctly outlined, occurring in a group on the highest part of the hill. Standing walls are here seen, but they are apparently receut, one room showing traces of a chimney (PL Liv). Some of tlie more distinct inclosures, built from fallen masonry of the old village, seem to have been intended for corrals. This is the case also with the remains found on the clitts to the north of the village, whose position is shown on the plan (PI. Liii). Here nearly all the scattered stones of the original one-story buildings, have been utilized for these large in- closures. It is quite possible that these smaller structures on the ledge of the mesa were built and occupied at a nuu'h later date than the prin- cipal village. PI. LIII illustrates a portion of the base of Taaaiyalana where these inclosures apjiear. , A striking featiire of this I'uin is the occurrence in the northeast cor- ner of the village of large upright slabs of stone. The largest of these is about 3 feet wide and stands r>k feet out of the ground. One of the slabs is of such symmetrical form that it suggests skillful artificial treatment, but the stone was used just as it came from a seam in the clift' above. From the same seam many slabs of nearly equal size and symmetrical form have fallen out and now lie scattered about on the talus below. Some are remarkable for tluMr perfectly rectangular form, while all are distinguished by a notable uniformity in thickness. Close by, and apparently forming part of the same group, are a number of stones imbedded in the ground with their upper edges exposed and placed at right angles to the faces of the vertical monuments. The taller slabs are said by the Indians to have been erected as a defense against the attacks of the Apache u])on this j)ueblo, but only a portion of the group could, from their position, have been of any use for this 86 rri'.Bi.o akciiitkctire. purpose. Tlic stones probably mark j^raves. Although thorough ex- cavation of the hard soil could not be undertaken, digging to the dejith of 18 inches revealed the same character of pottery fragments, ashes, etc., found in many of the pueblo graves. Mr. E. W. Nelson found identical remains in graves in the Kio San Francisco region which lie excavated in collectiug ])Ottery. .Comparatively little is known, liow- ever, of the burial practices of this region, so it would Ije difficult to decide whether this was an ordinary nu'thod of burial or not. This pueblo has been identified by Mr. Cushiiig, thiDugJi Zuni tra- dition, as the scene of the death of Estevanico. the negro who accom- panied the first Spani.sh expedition to Oiljola. JIATSAKI. Matsaki is situatetl on a foothill at the base of Taaaiyalana, near its northwestern extremity. This pueblo is in about the .same state of preservation as K'iakinni, no comijlete rooms being traceable over mo.st of the area. Traces of walls, where seen, are not uniform in direction, .suggesting irregular grouping of the village. At two points ou the ])Ian rooms partially bounded by standing walls are found. These ap- pear to owe their preservation to their occupation as outlooks over fields in the vicinity long after the destruction of the pueblo. One of the two rooms .shows only a few feet of rather rude masonry. The walls of the other room, in one corner, stand the height of a full story above the surrounding debris, a low room under it having been par- tially filled up with fallen masonry and earth. The well preserved inner corner of the exjjosed room shows lumps of clay adhering here and there to the walls, the remnants of an interior corner chimney. No ti'ace of the supports for a chimney hood, such as occur in the modern fireplaces, could be fomid. The form outlined against the wall by these slight renuiins indicates a rather rudely constructed feature which was added at a late date to the room and formed no part of its original construction. It was probably built while the room was used as a I'arniing outlook. As shown on the ground i>lan (PL LV), a small cluster of houses once stood at some little distance to the southwest of the main i)U(>bl() and was connected with the latter by a series of rooms. The intervening space may liave been a court. At the northern edge of the villager a inimitive shrine has been erected in recent times and is still in use. It is rudely constructed by sim])ly idling up stones to a height of 2i or 3 feet, in a rudely rectangular arrangement, with an opening on the east. This shrine, facing east, contains an uju'Ight slab of thin sandstone on which a rude sun-symbol has been engraved. The governor of Zufii, in exjilaining the purpose of this shrine, c(mii)ared its use to that of our own astronomical observatories, wliich he had .seen. I'INAWA. The ruins of the small pueblo of I'inawa occu])y a slight rise on the south side of the Zuni Kiver, a short distance west of Zuui. The road MIN'DRI.EFF.] I'lXAWA. 87 from Zuiii to OJo Oalieiitt' traverses the ruin. Over most of the area rooms can not be traced. One complete room, however, has heiMi jne- served and appears to t>(> still oeeupied during;- the cultivation of the neighboring "milpas." It is roofed over and iu good condition, though the general character of the masonry resembles the older work. On the plan (Fig. 10) it will be seen that tlie stones of the original masonry have been collected and built into a number of large inclosures, which have in turn been partly destroyed. The positions of the entrances to these inclosures can be traced by the absence of stones on the surface. The general outline of the corral-like inclosures appears to have fol- lowed comparatively well pieserved ])ortions of the original wall, as was the case at Ketclii])auan. (PI. i.vr.) Scaio O M 50 7S lOO i23rCrT Fir.. K). IMiiuwii. plan. On the. southwest side of the i>iiehl(), porticms 4»f' the outer wall are distinctly traceable, some ot* tlie stones beinjj;' still in position. This 88 PUEBLO ARCHITECTUKE. portion of till' oiitliiK' is (listiu^uisiu'd by acniious series of curves, re- sembling portions of Nutria and I'escado, liut intersecting in an un- usual manner. The OJo ('aliente road i)asses between the main ruin and the stand- ing room above described. The remnants of the fiillen masonry are so few antl so promiscuously scattered over this area tluit the continuity of remains can not l)e fully traced. An ancient pueblo called Halona is said to have belonged to the Cibolan grouj), and to have been inhabited at the time of the conquest. It occu- ])ie(l a jiortion of the site upon which the present ])ueblo of Zuni stands. A part of this pueblo was built on the opposite side of the river, where the r(>mains of walls were encountered at a slight depth below the surface of the ground in excavating for the foundations of Mr. Cushing's house. At that time only scattered remains of masonry were met with, and they furnished but little indication of details of plan or arrange- ment. Later—during the summer of 1888—Mr. Gushing made exten- sive additions to his house on the south side of the river, and in exca- vating for the foundations laid bai'c a number of small rooms. Excava- tion was continued until December of that year, when a large part of the ancient village had been exposed. PI. LVir, ft-om a photograph, illustrates a portion of these remains as seen from the southwest corner of Zuni. The view was taken in the morning during a light fall of snow which, lightly covering the tops of the walls left standing in the excavations, sharply defined their outlines against the shadows of the rooms. It seems impossible to restore the entire outline of the portion of Ha- lona that has served as a nucleus for modern Zuiii from such data as can be procured. At several points of the jiresent village, however, vestiges of the old pueblo can be identified. Doubtless if access could be ob- tained to all the innermost rooms of the pueblo some of them would show traces of ancient methods of construction sutticient, at least, to admit of a restoration of the general form of the ancient ])ueblo. At the time the village was surveyed such examination was not practica- ble. Tlie ])ortion of the old jiMcblo serving as a nucleus for later coii- structiou would probably be found under liouses Xos. 1 and 4, forming practically one mass of rooms. Strangers and outsiders are not ad- mitted to these innermost rooms. Outcrops Iti the small cluster l^o. 2 indicate by their position a continuous wall of the old jiueblo, probably the external one. Portions of the ancient outer wall are probably in- corjiorated into the west side of cluster Ko. 1. On the north side of cluster No. 2 (see PI. LXXVi) may be seen a buttress-like projection whose construction of small tabular stones strongly contrasts with the character of the surrounding walls, and indicates that it is a fragment of the ancient puebh). This projecting buttress answers no purpose whatever in its present position. MINDELEFF.] TA AA I V Al. ANA. 89 The above suggestious are coutiiuieil by another feature in the same house-cluster. On continuing the line of this buttress through the governor's house we fliid a projecting fragment of second story wall, the character and finish of which is clearly shown in PI. LViii. Its general similarity to ancient masonry and contrast with the present careless methods of construction are very noticeable. The height of this fragment above tlic ground suggests that the original pueblo was in a very good state of preservation when it was first utilized as a nucleus for later additions. That portion under house No. 1 is ])robably equally well preserved. The fi('(iucnt renovation of rooms by the ap- plication of a mud coating renders the task of determining the ancient portions of the cluster by the character of the masonry a very difficult one. Ceilings would probably longest retain the original appearance of the ancient rooms as they are not subjected to such renovation. Mr. Gushing thought that the outer western wall of the ancient pueblo was curved in outline. It is more probable, however, that it regulated the lines of the present outer rooms, and is reflected in them, as the usual practice of these builders was to put one partition directly over another in adding to the height of a building. Tlus would suggest a nearly rec- tangular form, perhaps witli Jogs and offsets, for the old builders could not incorx)orate a curved outer wall into a mass of rectangular cells, such as that seen in the present pueblo. On the other hand, the outer wall of the original i)ueblo may have been outside of rooms now occu- pied, for the villager had been abandoned for some time before the colony returned to the site. TAAAIYAI.ANA. On the abandonment of the pueblos known as the Seven Cities of Cibola, supposed to have occurred at the time of the general uprising of the pueblos in 1080, the inhabitants of all the Cibolan villages sought refuge on the summit of Taaaiyabina, an isolated mesa, 3 miles south- east from Zuiii, and there built a number of pueblo clusters. This mesa, otherwise known as "Thunder Mountain," rises to the height of 1,000 feet above the ])lain, and is almost iiiucccssible. There are two foot trails leading to the summit, each of which in places tra- verses abrupt slopes of sandstone where holes have been i)ecked into the rock to furnish foot ami hand holds. From tlu^ northeast side the summit of the mesa can be readied by a rough and tortuous burro trad. All the rest of the mesa rim is too precipitous to be scaled. Its appear- ance as seen from Zuiii is shown in PI. Lix. On the southern portion of this impregnable site and grouped about a point where nearly the whole drainage of the mesa top collects, are found the village remains. The Zufiis stated that the houses were dis- tributed in si.x: groups or clusters, each taking the place of one of the abandoned towns. Mr. Frank H. Cushing ' was also under the impres- ' Sen Millstono fi)r April, 1884, Iii(li:in;iii..lis, Iijilimia. . 90 ITKIiLO ARCHITECTrRK. siou that these houses had been hiiilt as six distiuct clusters of one vil- lage, aud he has found that at the time of the Puebh) rebellion, but six of the Cibolau villages were occupied. Au examination of the jilan, how- ever, will at once show that no such definite scheme of arrangement governed the builders. Thei-e are but three, or at most four groups that could be defined as distinct clusters, and even in the case of these the dis- position is so iiTegular aud their boundaries so ill defined, through the great number of outlying small groups scattered about, that they can hardly be considered distinct. There are really thirty-eight separate buildings ( PI. LX) ranging iu size from one of two rooms, near the soutlu'ru extremity to one of one liundred andtliree rooms, situated at the south- western corner of the whole group and close to the western edge of the mesa where the foot trails reach the summit. There is also great diver- sity in the arrangement of rooms. In some cases the clusters are quite compact, and iu others the rooms are distributed iu narrow rows. In the large cluster at the northwestern extremity the houses are arranged around a court; with this exception the clusters of rooms are scattered about iu an irregular manner, regardless of any defensive arrangenu'ut of the buildings. The builders evidently placed the greatest reliance on their impregnable site, and freely adopted such arrangement as con- venience dictated. The masonry of these villages was roughly constructed, the walls be- ing often less than a foot thick. Very little adobe mortar seems to have been used; some of the thickest and best preserved walls have appar- ently been laid nearly dry (PI. LXi). The few openings still preserved also show evidence of hasty and careless construction. Over most of the area the debris of the fallen walls is verly clearly marked, and is but little encumbered with earth or drifted sand. This imparts an odd effect of newness to tliese ruins, as though the walls had recently fallen The small amount of debris suggests that the majority of these buildings never were more than one story liigh, thimgli iu four of the broadest clusters (see plan, PI. LX) a lieight of two, and possibly three, stories may have been attained. All the ruins are thickly covered by a very luxurious growth of braided cactus, l>ut litth' of which is found else- where in the neighborhood. Tiu' extreme soutlu'asteru cluster, cousisl- ing of four large i-ooms, differs greatly in character from the rest of the ruins. Here the rooms or inclosures are defined only by a few stones on the surfa<-e of tiie ground and partly embedded in the soil. There is no trace of the debris of fallen walls. These outlined inclosures ajjpear never to have been walled to any consideralile height. Within one of the rooms is a slab of stcme, about which a few cerenumial plume sticks have been set on end within recent times. The motive tliat led tt) the occupation of this mesai was defense; the cause that led to the selection of the partit-ular site was facility for ])rocuring a water sup])ly. The trail on the west side passes a siiring half way down the mesa. There was another spring close to the foot M1NI>K1,EFF.1 KIN-TIEL. 91 trail on the south side; this, however, wns h)wei', l)eiiig iiiiiiost at the foot of the talus. In addition to these water sources, the builders collected and stored the drainage of the mesa summit near the southern gap or recess. At this point are still seen the remains of two reservoirs or dams built of hea\'y masonry. Only a few stones are now in place, but these indicate unusually massive construction. Another reservoir occurs farther along the mesa rim to the southeast, beyond the limits of the plan as given. As may be seen from the plan (PI. lx) the two reservoirs at the gap are cpiite close together. These receptacles have been much filled ui> with sediment. PI. LXii gives a view of the principal or western- most reservoir as seen from the northeast. On the left are the large stones once incorporated in the masonry of the dam. This masonry appears to have originally extended around three-fourths of the circum- ference of the reservoir. As at Ketchipauan, previously described, the upper portion of the basins merged insensibly into the general drainage and had no definite limit. The Zuni claim to have here practiced a curious method of water storage. They say that whenever there was snow on the ground the villagers would turn out in fin-ce and roll up huge snowballs, which were finally collected into these basins, the gradually melting snow furnishing a considerable quantity of water. The desert environment has taught these people to avail themselves of every expedient that could increase their supply of water. It is proper to state that in the illustrated plan of the Taaaiyalana ruius the mesa margin was sketched in without the aid of instrumental sights, and hence is not so accurately recorded as the plans and relative positions of the houses. It was all that could be done at the time, and will sufficiently illustrate the general relation ot the buildings to the surrounding topography. All the ruins above described bear close traditional and historic rela- tionship to Zuni. This is not the case with the splendidly preserved ancient pueblo of Kintiel, but the absence of such close historic con- nection is compensated for by its architectural interest. Differing rad- ically in its general plan from the ruins already examined, it still sug- gests that some resemblance to the nnn-e ancient portions of Nutria and Pescado, as will be seen by comparing the ground plans (Pis. lxvii and LXix). Its state of preservation is such that it throws light on details which have not survived the general destruction in the other ])ueblos. These features will be referred to in the discussion and com- ]>arison of these architectural groups by constructional details in Chap- ter IV. This pueblo, located nearly midway between Cibola and Tusayan, is given on some of the nuips as Pueblo Grande. It is situated on a small 92 PUEBLO ARCHITECTTTRE. arm of the Pueblo (Jolorado wash, 22 or 23 uiih^.s north ofNavajo Springs, and about the same distance south from Pueljlo Colorado (Ganado jiost- offlce). Geographically the ruins might belong to either Tusayau or Cibola, but Mr. Cushiug has collected traditional references among the Zuiii as to the occupation of this pueblo by related peoples at a time not far removed from tlie first Spanish visit to this region. The plan (J'l. LXiii) shows a marked contrast to the irregularity seen in the ruins previously described. The pueblo was clearly defined by a c(mtinuous anil unbroken outer wall, which prol)ably extended to the full height of the highest stories (PL lxiv). This symmetrical form is all the more remarkable in a pueblo of such large dimensions, as, with the excei)tion of Pueblo Ronito of the Chaco grou]i, it is the largest ancient pueblo examined by this Bureau. This village seems to belong to the same type as the Chaco examples, representing the highest de- velopment attained in building a large defensive pueblo practically as a single house. All the terraces faced upon one or more inclosed courts, thrcmgh which access was gained to the rooms. The openings in this outer wall, especially near the ground, were few in number and very small in size, as shown in PI. CiV. The pueblo was built in two \viugs of nearly equal size on the opposite slopes of a large sandy wash, trav- ersing its center from east to west. This wash doubtless at one time furnished peculiar facilities for storage of water within or near the vil- lage, and this nuist have been one of the inducements i'or the selection of the site. At the time of our survey, however, not a di'op of water was to be found about the ruin, nor could vestiges of any construction for gathering or storing water be traced. Such vestiges would not be likely to remain, as they must have been washed away by the violent summer torrents or buried under the accumulating sands. Two seasons subsequent to our work at this point it was learned that an American, digging in some rooms on the arroyo margin, discovered the remains of a well or reservoir, which he cleared of sand and di^bris and found to be in good condition, furnishing so steady a water supply that the dis- coverer settled on the sjjot. Tliis was not seen by the wi-iter. There is a small spring, perhaps a nule from the pueblo in a northeasterly direction, but this source would have been wholly insufficient for the needs of so large a village. It may have furnished a much more abun- dant supply, however, when it was in constant use, for at the time of imr visit it seemed to be choked up. About a mile and a half west ([nite a lagoon forms from the collected drainage of several broad valleys, and contains water for a long time after the cessation of the rains. About 6 miles to the north, in a depression of a broad valley, an extensive lake is situated, and its supply seems to be constant throughout the year, excei)t, perhai>s, during an unusually dry season. These vaiious Ijodies of water were undoubtedly utilized in the liorticulture of the occajjants of Kin-tiel; in fact, near the borders of the larger lake referred to is a small house of two rooms, much similar in workmanship to the maiu -^'Wfil fffl: h li 1^ ^> I f .A' I- '/lip ^-||^:Z^ MINDELEFF.l KIN-TIEL 93 pueblo, evidently designed as an outlook over fields. This building in illustrated in PI. Lxvi. The arrangement of the inner houses ditt'ers in the two halves of the ruin. It will be .seen that in the north half the general arrangement is roughly parallel with the outer walls, with tlus exce])tion of a small group near the east end of the arroyo. lu the soutli half, on the other hand, the inner rows are nearly at right angles to the outer room clus- ters. An examination of the eontonrs of tlu^ site will reveal the cause of this difference in the different configaration of the slopes iu the two cases. In the south half the rows of rooms Inive been built on two long projecting ridges, and the diverging small cluster in the nortii lalf owes its direction to a similar cause. The line of outer wall being once fixed as a defensive bulwark, there seems to have been but little restric- tion in the adjustment of the inner buildings U) conform to the irregu- larities of the site. (LM. lxiii.) Only three clearly deflned means of access to the interior of the pueblo could be found in the outer walls, and of these only two were suitable for general use. One was at a reentering angle of the outer wall. Just south of the east end of the arroyo, where the north wall, continued across the arroyo, overlaps the outer wall of the south half, and the other one was near the rounded northeastern corner of the pueblo. The third opening was a doorway of ordinary size in the thick north wall. It seems probable that other gateways once existed, especially in the south half. From its larger size and more compact arrangement this south half would seem to have greatly needed such fixcilities, but the preserved walls show no trace of them. The ground plan furnishes indications, mostly in the nortli half, of several large rooms of circular form, but broken down remains of scpiare rooms are so much like those of round ones in appearance, owing to the greater amount of debris that collects at the corners, that it could not be definitely determined that the ceremonial rooms here were of the circular form so counuon in the ancient pueblos. While only circular kivas have been found associated with ancient pueblos of this type, the kivas of all the Cibola ruins above described are said by the Zufiis to have been rectangular. The ((uestion can be decided for this pueblo (mly by excavation on a larger scale than the party was prepared to under- take. Slight excavation at a i)oiut where a round room was iiulicated on the surfiicc, revealed portions of straight walls only. The large size of the refuse heap on the south side of the village indi- cates that the site had been occupied for numy generatiims. Notwith- standing this long period of occupation, no important structure of the village seems to have extended beyond the i)lan. On the north side, outside the main wall, are seen several rectangles faintly outlined by stones, but these do not appear to have been rooms. They resemble similar inclosures seen in connection with ruined pueblos farther south, which proved on excavation to contain graves. 94 IMKHLO AliCHITKCTl'RE. The positions oC tlui few cxciivations iinidc arc imlicatcd oi: t'lc plan (PI. LXiii). Our facilities for such woik were most meager, and what- ever results were secured were reached at no };reat distance from the surface. One of these excavations, illustrated in PL c, will be described at greater length iu Chapter iv. PLANS AND DBSCEIPTIONS OF INHABITED VILLAUK8. Nutria is the smallest of the three farming puel)los of Zuui, and is located about 2.i miles by trail northeast from Zuui at the head of Nutria valley. The water supply at this point is abundant, and fur- uishes a running stream largely utilized in irrigating fields in the vicinity. Most of the village is comi)actly arranged, as may be seen from the l»lan (PI. Lxvii and Fig. 17), l)ut a few small clusters of late construc- tion, containing two or three rooms eacli, are situati'd toward the east at quite a di.stance from the principal group. It is now occupied solely as a farming pueblo during the planting and harvesting season. The outline of this small pueblo differs greatly from those of most of the Cibolan villages. The village (PL lxviii), particularly in its north- ernmost cluster, somewhat approximates the form of the ancient pueblo of Kiii-ticl (PI. Lxm), and has ai)parently been built on the remains of an older village of somewhat corresponding form, as indicated by its curved outer wall. Fragments of carefully constructed masonry of the ancient t>i>e, contrasting noticeably with the surrounding modern con struction, afford additional evidence of this. The ancient village must have been pi'ovided originally with ceremonial rooms or kivas, but no traces of such rooms are now to be found. nT ^^*^ONflv Vm. 17. Ntirriii. plan; small (lia^ratu. old wall. At the close of the harvest, when the season of feasts and ccreinonials begins, lasting through most of the winter, the occupants of these farm- BUREAU OF ETHNOLOGY EIGHTH ANNUAL REPORT PL. XLIII MOEN-KOPI. PESCADO. 95 iiij,' villages close, up their houses and move back to the luaiu i)uebl(> leaving them uuteiiaiiteil until the succeeding spring. The great number of abandoned and ruined rooms is very noticeable in the farming pueblos illustrated in this and two of the succeeding jdans (Pis. LXix and Lxxiii). The families that farm in their vicinity seem to occupy scarcely more than half of the available rooms. This village, also a Zuhi farming pueblo, is situated in a large valley, about 12 miles northeast from Zuiii. Although it is much larger than Nutria it is wholly comprised within the c(nniiact group illustrated. The tendency to build small detached houses noticed at Nutria and at Ojo Caliente has not manifested itself here. The prevalence of aban- doned and roofless houses is also noticeable. .^',«»' *^#-'::?33'=^ FiG. 18. PescaUo, piau, old wall diagram. The outlines of the original court inclosing pueblo (PI. Lxx) are very clearly marked, as the farming Zuiiis in their use of this site liave scarcely gone outside of the original limits of the ancient pueblo. The plan, PI. LXIX and Fig. 18, shows a small irregular row built in the large inclosed court; this row, with the inclosures and corrals that surround it, probably formed no part of the original plan. The full curved out- line is broken only at the west end of the village by small additions to the outer wall, and the north and east walls also closely follow the boundary of the original jiueblo. In fact, at two points along the north wall fragments of carefully executed masonry, jnobably forming part of the external wall of the ancient pueblo, are still preserved (PI. lxxii). This outer wall was probably once continuous to the full height of the 96 PUEBLO ARCHITECTURE. puchid, but the partial restorations of the buildings by the Zufii fanners reseiiil)l(' more closely the modern arrangement. Small rooms have been added to the outside of the eluster and in some cases the terraces are reached by external stone steps, in contrast with the defensive arranj^ement pvevailinjjf fjciicrally in ])U('bl()s of tliis form. A number of dome-shaped ovens have been built outside the walls. The principle of pueblo plan embodied in Kin-tiel, before referred to, is tra(^eable in this village with jwirticular clearness, distinguishing it from most of the Cibolan pueblos. No traces of kivas w.ere met with in this village. OJI) CALIEXTE. The farming village of Qjo Caliente is located near the dry wash of the Zuiii lliver, and is about 15 miles distant from Zuni, in a southerly direction. It is about nudway between Hawikuh and Ketchi]iauan, two of the seven cities of (Jibola above described. Though situated in fer- tile and well watered country and close to the remains of the ancient villages, it bears indications of having been, built in comparatively recent times. There are no such evidences of connection with an older village as were found at Nutria and Pescado. The irregular and small clusters that form this village are widely scattered over a rather rough and broken site, as shown on the plan (PI. lxxiii). Here again a large portion of the village is untenanted. The large cluster toward the eastern extremity of the group, and the adjoinin;' houses situated on the low, level ground, compose the present inhabited village. The houses occupying the elevated rocky sites to the west (PI. lxxiv) are in an advanced stage of decay, and have been for a long time abandoned. This southern portion of the Cibola district seems to have been much exposed to the inroads of the Apache. One of the eftects of this has already been noticed in the defensive arrangement in the Ketchipauan church. On account of such danger, the Zuni were likely to have built the first house-clusters here on the highest points of the rocky promon- tory, notwithstanding the comparative inconvenience of such sites. Later, as the fiirmers gained confidence or as times became safer, they built houses down on the fiat now occui)ied ; but this apparently was not done all at once. The distribution of the houses over sites of vary- ing degrees of inaccessibility, suggests a succession of approaches to the occupation of the open and uii])rotected valley. Some of the masonry of this village is carelessly constructed, and, as in the other farming pueblos, there is much less adobe plastering and smoothing of outer walls than in the home pueblo. At the time of the survey the occui)iition of this village throughout the year was proposed by several families, who wished to resort to the parent village only at stated ceremonials and important festivals. The comi»arative security of recent times is thus tending to the disintegra- tion of the huge central pueblo. This result must be inevitable, as the 1^ *. MLN-DELEFF.] ZUNI. 97 dying out of tlic defeusive motive brings about a realizatiou of tlie great iucouvenience of the present centralized system. ZUSl. Tlie pueblo of Zuiii is built upon a small knoll on the north bank of the Zuili River, about three miles west of the conspicuous mesa of Taa- aiyalaua. It is the successor of all the original " Seven Cities of Cibola" of the Spaniards, and is the largest of the modern pueblos. As before stated, the remains of Halona, one of the "seven cities," as identified by Mr. Gushing, have served as a nucleus for the construction of the modern pueblo, and have been incorporated into the most densely clus- tered portions, represented on the plan (PI. lxxvi) by numbers 1 and 4. Some of the Cibolan villages were valley pueblos, built at a distance ti-om the rocky mesas and canyons that must have served as quarries for the stone used in building. The Halona site was of this tyi^e, the nearest supply of stone being 3 miles distant. At this point (Halona) the Zuiii River is perennial, and furnishes a plentiful supply of water at all seasons of the year. It disappears, however, a few miles west in a broad, sandy wash, to appear again 20 miles below the village, prob- ably through the accession of small streams from springs farther down. The so-called river furnishes the sole water supply at ZuQi, with the exception of a single well or reservoir on the north side of the village. Zuiii has been built at a point having no s])ecial advantages for de- fense; convenience to large areas of tillable soil has apparently led to the selection of the site. This has subjected it in part to the same influences that had at an earlier date produced the carefully walled fortress pueblos of the valleys, where the defensive efticiency was due to well planned and constructed buildings. The result is that Zuiii, while not comparable in symmetry to many of the ancient examples, displays a remarkably compact arrangement of dwellings in the por- tions of the pueblos tirst occupied, designated on the plan (PI. lxxvi) as houses 1 and 4. Owing to this restriction of lateral expansion this portion of the pueblo has been carried to a great height. PI. Lxxvm gives a general view of these higher terraces of the village from the southeast. A height of five distinct terraces from the ground is attained on the south side of this cluster. The same point, however, owing to the irregularity of the site, is only three terraces above the ground «^n the north side. The summit of the knoll upon which the older portion of Zuni has been built is so uneven, and the houses them- selves vary so much in dimensions, that the greatest disparity i)revails in the height of terraces. A three-terrace portion of a cluster may have but two terraces immediately alongside, and throughout the more closely built portions of the village the exposed height of terraces varies from 1 foot to 8 or 10 feet. PI. Lxxix illustrates this feature. The growth of the vilhige has apparently been far beyond the origi- nal expectation of the builders, and the crowded additions seem to have 8 ETH 7 98 PUEBLO ARCHITECTURE. been joined to the clusters wlierever the demand for more space was most urgent, witlumt following any delinite x'lau in their arrangement. In such of the ancient puebh) ruins as afford evidence of having passed through a similar experience, the crowding of additional cells seems to have been made to conforin to some extent to a predetermined plan. At Kin-tiel we have seen how such additions to the number of habitable rooms could readily be made within the open court without ali'ectiug the symmetry and defensive efficieucy of the pueblo; but here the niu'lcus of the large clusters was small and compact, s-o that (Milarge- mcnt has taken place only by the addition of rooms on the outside, both on the ground and on upper terraces. The highestpoint of Zufii, now showing five terraces, is said to havehad a height of seven terraces as late as the middle of the present century, but at the time of the survey of the village no traces were seen of such additional stories. The top of the ])resent fifth terrace, however, is more than oO feet h)ng, and aflbrds sufficient space for the addition of a sixth and seventh story. The court or plaza in which the church (PI. Lxxx) stands is so much lai'ger than such inclosures usually are when incori)orated in a pueblo plan that it seems unlikely to have formed part of the original village. It probably resulted from locating the church prior to the construction of the eastern rows of the village. Certain features in the houses them- selves indicate the later date of these rows. The arrangement ofdwelliugs about a court (PI. LXXXii), characteristic of the ancient pueblos, is likely to have prevailed in the small pueblo of Ilalona, about which clustered the many irregular houses that consti- tute modern Zufii. Occasional traces of such au arrangement are still met with in portions of Zuni, although nearly all of the ancient pueblo has been covered with rooms of later date. In the arrangement of Zuiii himses a noticeable difierence in the manner of clustering is found in different parts of the pueblo. That portion designated as house 'So. 1 on tlie plan, built over tlie remains of the original small ]iueblo, is un- (luestionably the oldest portiou of the village. The clustering seems to have gone on around this center to an extraordinary and excejitioual extent before any houses were built in other portions. House No. 4 is a portion of the same structure, for although a. street or passageway intervenes it is covered with two or three terraces, indicating that such connection was established at an early date. The rows on the lower ground to the east (PI. lkxxi), where the rooms are not so densely clustered, were built after the I'emoval of the defensive motive that in- fluenced the construction of the central pile. These portions, arranged approximately in rows, show a marked resemblance to pueblos of known recent date. That they were built subseiiuently to the main clusters is also indicated by the abundant use of obliijue openings and roof holes, where there is very little necessity for such contrivances. This feature was originally devised to meet the exceiitional conditions of lighting MINDELEFF.] ZUNI. 99 imposed by dense crowding of the living rooms. It will be referred to again in examining the details of openings, and its wide dt'partnre from the arrangement found to prevail generally in pueblo eonstruetions will there be noted. The habit of making such provisions for lighting inner rooms became fixed and was applied generally to many clusters much smaller in size than those of other pueblos where this feature was not developed and where the necessity for it was not felt. These less crowded rooms of more recent construction form the eastern portion of the pueblo, and also include the governor's house ou the south side. The old ceremonial rooms or kivas, and the rooms for the meeting of the various orders or secret societies were, during the Spanish occu- pancy, crowded into the innermost recesses of this ancient portion of Zuni under house No. 1. But the kivas, in all likelihood, occupied a more marginal position before such foreign influence was brought to bear ou them, as do some of tiie kivas at the present time, and as is the general practice in other modern pueblos. CHAPTER IV. ARCHITECTURE OF TUSAYAN AND CIBOLA COMPARED BY CON- STRUCTIONAL DETAILS. INTRODUCTION. In the two preceding chapters the more general features of form and distribution in the ruined and inhabited pueblos of Tusayan and Cibola have been described. In order to gain a full and definite idea of the architectural acqnirements of the pueblo Ituilders it will be necessary to examine closely the constructional details of their present houses, endeavoring, when practicable, to compare these details with the rather meager vestiges of similar features that have survived the destruction of the older villages, noting the extent to which these have departed from early types, and, where practicable, tracing the causes of such deviation. For convenience of comparison the various details of house- building for the two groups will be treated together. The writer is indebted to Mr. A. M. Stephen, the collector of the tra- ditionary data already given, for information concerning the rites con- nected with house building at Tusayan incorporated in the following pages, and also for the carefully collected and valuable nomenclature of architectural details appended hereto. Material of this class per- taining to the Cibola group of pueblos unfortunately could not be pro- cured. HOUSE BUILDING. lUTKS AND METHODS. The ceremonials connected with house building in Tusayan are quite meager, but the various ste])s in the ritual, described in their proper connection in the following paragrajihs, are well defined and definitely jassigned to those who participate in the construction of the buildings. So far as could be ascertained there is no prearranged i)lan for an entire house of several stories, or for the arrangement of contiguous houses. Most of the ruins examined emphasize this absence of a clearly defined general plan governing the location of rooms added to the original cluster. Two notable exceptions to this want of definite j)lan occur among the ruins described. In Tusayan the Fire House (Fig. 7) is evidently the lesult of a clearly defined purpose to give a definite form to the entire cluster, just as, on a very much larger scale, does the ruin of Kin-tiel, belonging to the Cibola group (PI. lxiii). In both these cases the fixing of the t)uter wall on a definite line seems to have 100 BUREAU or ETHNOLOGY EIGHTH ANNUAL REPORT PL. XLVl MWDELEPP.] HOUSE BUILDING. 101 beeu regarded as of more importance than tbe specific locations of in- dividual rooms or dwellings within this outline. Throughout that part of Tusayan which has been examined, however, the single room seems now to be regarded as the pueblo unit, and is spoken of as a complete house. It is the construction of such a house unit that is here to be described. A suitable site having been selected, the builder considers what the dimensions of the house should be, and these he measures by paces, placing a stone or other mark at each corner. He then goes to the woods and cuts a sufficient number of timbers for the roof of a length corresponding to the width of his house. Stones are also gathered and roughly dressed, and in all these operations he is assisted by his friends, usually of his own geus. These assistants receive no compensation except their food, but that of itself entails considerable expense on the builder, and causes him to build his house with as few helpers as possible. The material having been accumulated, the builder goes to the village chief, who prepares for him four small eagle feathers. The chief ties a short cotton string to the stem of each, sprinkles them with votive meal, and breathes upon them his prayers for the welfare of the pro- posed house and its occupants. These feathers are called Nakwa kwoci, a term meaning a breathed prayer, and the prayers are addressed to M4sauwu, the Sun, and to other deities concerned in house-life. These feathers are placed at the four corners of the house and a large stone is laid over each of them. The buUder then decides where the door is to be located, and marks the place by setting some food on each side of it ; he then passes around the site ti'om right to left, spriukling piki crumbs and other particles of food, mixed with native tobacco, along the lines to be occupied by the walls. As he sprinkles this offer- ing he sings to the Sun his Kitdauwi, house song: "Si-ai, a-hai, si-ai, a-hai." The meaning of these words the people have now forgotten. Mr. Stephen has beeu informed by the Indians that the man is a ma- son and the woman the plasterer, the house belonging to the woman when finished; but according to my own observation this is not the universal practice in modern Tusayan. In the case of the house in Oraibi, illustrated in PI. xl from a photograph, much, if not all, of the masonry was laid, as well as finished and plastered, by the woman of the house and her female relatives. There was but one man ]jreseut at this house-building, whose grudgingly performed duty consisted of lifting the larger roof beams and lintels into place and of giving occa- sional assistance in the heavier work. The grcmnd about this house was strewn with quantities of broken stone for masonry, which seemed to be all iirepared and brought to the spot before building began ; but often the various divisions of the work are carried on by both men and women simultaneously. Wliilo the men were dressing the stones, the women brought earth and water and mixed a miul plastei'. Then the walls were laid in irregular courses, using the mortar very sparingly. 102 PUEBLO ARCHITECTURE. TUe house is always built iu the form of a parallelogram, the walls being from 7 to 8 feet high, and of irregular thickness, sometimes vary- ing from 15 to 22 inches iu different jiarts of the same wall. Pine, piilou, juniper, cottonwood, willow, and indeed all the available trees of the region are used in house construction. The main beams for the roof are usually of pine or cottonwood, from which the bark has been stripped. The roof is always made nearly level, and the ends of the beams are placed across the side walls at intervals of about 2 feet. Above these are laid smaller poles jjarallel with the side.walls, and not more than a foot apart. Across these again are laid reeds or small willows, as close together as they can be placed, and above this series is crossed a layer of grass or small twigs and weeds. Over this frame- work a layer of mud is spread, which, after drying, is covered with earth and firmly trodden down. The making of the roof is the work of the women. When it is finished the women ijroceed to spread a thick coating of mud for a floor. After this follows the application of plaster to the walls. Formerly a custom prevailed of leaving a small space on the wall nuplastered, a belief then existing that a certain Kat^hiua came and finished it, and although the space remained bare it was con- sidered to be covered with an invisible plaster. The house being thus far completed, the builder prepares four feath- ers similar to those prepared by the chief, and ties them to a short piece of willow, the end of which is inserted over one of the central roof beams. These feathers are renewed every year at the feast of Soyal- yina, celebrated in December, when the sun begins to return north ward. The builder also makes an offering to M4sauwu (called "feed- ing the house") by x>lacing fragments of food among the rafters, be- seeching him not to hasten the departure of any of the family to the under world. A hole is left in one corner of the roof, and under this the woman builds a fireplace and chimney. The former is usually but a small cavity about a foot square iu the corner of the floor. Over this a chim- ney hood is constructed, its lower rim being about 3 feet above the floor. As a rule the house has no eaves, the roof being finished with a stone coi)ing laid flush with the wall and standing a few inches higher than the roof to preserve the earth covering from being blown or washed away. Roof-drains of various materials are also commonly inserted in the cojiings, as will be described later. All the natives, as far as could be ascertained, regard this single- roomed house as being complete in itself, but they also consider it the nucleus of the larger structure. When more space is desired, as when the daughters of the house marry and rt^quire room for themselves, another house is built in front of and adjoining the first one, and a sec- ond story is often added to tlie original house. The same ceremony is observed in building the ground story iu front, but there is no cere- mony for the second and additional stories. MINDELEFF.] HOUSE BUILDING. 103 Auawitii (vyar-cliief of Sicbuiaovi) ilescribess the house in Walpi in which he was bom as having IukI five rooms on the ground floor, and as being four stories high, but it was terraced both in front and rear, his sisters and their families occupying the rear portion. Tlie fourth story consisted of a single room and had teri;K'es on two opijosite sides. This old house is now very dilapidated, and the greater portion of the walls have been carried away. There is no prescribed position for com- municating doorways, but the outer doors are usually placed iu the lee walls to avoid the prevailing southwest winds. Formerly on the approach of cold weather, and to some extent the custom still exists, people withdrew from the upper stories to the ki- koli rooms, where they huddled together to keep warm. Economy in the consumption of fuel also prompted this ex]iedieut ; but these ground- floor rooms forming the first terrace, as a rule having no external door- ways, and entered from without by means of a roof hatchway provided with a ladder, are ordinarily used only for purposes of storage. Even their roofs are largely utilized for the temporary .storage of many house- hold articles, and in the autumn, after the harvests have been gathered, the terraces and copings are often covered with drying peaches, and the peculiar loug strips into which pumjikins and squashes have been cut to facilitate their desiccation for winter use. Among other things the household supply of wood is sometimes piled up at one end of this ter- race, but more commonly the natives have so many other uses for this space that the sticks of fuel are piled up on a rude projecting skeleton of poles, supported on one side by two upright forked sticks set into the grounil, and ed on one side, a type of opening which is quite common in Tusayau. This form is illustrated in Fig. 84. This rttom has three windows, all of very small size, but it has no interior communication with any other room. In this respect it is ex- cej)tional. Ordinarily rooms communicate with others of the cluster. PI. Lxxxv shows another tyi)ical Tusayan interior in perspective. It illustrates essentially the same arrangement as does the preceding ex- anijile. The room is much larger than the one above described, and it 110 PUEBLO ARCHITECTURE. is divided midway of its length by a similar buttress. This buttress supports a heavy girder, thus admitting of the use of two tiers of lioor beams to span the whole length of the room. The firej)lace and ehiui- ney are similar to those described, as is also the single compartment for mealing stones. In this ease, however, this ])ortion of the room is quite large, and the row of mealing stones is built at right angles to its back wall and not parallel with it. The right-hand ])oition of the room is provided with a long, straight pole suspended from tlie roof beams. This is a common feature in both Tusayan and Zuui. The pole is used for the suspension of the house- hold stock of blankets and other garments. The windows of this house are small, and two of them, in the right-hand division of the room, have been roughly sealed up with masonry. PI. LXXXVI illustrates a typical Zuni interior. In this instance the example happens to be rather larger than the average room. It will be noticed that this apartment has many features in common with that at Tusayan last described. The pole upon which blankets are sus- pended is here incorporated into the original construction of the house, its two ends being deeply embedded in the masonry of the wall. The entire floor is paved with slabs of much more regular form than any used at Tusayan. The Zurii have access to building stone which is of a much better grade than is availalde in Tusayan. This room is furnished with long, raised benches of masonry along the sides, a feature much more common at Zuni than at Tusayan. Usually such benches extend along the whole length of a wall, but here the pi'o- jection is interrupted on one side by the fireplace and chimney, and on the left it terminates abrui>tly near the beginning of a tier of mealing stones, in order to afford floor space for the women who grind. The metates are arranged in the usual manner, three in a row, but there is an additional detached section placed at right angles to the main series. The sill of the doorway by which this room communicates with an ad- joining one is raised about 18 inches above the floor, and is provided with a rudely mortised door in a single panel. Alongside is a small hole through which the occupant can prop the door on the inside of the communicating room. The subsequent sealing of the small hand-hole with mud effectually closes the house against intrusion. The unusual height of this door sill from the floor has necessitated the construction of a small step, which is built of masonry and covered with a single slab of stone. All the doors of Zuiii are more or less raised above the ground or floor, though seldom to the extent shown in the present example. This room has no external door and can be directly entered only by means of the hatchway and ladder shown in the drawing. At one time this room was probably bounded by outer walls and was i)ro- vided with both door and windows, though now no evidence of the door remains, and the windows have become niches in the wall utilized for the reception of the small odds and ends of a Zuiii household. The ir MIXDELEFF.l KIVAS IN TUSAYAN. Ill chimney of this house will be noticed as differing matei'ially, both iu form and iu its position in the room, from the Tusayan examples. This form is, however, the most common type of chimney used in Zuiii at the present time, although many examples of the curved type also occur. It is built about midway of the long wall of the room. The Tusayan chimneys seldom occupy such a position, but are nearly always built in corners. The use of a pier or buttress-i^rojectiou for the support of a roof girder that is characteristic of Tnsayan is not practiced at Zuiii to any extent. Deer horns have been built into the wall of the room to answer the purpose of pegs, upon which various household articles are suspended. The various features, whose positions iu the pueblo dwelling house have been briefly described above, will each be made the subject of more exhaustive study in tracing the various modifications of form through which they have passed. The above outline will furuisli a general idea of the place that these details occupy in the house itself. KIVAS IN TUSAYAN. General use of kivas.—Wherever the remains of iiueblo architecture occur among the plateaus of the southwest there appears in every im- portant village throughout all changes of form, due to variations of environment and other causes, the evidence of chambers of exceptional character. The chambers are distinguishable from the typical dwelling rooms by their size and position, and, generally, iu ancient examples, by their circular form. This feature of i)ueblo architecture has survi^'ed to the present time, and is prominent in all modern pueblos that have come under the writer's notice, including the villages of Acoma and Jeniez, belonging to the Eio Grande gnmp, as well as in the pueblos under discussion. In all the i)ueblos that have been examined, both ancient and modern, with the exception of those of Tusayan, these sjiecial rooms, used for ceremonial purposes, occupy marginal or semi- detached positions in the house clusters. The latter are wholly de- tached from the houses, as may be seen from the ground iilans. Origin of the name.—Such ceremonial rooms are known usually by the Sj^auish term " estiifa," meaning literally a stove, and here used iu the sense of " sweat house," but the term is misleading, as it more prop- erly describes the small sweat houses that are used ceremonially by lodge-building Indians, such as the Navajo. At the suggestion of Major Powell the Tusayan word for this everpresent feature of ijueblo archi- tecture has been adopted, as being much more appropriate. The word " kiva," then, will be understood to designate the ceremonial (chamber of the pueblo building peoples, ancient and modern. Antiquity of the kiva.—The Avidespread occurrence of this feature and its evident antiquity distinguish it as being especially worthy of ex- haustive study, especially as embodied in its construction maybe found survivals of early methods of arrangement that have long ago become 112 PUEBLO ARCHITECTURE. extiiu-t ill tlic constantly imijroviiig art of housebuilding, luit wliicli art' preserved through the well kno'mi tendency of the survival of ancient practice in matters pertaining to the religious observances of a primitive people. Unfortunately, in the past the Zufii have been ex- posed to the repressive policy of the Spanish authorities, and this has probably seriously afi'ected the purity of the kiva tyi)e. At one time, when the ceremonial observances of the Zuui took jilace in secret for fear of incurring the wrath of the Spanish priests, the original kivas must have been wliolly abandoned, and though at the i>resent time some of the kivas of Zuiii occupy marginal positions in the cell clus- ters, just as in many ancient examples, it is doubtful Avhether these rooms faitlifully represent the original type of kiva. There seems to be but little structural evidence to distinguish the ijresent kivas from ordinary large Zuui rooms beyond the sjiecial character of the fireplace and of the entrance tiap door, features which will be fuUy described later. At Tusayan, on the other hand, we find a distinct and charac- teristic structural i>lan of the kiva, as well as many si)ecial constructive devices. Altliough the position of the ceremonial room is here excep- tional in its eiitire separation from the dwelling, this is due to clearly traceable influences in the immediate orographic environment, and the wholly subterranean arrangement of most of the kivas in this group is also due to the same local causes. Excavation of the kiea.—The tendency to depress or partly excavate the ceremonial chamber existed in Zuui, as in all the ancient pueblo buildings which have been examined; but the solid rock of the mesa tops iu Tusayan did not admit of the necessary excavati(ui, and the persistence of this requirement, which, as I shall elsewhere show, has an important connection with the early tyijes of pueblo building, compelled the occupants of these rocky sites to locate their kivas at points where depressions already existed. Such facilities were most abundant near the margins of the mesas, where in many places large blocks of sandstone have fallen out from the edge of the surface stra- tum, leaving nearly rectangular spaces at the summit of the clift' wall. The construction of their villages on these rocky promontories forced the Tusayan builders to sacrifice, to a large extent, the tradi- tional and customary arrangement of the kivas within the house- inclosed courts of the pueblo, in order to obtain properly depressed .sites. This accidental effect of the immediate environment resulted in giving unusual [irominence to the sinking of the ceremonial room below the ground surface, but a certain amount of excavation is found as a constant accompaniment of this feature throughout the pueblo region in both ancient and modern villages. Even at Zuiii, where the ki\'as appear to retain but few of the si)ecialized features that distinguish them at Tusayan, the floors are found to be below the general level of the ground. But at Tusayan the development of this single require- meut has been carried to such an extent that iuau>- of the kivas are BUREAU OP ETHNOLOGY ScEiIe, > 50 100 ooFeet. I I I I 1 I I I I I I I I I I I I I I I M I 1 I |-rrrl EIGHTH ANNUAL REPORT PL. Lll '1'***^ 'uuiuAt.a "UKIU^. MiNDELEFF] KIVAS IN TUSAYAN. 113 wholly subterraueaii. This is partii-ulaily the case with those that occupy luaig'iiial sites on the mesas, such as have been referred to above. lu such iustauces the broken-out recesses in the upper rocks have been walled up on the outside, roughly lined with uiasoury within, and roofed over in the usual niauiier. In many cases the depth of these rock niches is such that the kiva roof wheu finished does not project above the general level of the mesa summit, and its earth cov- ering is indistinguishable from the adjoining surface, except for the presence of the box-like projection of masonry that surrounds the en- trance trap door and its ladder (see PI. lxxxvii). Frequently in such cases the surface of the ground shows no evidence of the outlines or dimeusicms of the underlying room. Examples of such subterranean kivas may be seen in the foreground of the general view of a court in Oraibi (PI. xxxviii), and in the view of the dance rock at Walpi (PI. XXIV). But such wholly subterranean arrangement of the ceremonial chamber is by no means universal even at Tusayau. Even when the kiva was placed within the village courts or close to the houses, in con- formity to the tratlitional plan and ancient practice as evidenced in the ruins, naturally depressed sites were still sought; but such sites as the mesa margin aftbrds were rarely available at any distance from the rocky rim. The result is that most of the court kivas are only partly depressed. This is particularly noticeable in a court kiva in Shumo- pavi, an illustration of which is given in Fig. 14. The numgkiva or principal kiva of Shupaulovi, illustrated in PI. xxxiii, is scarcely a foot above the ground level on the side towards the houses, but its rough walls are exposed to a height of several feet down on the declivity of the knoll. The view of the stone corrals of Mashongnavi, shown in PI. CIX, also illustrates a kiva of the type de- scribed. This chamber is constructed on a sharp slope of the declivity where a natural depression favored the builders. On the upper side the roof is even with the ground, but on its outer or southern side the masonry is exposed to nearly the whole depth of the chamber. At the nortli end of Shumopavi, just outside the houses, are two kivas, one of which is of the semi-subterranean type. The other shows scarcely any masonry above the ground outside of the box-like entrance way. PL Lxxxviii illustrates these two kivas as seen from the northeast, and shows their relation to the adjacent houses. The following (Fig. 21) illusti'ates the same group from the opposite point of view. Access.—The last described semi-subterranean kiva and the .similar one in the court of the village, show a short flight of stone steps on their eastern side. Entrance to the ceremonial chamber is prevented when necessary by the removal of the ladder from the outside, or in some instances by the withdrawal of the rungs, which are loosely inserted into holes iu the side pieces. There is no means of preventing access to the exposed trap doors, which are nearly on a level with the ground. As a matter of convenience and to facilitate the entrance into 8 ETH 8 114 PUEBLO ARCHITECTURE. the kivii of costtiiued and masked dancers, often encumbered with clumsy paraphernalia, steps are permanently built into the outside wall of the kiva in dii-ect contradiction to the ancient principles of construc- tion; that is, in liaving' no permanent or fixed means of access from the ground to the first roof. These are the only cases in which stone steps sprinj;: directly from the ground, although they are a very important feature iu Tusayaii house artrhitecture above the first story, as may be seen in any of the general views of the villages. Tlu' justification of such an arrangement in connection with the indefensible 4viva roof lies obviously iu the different conditions here found as compared with the dwclliuffs. Fig. 21. North kivas of Slmnio]):ivi, aei-n from the southwest. Tli(>, siibtei laneaii kiva of the iShumopavi group, above illustrated, is exceptional as occurring at some distance from the mesa rim. Prob- ably all such exceptions to the rule are located in natural fissures or crevices of the sandstone, or where there was some unusual facility for the excavation of thi^ site to tiie required depth. The most noteworthy example of such inner kiva being located with reference to favorable rock fissures has been already described in discussing the ground plan of Walpi and its southern court-inclosed kiva (]>. 65). M(isi>)tr)/.—The exterior masonry of these cliambers seems in all cases to be of ruder construction than that of the dwelling hou.ses. This is particularly noticeable in the kivas of Wali)i on the mesa edge, but is api)arent even iu some of the ZuPii examples. One of the kivas of house No. 1 in Zufd, near the churchyard, has small openings iu its wall that are rudely framed with stone slabs set in a stone wall of ex- ceptional roughness. A])pa-renrly there has n(^ver been any attempt to smooth or reduce this wall to a finished surface with the usual coating of adobe mud. I mk MLVDELEFFI KIVAS IN TUSAYAN. 115 In Tusayaii also some of the kiva walls look as thonsli tliey had Ix-en built of the first material tliat came to haud, i)iled up nearly dry, and with no attempt at the chinking of joints, that imparts some decree of finish to the dwelling-house masonry. The inside of tliese kivas, how- ever, is usually plastered smoothly, but the interior plastering is applied on a base of masonry even in the case of the kivas that are wliolly subterranean. It seems to be the Tusayan practice to line all sides of the kivas with stone masonry, regardless of the eomi)leteness and fitness of the natural cavity. It is impossible, therefore, to ascertain from the interior of a kiva how much of the work of excavation is artificial and how much has been done by nature. The lining of masonry probably holds the i)la-*tering of adobe mud nuu-h better than the naked surface of the rock, but the Tusayan builders would liardly resort to so labor- ious a device to gain this small advantage. The explanation of this apparent waste of labor lies in tlie fact that kivas liad been built of masonry from time immemorial, and that the changed conditions of the present Tusayan environment have not exerted their influence for a sufik'ient length of time to overcome the traditional i)ractice. As will be seen later, the building of a kiva is accompanied by certain rites and ceremonies based on the use of masonry walls, additional testimony of the comparatively recent date of the present subterranean tyi)es. Orientation.—In questioning the Tusayan ou this subject Mr. Stei)hen was told that no attention to the cardinal points was observed in tlie plan, although the walls are spoken of according to the direction to which they most closely :ii)proximate. An examination of the village plans of the preceding chapters, however, will show a remarkable de- gree of uniformity in the directions of kivas which can scarcely be due to accident in rooms built ou such widely differing sites. Tlie intention seems to have been to arrange these ceremonial chambers approxi- mately on the north and south line, though none of the examples ap- proach the meridian very closely. Most of them face southeast, though some, particularly in Walpi, face west of south. In Wal]ii four of the five kivas are planiied on a southwest and northeast line, following the general direction of the mesa edge, while the remaining one faces southeast. The difference in this last case may have been brought about by exigencies of the site on the mesa edge and the form of the cavity in which the kiva was built. Again at llano and Sichumovi (Pis. XVI and xviii) on the first mesa this uniformity of direction pre- vails, but, as the plans show, the kivas in these two villages are few in number. The two kivas of Shupaulovi will be seen (PI. xxx) to have the same direction, viz, facing southeast. In Shumopavi (PI. xxxiv) there are four kivas all facing southeast. In Mashongnavi, however (PI. XXVI), the same uuiformity does not prevail. Three of the kivas face south of east, and two others built in the edge of the rocky bench on the south side of the village face west of south. In the large village 116 PUEBLO ARCHITECTURE. of Oraibi there is lenuxrkable uniformity in the direction of the many kivas, there being a variation of only a few degrees in direction in the whole number of thirteen shown on the plan (PI. xxxvi). But in the case of the large kiva partly above ground de.signated as the Coyote kiva, the direction from which it is entered is the reverse of that of the other kivas. No explanation is offered that will account for tliis curious single exception to the rule. The intention of the builders has evi- dently been to make the altar and its attendant structural features con- form to a definite direction, fixed, perhaps, by certain requirements of the ceremonial, l)ut the irregularity of the general village plan in many cases resulting from its adaptation to restricted sites, has given rise to the variations that are seen. In Zufii there was an e\'ideut purpose to preserve a certain uniform- ity of direction in the kiva entrances. In house No. 1 (Pis. Lxxvi and Lxxvii) there are two kivas, distinguishable on the x'lan by the large divided trap door. The entrance of these both face southeast, and it can readily be seen that this conformity has been piovided intention- ally, since the rooms themselves do not correspond in arrangement. The roof opening is in one case across the room and in the other it is idaced longitudinally. As has been pointed out above, the general plan of arranging the kivas is not so readily distinguished in Zufii as in Tusayan. Uniformity, so far as it is traceable, is all the more striking as occurring where there is so much more variation in the dir(M'tions of the walls of the houses. Still another confirmation is fur- nished by the pueblo of Acoma, situated about (50 miles eastward from Zuui. Here the kivas are six in number and the directions of all the examples are found to vary but a few degrees. These also face east of south. There are reasons for believing that the nse of rectangular kivas is of later origin in the pueblo system of building than the use of the circular form of ceremonial chamber that is of such frequent occurrence among the older ruins. Had strict orientation of the rectangular kiva prevailed for long jieriods of time it would undoubtedly have exerted a strong influence towards the orientation of the entire pueblo clusters in which the kivas were incorporated; but in the earlier circular form, the constructional ceremonial devices could occupy definite positions in relation to the cardinal points at any part of the inner curve of the wall without necessarily exerting any influence on the directions of adjoin- ing dwellings. The ancient form of Mva.—In none of the ruins examined in the province of Tusayan have distinct traces of ancient kivas been found, nor do any of them aftbrd evidence as to the character of the ceremonial rooms. It is not likely, however, that the present custom of building these chambers wholly under ground i)revailed generally among the earlier Tusayan villages, as some of the remains do not occupy sit«s that would suggest such arrangement. The typical circular kiva char- MiNDELEFP.] KIVAS IN TUSAYAN. 117 acteristic of most of the aucieut pueblos has not been seen within the limits of Tusayan, although it occurs constantly in the ruins of Canyon de Chelly which are occasionally referred to in Tusayan tradition as having been occu])ied by related peoples. Mr. Stephen, however, found vestiges of such ancient forms among the debris of fallen walls occupy- ing two small knolls on the edge of the first mesa, at a point that over- looks the broken-down ruin of Sikyatki. On the .southeast shoulder of one of the knolls is a fragment of a circular wall which was originally 12 feet in diameter. It is built of tlat stones, from 2 to i inches thick, 6 to 8 inches wide, and a foot or more in length, nearly all of which have been pecked and dressed. Mud mortar has been sparingly used, and the masonry shows considerable care and skill in execution; the curve of the wall is fairly true, and the interstices of the masonry are neatly filled in with smaller fragments, in the manner of some of the best work of the Canyon de Chelly ruins. The knoll farther south shows similar traces, and on the southeast slope is the complete ground plan of a round structure 16i feet in diameter. At one point of the curved wall, which is about 22 inches thick, occurs the characteristic recessed katchinkihu (described later in discussing the interior of kivas) indicating the use of this chamber for ceremonial purposes. Although these remains probably antedate any of the Tusayan ruins discussed above (Chapter ii), they suggest a connection and relationshi]) between the typical kiva of the older ruins and the radically different form in use at the present time. Native e.vpla nations of position.—iSTotwithstanding the present prac- tice in the location of kivas, illustrated in the plans, the ideal village plan is still acknowledged to have had Its house-clusters so distributed as to form inclosed and protected courts, the kivas being located within these courts or occupying marginal positions in the house-clusters on the edge of the inclosed areas. But the native explanations of the traditional plan are vagiie and contradictory. In the floor of the typical kiva is a sacred cavity called the sipapah, through which comes the beneficent influence of the deities or powers invoked. According to the accounts of some of the old men the kiva was constructed to inclose this sacred object, and houses were built on every side to surround the kiva and form its outer wall. In earlier times, too, so the priests relate, i)eople were more devout, and the houses were planned with their terraces fronting ujyon the court, so that the women and children and all the people, could be close to the masked dancers (katchinas) as they issued from the kiva. The spectators filled the terraces, and sitting there they watched the katchinas dance in the court, and the women sprinkled meal upon them, while ilu'y listened to their songs. Other old men say the kiva was excavated in imitation of the original house in the interior of the earth, where the human family were created, and from which they climbed to the surface of the ground 118 PUEBLO ARCIHITECTURE. by means of a ladder, and tliroiigli just such an opening as the hatch- way of tlie kiva. Another exphmation commonly offered is that they are made underground because they are thus cooler in summer, and more easily warmed in winter. All these factors may have had some influence in the design, but we have already seen that excavation to the extent here practiced is wholly exceptional in pueblo building and the unusual d(>vel()])ni('nt of this requirement of kiva construction has been due to purely, local causes. lu the habitual practice of such an ancient and traditional device, the Indians have lost all I'ecord of the real causes of the jierjictnation of this rc([uircment. At Zuiii, too, a curious explanation is ottered for the l)artial depression ofthe kiva floor below tlie general surrounding level. Here it is naively explained that the floor is excavated in order to attain a liberal height for the ceiling within the kiva, this being a room of great importance. Apparently it does not occur to the Zuni archi- tect that the result could be achieved in a more direct and much less laborious manner by making the walls a foot or so higher at the time of l)uilding the kiva, after the manner in which the same problem is solved when it is encountered in their ordinary dwelling house con- struction. Such explanations, of course, originated long after the prac- tice became established. METHODS OF KIVA BUILDIXG AND RITES. The external ajjpearance of the kivas of Tnsayan lias been described and illustrated ; it now remains to examine the general form and method of construction of these subterranean rooms, and to notice the at- tendant rites and ceremonies. Typical plans.— All the Tusayan kivas are in the form of a paralello- gram, usually about 25 feet long and half as wide, the ceiling, which is from oj to S feet high, being slightly higher in the middle than at either end. There is no piescribed rule for kiva dimensions, and seemingly the size of the chamber is determined according to the number who are to use it, and who assume the labor of its construction. A list of tyid- cal measurements obtained by Mr. tStephen is ai)i)cnded (p. 130). An excavation of the destred dimensions having been made, or an existing one having been discovered, the person who is to be chief of the kiva performs the same ceremony as that jirescribed for the male head of a family when the building of a dwelling house is nndertaken. He takes a handful of meal, mixed with jiiki crumbs, and a little of the crumbled herb they use as' tobacco, and these he sprinkles upon the ground, beginning on the west side, passing southward, and so around, the s]iiinkled line he describes marking the iiosition to be occupied by the walls. As lie thus marks the compass of the kiva, he sings in a droning tone "Si-ai, a-hai, a-hai, si-ai, a-hai''"—no other words but these. The meaning of these words seems to be unknown, but all the ]niests agiee in saving that the archaic chant is addressed to the sun, and it BUREAU OF ETHNOLOGY Scale. 50 100 I 'Ill I I CIGHTH A-;';UAL REPORT PL. LV BUREAU or CTHNOLOCVV CIOHTM AWUAL ftEP{ MINBELEFF.) KIVA BUILDING. 119 is called Kitdaiiwi—the House Song. The chief then selects four good- sized stones of hard texture for corner stones, and at each corner he lays a baho, previously i(rci)ared, sprinkles it with the mixture with which he has described the line of the walls, and then lays the corner stone upon it. As he does this, he expresses his hope that the walls "will take good root hold," and stand Arm and secure. The men have already (juarried or collected a sufticient quantity of stone, and a wall is built in tolerably regular courses along each side of tiie excavation. The stones used are roughly di-essed by fracture; they are irregular in shape, and of a size convenient for one man to handle. They are laid with only a very little mud mortar, and carried uj), if the ground be level, to within 18 inches of the surface. If the kiva is built on the edge of the cliff", as at Walpi, the outside wall con- nects the sides of the gaj), conforming to the line of the cliff". If the surface is sloping, the level of the roof is obtained by building up one side of the kiva above the ground to the requisite height as illustrated in Fig. 21. One end of the "Goat" kiva at Walpi is 5 feet above ground, the other end being level with the sloping surface. When the ledge on the precipitous face of the mesa is uneven it is filled in with rough masonry to obtain a level for the floor, and thus the outside wall of some of the Walpi kivas is more than 12 feet high, although in the interior the measurement from floor to ceiling is much less. Both Cottonwood and pine are used for the roof timbers; they are roughly dressed, and some of them show that an attempt has been made to hew them with four sides, but none are square. In the roof of the "Goat" kiva, at Walpi, are four well hewn pine timbers, measuring exactly 6 by 10 inches, which are said to have been taken from the mission house built near Waljii by the Spanish priests some three cen- turies ago. The ceiling plan of the numgkiva of Shupaulovi (Fig. 2.3) shows that four of these old Spanish squared beams have been utilized in its construction. One of these is covered with a rude decoration of gouged grooves and l)ored holes, forming a curious line-and-dot orna- ment. The other kiva of this village contains a single undecorated square Spanish roof beam. This beam contrasts \'ery noticeably with the rude round poles of the native work, one of which, in the case of the kiva last mentioned, is a forked trunk of a small tree. Some of the Indians say that the timbers were brought by them fi-om the Shumopavi spring, where the early Si)anish priests had established a mission. According to these accounts, the home mission was established at Walpi, with another chapel at Shumoi)avi, and a third and important one at Awatubi. One man, Sikapiki by name, stated that the squared and carved beams were brought from the San Francisco Mountains, more than a hundred miles away, under the direction of the priests, and that they M'ere carved and finished i)rior to traiisjjortation. They were intended for the chapel and cloister, but the latter building was never fliiished. 120 PUEBLO ARCHITECTURE. Tlif roof timbers were ttually distributed uiuoug tlie people of Sliumo- piivi and Shupaulovi. At Shumopavi one of the kivas, known as the Nn\ watikyuobi (Tlie-higii-plaee-of-snow — San Franciseo Mountains) kiva, was built only 8 years ago. The main roof timbers are seven in number. Four of them are bewu with flat sides, 8 by 12 inehes to 9 by 13 inehes; the other three are round, the under sides slig-htly hewn, and they are 12 inches in diameter. These timbers were brought from the San Franeiseo Mountains while the Spaniards were here. The Shumopavi account states that the people were compelled t« drag most of the timbers with ropes, although oxen were also used in some cases, and that the Spaniards used them to roof their mission buildings. After the destruction of the mission these timbers were used in the construction of a dwelling house, which, falling into ruin, was aban- doned and pulled down. Subsequently they were utilized as described above. In the Tcosobi, Jay, the main timbers were taken out of it many years ago and used in another kiva. The timbers now in the roof are quite small and aie laid in pairs, but they are old and much de- cayed. In the Gyarzobi, Paroquet, are six squared timbers from the Spanish mission buildings, measuring 9 by 13 inches, 8 by 12 inches, etc. These have the same curious grooved and dotted ornamentation that occurs on the square beam of Shupaulovi, above described. At the other end of the kiva are also two unusually perfect round timbers that may have come fi'om the mission ruin. All of these show marks of Are, and are in jtlaces deeply charred. In continuationof the kiva building ])rocess, the tops of the walls are brought to an ai)i)roximate level. The main roof timbers are then laid ])arallel with the end walls, at irregular distances, but less than 3 feet a])art, except near the middle, where a space of about 7 feet is left be- tween two beams, as there the hatchway is to be built. The ends of the timbers rest upon the side walls, and as they are placed in position a small feather, to which a bit of cotton string is tied (nakwakwoci) is also placed uuder each. Stout poles, from whicli the bark has been stripped, are laid at right angles upon the timbers, with slight spaces between them. Near the center of the kiva two short timbers are laid across the two main beams about 5 feet apart; this is done to preserve a space of 5 by 7 feet for the hatchway, which is made with walls of stone laid in luud ])laster, resting upon the two central beams and upon the two side piecc^s. This wall or combing is carried up so as to be at least 18 inches above the level of the finished roof. Across the poles, covering the rest of the roof, willows and straight twigs of any kind are laid close together, and over tliese is placed a layer of dry grass arranged in regular rows. Mud is then carefully S])read over the grass to a depth of about 3 inches, and after it has nearly dried it is again gcme over so as to fill up all the cracks. A layer of dry earth is then spread over all and lirmly trodd(>ii down, to I'ender the roof water-tight and bring its surface level with the surrounding ground, following the same method and order of construction that prevails in dwelling-house buildings. i^^V f^^ I ^ ^'l y ' ^^ ;>!*' r/2lOT» ,|mi ifflirr II MINDELEFF.) KlVA BUILDING. 121 Short timbers are i^laced across the top of the hatchway wall, one end of which is raised higher than the other, so as to form a slope, and upon these timbers stone slabs are closely laid for a cover. (See PI. Lxxxvii.) An open space, usually about 2 by 4|^ feet, is preserved, and this is the only outlet in the structure, serving at once as doorway, window, and chinniey. The roof being finished, a floor of stone flags is laid ; but this is never in a continuous level, for at one end it is raised as a platform some 10 or 12 inches high, extending for about a third of the length of the kiva and terminating in an abrupt step .just before coming under the hatch- way, as illustrated in the ground plau of the muugkiva of ShuiTaulovi (Fig. 22, and also in Figs. 25 and 27). On the edge of the platform rests the foot of a long ladder, which leans against the higiier side of the hatchway, and its tapering ends project 10 or 12 feet in the air. Upon this platform the women and other visitors sit when admitted to witness any of the ceremonies observed in the kiva. The main floor in a few of the kivas is composed of roughly liewu planks, but this is a comparatively recent innovation, and is not generally deemed desirable, as the movement of the dancers on the wooden floor shakes the fetiches out of position. On the lower or main floor a shallow pit of varying dimensions, but usually about a foot square, is made for a fireplace, and i.s located immediately under the opening in the hatchway. The intention in raising the hatchway above the level of the roof and in elevating the ceiling in the middle is to prevent the fire from igniting them. The ordinary fuel used in the kiva is greasewood, and there are always several bundles of the shrub in its green state suspended on pegs driven in the wall of the hatchway directly over the fire. This shrub, when green, smolders and emits a dense, pungent smoke, but when perfectly dry, burns with a bright, sparkling flame. Across the end of the kiva on the main floor a ledge of masonry is built, usually about 2 feet high and 1 foot wide, which serves as a shelf for the display of fetiches and other paraphernalia during stated observ- ances (see Fig. 22). A small, niche-like aperture is made in the middle of this ledge, and is called the katchiu kihu (katchina house). During a festival certain masks are placed in it when not in use by the dancers. Some of the kivas have low ledges built aloug one or both sides for use as seats, and some have none, but all except two or three have the ledge at the end containing the katchina house. In the main floor of the kiva there is a cavity about a foot deep and 8 or 10 inches across, which is usually covered with a short, thick slab of Cottonwood, whose upper surface is level with the floor. Through the middle of this short plank and immediately over the cavity a hole of 2 or 2i inches in diameter is bored. This hole is tapered, and is accurately fitted with a movable wooden plug, the top of which is flush with the surface of the plank. The plauk and cavity usually occupy a position 122 PUEBLO ARCHITECTURE. in the iiiaiu floor near the eucl of the kiva. Tliis feature is the sipapuh, the phiue of the gods, and the most sacred portion of the ceremonial chamber. Around this spot the fetiches are set during a festival; it typifies also the first world of the Tusayan genesis and the opening through which the people first emerged. It is freqixeutly so spoken of at the present time. Other little apertures or niches are constructed in the side walls ; they usually open over the main floor of the kiva near the edge of the dais that forms the second level, that ui)oii which the foot of the ladder rests. These are now dedicated to any special purpose, but are used as recep- tacles for small tools and other ordinary articles. In early days, how- ever, these niches were used exclusively as receptacles for the sacred pipes and tobacco and other smaller paraphernalia. In order to make clearer the lelative positions of the varioirs features of kiva construction that have been described several typical examples are here illustrated. The three ground plans given are drawn to scale and represent kivas of average dimensions. Mr. Stei)hen has made a series of typical kiva measurements, which is appended to this sec- tion, and comparison of these with the plans will show the relation of the examjjles selected to the usual dimensions of these rooms. Fig. 22 is the ground plan of the mungkiva, or chief kiva, of Shupaulovi. Fig. 22. Grunud plan of the chief kivji of Shupaulovi. It will be observed that the second level of the kiva floor, forming the dias before referred to, is about 15 inches narrower on each side than the main floor. The luxrrowiug of this portion of the kiva floor is not universal and does not seem to be regulated by any rule. Sometimes the narrowing is carried out on one side only, as in the mungkiva of Mashonguavi (Fig. 27), sometimes on both, as in the present example, KIVA BUILDING. 123 and in other cases it is absent. In the second kiva of Shupaulovi, il- kistrated in Fig. i!.'5, there is only one small jog' that has been bnilt mid way along the wall of the upper level and it bears no relation to the point at which the change of floor level occnrs. The ledge, or dias, is free for the nse of spectators, the Indians say, jnst as the women stand on the honse terraces to witness a dance, and do not step into the conrt. The ledge in this case is about a foot above the main floor. Benches of masonry are built along each side, though, as the plan shows, they are not of the same length. The bench on the eastern side is about -i feet shorter than the other, which is cut oft' by a continuation of the high bench that contains the katchinkihu beyond the corner of the room. These side benches are for the use of participants in the ceremonies. "When young men are initiated into the various societies during the feasts in the fall of the year they occupy tlie floor of the sacred divi- sion of the kiva, while the old members of the order occupy the benches along the wall. The higher bench at the end of the room is used as a shelf for paraphernalia. The hole, or recess, in this bench, whose po-' sition is indicated by the dotted lines on the plan, is the sacred orifice from which the katchina is said to come, and is called the katchinkihu. Fig. 23. Ceiling plan of the chief kiva of Shupaulovi. In the floor of the kiva, near the katchinkihu, is the sipapuh, the cot- tonwood plug set into a Cottonwood slab over a cavity in the floor. The plan shows how this plank, about IS inches wide and ()i feet long, has been incorporat(Ml into the paving of the main floor. The paving is composed of some quite large slabs of sandstone whose irregular edges have been skillfully fitted to form a smooth and well finished pavement. The position of the niches that form pipe receptacles is shown on the l>lan opposite the fireplace in each side wall. The position of the foot of the ladder is indicated, the side poles resting upon the paved sur- face of the second level about lii inches from the edge of the step. Fig. i'.3 gives a ceiling plan of the same kiva, illustrating the arrangement 124 PUEBLO ARCHITECTURE. of vsuch of the roof beams and sticks as are visible from inside. The plau shows the position of the four Spanish beams before referred to, the northernmost being the one that has the line and dot decoration. The next two beams, hiid in contact, are also square and of Spanish make. The fourth Spanish beam is on the northern edge of the hatch- way dome and supports its wall. The adjoining beam is round and of native workmanship. The position and dimensions of the large hatch- way projection are here indicated in ]>lau, but the general appearance of this ciu-ious feature of the Tusayan kiva can be better seen from the interior view (Fig. 24:). Various uses are attributed to this domelike Fig. 24. Interior view of a Tuaayan kiva. structure, aside li-om the explanation that it is built at a greater height in order to lessen the danger of ignition of the roof beams. The old men say that formerly they smoked and preserved meat in it. Others say it was used for drying bundles of wood by suspension over the fire preparatory to use in the fireplace. It is also said to constitute an upi^er chamber to facilitate the egress of smoke, and doubtless it aids in the performance of this good ofiice. The mnd i)laster that has been applied directly to the stone work of the interior of this kiva is very much blackened by smoke. From about half of the wall space the plaster has fallen or scaled off, and the ex- MTNDBLEPF.] SHUPAULOVI KIVAS. 125 posed stonework is much blackened as though the kiva had hjug been used with tlie wall in this uncovered conditiou. The fireplace is simply a shallow pit about 18 inches square that is placed directly under the opening of the combmed hatchway and smoke hole. It is usually situated from 2 to 3 feet from the edge of the second level of the kiva floor. The paving stones are usually finished quite neatly and smoothly where their edges enframe the flrepit. Fio. 25. Ground plan of a Shupanlovi kiva. Fig. 26. Ceiling plan of a Shupaulovi kiva. Figs. 25 and 26 illustrate the ground and ceiling plans of the second kiva of the same village. In all essential principles of arrangement it is identical with the preceding exam])le, but minor modifications will be noticed in several of the features. The bench at the katchina, or "alta,r" end of the kiva, has not the height that was seen in the mung- kiva, but is on the same level as the benches of the sides. Here the 126 PUEBLO ARCHITECTURE. sipapuli is at much j^rcater distance than iisnal t'loni tlic katrliina re- ces.s. It is also qnite cxceptioual in tliat the pliijf is let into an oriflee in one of the paving stones, as shown on the plan, instead of into a Cottonwood plank. Some of the paving stones forming the tloor of this kiva are quite regnlar in shape and of unusual dimensions, (me of them being nearly 5 feet long and 2 feet wide. The gray polish of long eon- tinued use imparts to these stones an appearance of gi'eat hardness. The ceiling plan of this kiva (Fig. 20) shows a single specimen of Span- ish beam at the extreme north end of the roof. It also shows a forked "viga" or ceiling beam, which is quite unusual. This kiva is better plastered than the mungkiva and shows in i)laces evidences of many successive coats. The general rule of tipidying the interior plastering of the kiva on a base of masonry has been violated in this example. The north end and part of the adjoining sides have been brought to an even fixce by tilling in the inc(iualities of the exca- vation with reeds which are applied in a vertical position and are held in place by long, slender, horizontal rods, forming a rude matting or Fig. 27. Ground plau of the chief kiva of Masbougnavi. wattling. The rods are fastened to tlie rocky wall at favorable points by means of small prongs of some hard wood, and the whole of the primitive lathing is then thickly plastered with adobe nnid. Mr. Stephen found the Ponobi kiva^ of Oraibi treated in the same manner. The walls are lined with a reed lathing over which mud is plastered. The reed used is the Bakabi {rhni< cdinmunis) whose stalks vary from a quarter of an inch to three-quarters of an inch in diameter. In this instance the reeds are also laid vertically, but they are api)lied to the ordinary mud-laid kiva wall and not directly to the sides of the natural excavation. The vertical laths are bound in place by hori- zontal re«ds laid upon them 1 or 2 feet apart. The horizontal reeds MIXDELEFF. THE KIVA HATCHWAY. 127 are held in place by pegs of greasewood driven into the wall at inter- vals of 1 or ii feet and are tied to the pegs with split yucca. These specimens are very interesting examples of aboriginal lathing and i)las- tering applied to stone work. The ground plan of the mungkiva of Mashongnavi is illustrated in Fig. '21. In this example the narrowing of the room at the second level of the floor is on one side. The step by which the upper level is reached fii'om the main floor is 8 inches high at the east end, rising to 10 inches at the west end. The south end of the kiva is provided ^-ith a small opening like a loop-hole, furnishing an outlook to the south. The east side of the main portion of the kiva is not provided with the usual bench. The jjortiou of the bench at the katchina end of the kiva is on a level witli the west bench and continuous for a couple of feet beyond the northeast corner along the east wall. The small wall niches are on the west side and nearer the north end than usual. The arrangement of the katcliinkihu is quite different from that described in the Shupau- lovi kivas. The orifice occurs in the north wall at a height of 3.i feet above the floor, and 2 feet 3 inches above the top of the bench that ex- tends across this end of the room. The flrepit is somewhat smaller than in the other examples illustrated. Fig. 28 illustrates the appear- FiG. 28. Interior view of a kiva hatchway in Tusayau. ance of the kiva hatchway fiom within as seen fi-om the north end of the kiva, but the ladder has been omitted from the drawing to avoid confusion. The ladder rests against the edge of the coping that caps the dwarf wall on the near side of the hatchway, its toi) leaning toward the spectator. The small smoke-blackened sticks that are used fin- the suspension of bundles of greasewood and other fuel in the hatchway are clearly shown. At the far end of the trapdoor, on the outside, is indicated the mat of reeds or rushes that is used for closing the open- ings when necessary. It is here shown rolled uj) at the foot of the slope of the hatchway top, its customary position when not in use. 128 PUEBLO ARCHITECTURE. When this mat is used for closiiii;- the kiva openino- it is usually held in place by several lari;e stime slabs laid (tver it. Fig. 29 illustrates a specimeu of the Tusayau kiva mat. Fig. 29. Mat used in i-iosiiij; tin- cntrauoL- . The kiva chief then proclaims the name by which the kiva will be known. This is often merely a term of his choosing, often without leference to its appropri ateness. Various uses of Jiinits.—Allusions occur in some of the traditions, sug- gesting that in earlier times one class of kiva was devoted wholly to the purposes of a ceremonial chamber, and was constantly occupied by a priest. An altar and fetiches were permanently maintaiiu'd, and approi)riate groups of these fetiches were displayed from month to niontli, as the different priests of the sacred feasts succeeded each other, each new moon bringing its prescribed feast. Many of the kivas were built by religious societies, which still hold their stated observances in them, and in Oraibi several still bear the names of the societies using them. A society always celebrates in a particular kiva, but none of these kivas are now preserved exclusively for religious purposes; they are all places of social resort for the men, especially during the winter, when they occui)y themselves with the arts conunon among them. The same kiva thus serves as a temple dur- ing a sacred feast, at other times as a council house for the discussion of public affairs. It is also used as a workshop by the industrious and as a lounging place by the idle. There are still traces of two classes of kiva, marked by the distinc- tion that only certain ones c(mtain the sipapuli, and in tliese theinore im])ortant ceremonies are held. It is said that no sipapuh has been made recently. The prescribed operation is performed by the chief and the assistant pinests or fetich keepers of the society owning the kiva. Some say the mystic lore pertaining to its preparation is lost and none can now be made. It is also said that a stone sipa])uli was formerly used instead of the cottonwood plank now commonly set'n. The use of stone for this xmr^jose, however, is nearly obsolete, though the second MTOTELEFF.] USES OF KIVAS. 131 kiva of Shupaulovi, illustrated in plan in Fig. 25, contains an example of this ancient form. In one of tlie newest kivas of Mashongnavi the plank of the sipapuh is pierced with a square hole, Avhich is cut with a shoulder, the shoulder supporting the plug with which the orifice is closed (see Fig. 30). This is a decided innovation on the traditional form, as the orifice from which the people emerged, which is symbolized in the sipa- puh, is described as being of circular form in all the versions of the Tusayan genesis myth. The presence of the sipapuh possi- bly at one time distinguished such kivas ^'o- ^"^ I'^'^^t'iusuiar siiiapuh in a . , 1 • 1 ,1 Maabongnavi kiva. as were . . considered strictly consecrated to reUgious observances from those that were of more general use. At Tusayan, at the present time, certain societies do not meet in the ordinary kiva but in an apartment of a dwelling house, each society having its own exclusive place of meeting. The house so used is called the Inmse of the " Sister of the eldest brother," meaning, probably, that she is the descendant of the founder of the society. This woman's house is also called the " house of grandmother," and in it is preserved the tiponi and other fetiches of the society. The tiponi is a ceremonial object about IS inches long, consisting of feathers set upright around a small disk of silicified wood, which serves as its base when set upon the altar. This fetich is also called iso (grandmother), hence the name given to the house where it is kept. In the house, where the order of warriors (Kuleataka) meets, the eldest son of the woman who owns it is the chief of the order. The ai)artment in which they meet is a low room on the ground floor, and is entered only by a hatchway and ladder. There is no sipapuh in this chamber, for the warriors appeal directly to Cotukinungwa, the heart of the zenith, the sky god. Large figures of animal fetiches are painted in difterent colors upon the walls. On the west wall is the Mountain Lion; on the south, the Bear; on the east, the Wild Cat, surmounted with a shield inclosing a star; on the north, the White Wolf; and on the east side of this figure is painted a large disk, rcpresoiiting the sun. The walls of the chambers of the other societies are not decorated permanently. Here is, then, really another class of kiva, although it is not so called by the people on the Walpi mesa. The ordinary term for the ground story rooms is used, "kikoli," the house without any opening in its walls. But on the second mesa, and at Oraibi, although they sometimes use this term kikoli, they com- monly apply the term "kiva" to the ground story of the dwelling house used as well as to the underground chambers. It is probable that a class of kivas, not specially consecrated, has existed from a very early peiiod. The rooms in the dwelling houses have always been small and dark, and in early times without chimneys. 132 PUEBLO ARCHITECTURE. * Within such cramped limits it was inconvenient for the men to practice any of the arts they Icnew, especially -weaving, wliich could ha\-e been carried on out of doors, as is done still occasionally, but subject to many interruiitions. It is possible that a class of kivas was designed for such ordinary x)urposes, though now one tj^ie of room seems to answer all these various uses. In most of the existing liivas there are planks, in which stout loops are secured, tixed in the floor close to the wall, for attaching the lower beam of a primitive vertical loom, and ])rojecting vigas or beams are inserted into the walls at the time of their construc- tion as a provision for the attachment of the upper loom ]X)les. The jilanks or logs to which is attached tlie lower part of the loom appear in some cases to be quite carefully worked. They are often partly buried in the ground and under the edges of adjacent paving stones in such a numuer as to be held in place very securely against the strain of the tightly stretched warp while the blanket is being made. The holes x)ierced in the upper surface of these logs are very neatly executed in the manner illustrated in Fig. 31, which shows one of the orifi(;es in section, together AN-itli the adjoining paving stones. The outward ap- pearance of the device, as seen at short intervals along the length of the log, is also shown. Strips of buckskin or bits of rope are passed through these U-^^lnvped cavities, and F.Q. 31. Loom post in kiva at Tusayau. ^^^^.^^ ^^^,^.j. jj^^ j^^^^.^.^. ^,,,j^. ,,f ^j,^ j^,^,,^^ at the bottom of the extended series of warp threads. The latter can thus be tightened preparatory to the operiition of tilling in with the woof. The kiva looms seem to be used mainly for weaving the dark- blue and black blankets of diagonal and diamond pattern, which form a staple article of trade with the Zuiii and the Eio Grande Pueblos. As an additional convenience for the practice of weaving, one of the kivas of Mashongnavi is provided with movable seats. These consist simply of single stones of suitable size and form. Usually they are 8 or 10 inches thick, a foot wide, and perhaps 15 or 18 inches long. Be- sides their use as seats, these stones are used in connection with the edges of the stone slabs that cap the pei'inanent benches of the kiva to support temporarily the upper and lower poles of the blanket loom while the warj) is gradually wound around them. Tln^ large stones that are incorporated into the side of the benches (if some of the Mashong- navi kivas have occasionally i-ound, cup-shaped cavities, of about an inch in dianu'ter, drilled into them. These holes receive one end of a warp stick, the other end being sui)poi'ted in a corresponding hole of the heavy, movable stone seat. The other warp stick is supported in a similar manner, while the thi'cad is ])assed around both in a horizontal direction ])rei)aratory to jilacing and stretching it in a vertical ])ositioii for the final working of the blanket. A number of these cup-shaped MINDELEFF ) RIVA OWNERSHIP. 133 pits are formed aloug the side of tlie stone bench, to provide for various lengtlis of warp that may be required. Ou the opposite side of this same kiva a number of similar holes or depressions are turned into the mud plastering of the wall. All these deviees are of common occur- rence at other of the Tusayan kivas, and indicate the antiquity of the practice of using the kivas for such industrial purposes. There is a suggestion of similar use of the ancient circular kivas in an example in Canyon de Chelly. At a small cluster of rooms, built partly ou a rocky ledge and partly on adjoining loose earth and rocky debris, a land shde had carried away half of a circular kiva, exposing a well-defined section of its tloor and the d6bris within the room. Here the writer found a number of partly finished sandals of yucca fiber, with the long, unwoven fiber carefully wrapped about the finished portion of the work, as though the sandals had been temporarily laid aside until the nuiker could again work on them. A number of coils of yucca liber, similar to that used in the sandals, and several balls of brown liber, formed fi'om the inner bark of the cedar, were found on the floor of the room. The condition of the rum and the debris that tilled the kiva clearly sug- gested that these specimens were in use just where they were found at the time of the abandonment or destruction of the houses. No traces were seen, however, of any structural devices like those of Tusayan that would serve as aids to the weavers, though the weaving of the par- ticular articles comprised in the collection from this s])ot would prob- ably not require any cumbrous apparatus. Kiva ownership.—The kiva is usually spoken of as being the home of the organization which maintains it. Diflereut kivas are not used in common by all the inhabitants. Every man has a membership in some particular one and he frequents that one only. The same person is often a member of different societies, which takes him to dilierent kivas, but that is only ou set occasions. There is also much informal visiting among tliein, but a man presumes to make a loitering place only of the kiva in which he holds membership. In each kiva there is a kiva mungwi (kiva chief), and he controls to a great extent all matters pertaining to the kiva and its membership. This ofUce or trust is hereditary and i)asses from uncle to nephew through the female line—that is, on the death of a kiva chief the eldest son of his eldest sister succeeds him. A kiva may belong either to a society, a group of gentes, or au in- dividual. If belonging to a society or order, the kiva. chief commonly has inherited his oflice in the manuer indicated from the "eldest brother" of the society who assumed its construction. l>ut tin? kiva chief is not necessarily chief of the society; in fact, usually he is but an ordinary member. A similar custom of inheritance prevails where the kiva be- longs to a group of geutes, only in that case the kiva chief is usually chief of the gentile group. As for those held by individuals, a couple of exami)les will illustrate the Tusayan practice. In Hauo the chief kiva was originally built 134 PUEBLO ARCHITECTURE. by a group of "Sun" gentes, but about 45 years ago, during an epi- demic of smallpox, all the people who belonged to the kiva died except one man. The room fell into I'uin, it.s roof timbers were carried off, and it became filled u]> with dust and rubl)isli. The title to it, however, rested with the old survivor, as all the more direct heirs had died, and lie, when about to die, gave the kiva to Kotshve, a "Snake" man from Walpi, who married a Tewa (Hano) woman and still lives in Hano. This man repaired it and renamed it Tokonabi (said to be a Pah-Ute term, meaning black mountain, but it is tlie only name the Tusayan have for Navajo Mountain) because his people (the "Snake") came from that place. He in turn gave it to his eldest son, who is therefore kiva mungwi, but the son says his successor will be the eldest son of his eldest sister. The membership is composed of men fiom all the Hano gentes, but not all nl' any one gens. In fact, it is not now customary for all the members of a gens to be members of the same kiva. Another somewhat similar instance occurs in Sicliumovi. A kiva, abandoned for a long time after the smallpox plague, was taken pos- session of by an individual, who repaired it and renamed it Keviny^p tshomo—Oak Mound. He made his friends its members, but he called the kiva his own. He also says that his eldest sister's son will suc- ceed him as chief. In each village one of the kivas, usually the largest one, is called (aside from its own special name) mungkivii—I'hief kiva. It is fre- quented by the kiinungwi—house or \all;ige cliief—and the tshaak- mungwi—chief talker, councillor—and in it also the more elaborate ceremonies are observed. No women frequent any of the kivas; in fact they never enter them except to plaster the walls at customary periods, or during the occasion of certain ceremonies. Yet one at least of the Oraibi kivas was built for the observances of a society of women, the Mamzr^ntiki. This and another female society—Lal6iikob4ki—exist in all the other villages, and on the occasion of their festivals the women are given the exclusive use of one of the kivas. Motipes fur huUding a lird.—Only two causes are mentioned for building a new kiva. Quarrels giving rise to serious dissensions among the occupants of a kiva are one cause. An instance of this occurred quite recently at Hano. The conduct of the kiva chief gave rise to dissensions, and the members oj)posed to him prepared to build a sep- arate loom of their own. They chose a gap on the side of the mesa cliff, close to Hano, collected stones for the walls, and brought the roof timbers from the distant wooded mesas; but when all was ready to lay the foundation their differences were adjusted and a complete reconcil- iation was effected. The other cause assigned is the necessity for additional room when a gens has outgrown its kiva. When a gens has increased in numbers sufiflcieutly to warrant its having a second kiva, the chief of the gen- BUREAU OF ETHNOLOGY EIGHTH ANNUAL REPORT PL. LXIII PLAN OF KIN-TIEL, ALSO SHOWING EXCAVATIONS. MiNDELEFP.] SIGNIFICANCE OF KIVA PLAN. 135 tile group, who in this case is also chief of the order, proposes to his kiu to build a separate kiva, and that being agreed to, he assumes the direction of the construi-tion and all the dedicatory and other cere- monies connected with the undertaking. An instauce of this kind occurred within the last year or two at Oraibi, where the members of the "Katchina" gentes, who are also members of the religious order of Katchina, built a spacious kiva for themselves. The construction of a new kiva is said to be of rare occurrence. On the other hand, it is common to hear the kiva chief lament the deca- dence of its membership. In the " Oak Mound " kiva at Sichumovi there are now but four members. The young men have married and moved to their wives' houses in more thriving villages, and the older men have died. The chief in this case also says that some 2 years ago the agent gave him a stove and pipe, which he set uj) in the room to add to its comfort. He now has grave fears that the stove is an evil innovation, and has exei'cised a deleterious influence upon the fortune of his kiva and its members; but the stove is still retained. Significance of structural plan.—The designation of the curious orifice of the sipapuh as "the place from which the people emerged" in con- nection with the peculiar arrangement of the kiva interior with its change of floor level, suggested to the author that these features might be regarded as typifying the foiu" w(jrlds of the genesis myth that has exercised such an influence on Tusayan customs; but no clear data on this subject were obtained by the writer, nor has Mr. Stei)hen, who is specially well equipped for such investigations, discovered that a defi- nite conception exists concerning the significance of the structural plan of the kiva. Still, from many suggestive allusions made by the various kiva chiefs and others, he also has been led to infer that it typifies the four "houses," or stages, described in their creation myths. The si- pajiuh, with its cavity beneath the floor, is certainly regarded as indi- cating the place of beginning, the lowest house under the earth, the abode of Myuiugwa, the Creator ; the main or lower floor represents the second stage; and the elevated section of the floor is made to denote the third stage, where animals were created. Mr. Stephen observed, at the New Year festivals, that animal fetiches were set in groups upon this platform. It is also to be noted that the ladder leading to the surface is invariably made of pine, and always rests upon the platform, never upon the lower floor, and in their traditional genesis it is stated that the people climbed \i\> from the third house (stage) by a ladder of pine, aud through such an opening as the kiva hatchway; only most of the stories indicate that the opening was round. The outer aii' is the fourth world, or that now occupied. There are occasional references in the Tusayan traditions to circular kivas, but these are so confused with fantastic accounts of early mythic structuies that their literal rendition would serve no useful purpose in the present discussion. 136 PUEBLO ARCHITECTURE. Typical measurements.—The following list is a record of a number of measiu-ements of Tusayau kivas collected by Mr. Steplieu. The wide difference between the end measurements of the same kiva are usually due to the interior offsets tliat have been noticed on the ])lans, but the differences in the lengths of the sides are due to irregularities of the site. The latter differences are not so marked as the former. ^ ^ v^ \ 1 / 5 f. iVji f^' -^- MiNDELEPF.] DETAILS OF CONSTRUCTION. 137 SHUPAULO\-I. 1. A'tkabi kira Place below. 2. Kokyaugobi kiv.a Place of spider. A'tkabi kiva is the luungkiva. SHUMOI'AVl. 1. NuTwa'tikyuobi Higb place of snow, San Francisco Mountain. 2. Al kiva A'la, Horn. 3. Gy'arzobi ...Gy'arzo, Paroquet, a gens. 4. Tco'sobi Blue Jay, a gens. Tco'sobi is the niungkiva. 1. Tdau kiva Tda'uollauwuh.The singers. 2. Ha'wiobi kiva Ha'wi, stair; High stair place. ol)i, high place. 3. Ish kiva Isa'uwnh Coyote, a gens. 4. Kwang kiva Kwa'kwanti Religious order. 5. Ma'zraii kiva Ma'nizrauti Female order. 6. Na'cabi kiva Half way or Central pl.ice. 7. Sa'kwalen kiva . .Sa'kwa le'na . - . Blue Flute, a religions order. 8. Po'ugobi kiva Pongo, a circle .An order who decorate themselves with circular marks on the body. 9. Hano' kiva Ha'nomuh A fashion of cutting the hair. 10. Mote kiva Mo'mtci The Warriors, an order. 11. Kwita'koli kiva. -Kwita, ord,ure; Ordure heap. ko'li, a heap. 12. Katcin kiva Katcina A gens. 13. Ten kiva Tcua, a snake . .Religious order. Tdaii kiva is the muugkiva. DETAILS OF TUSAYAN AND CIBOLA CONSTRUCTION. The complete operation of biiildiug a wall lias never been ob.served at Znfii by the writer, but a close examination of nnnierons finished and some broken-down walls indicates that the methods of construction adopted are essentially the same as those employed in Tusayan, which have been repeatedly observed; with the possible difference, however, that in the former adobe nmd mortar is more liberally used. A singular feature of pueblo masonry as observed at Tusayan is tlie very sparing use of mud in the construction of the walls; in fact, in some instances when walls are built during the dry season, the larger stones are laid up in the walls without the use of mud at all, and are allowed to stand in this condition until the rains come; then the mud mortar is mixed, the interstices of the walls filled in with it and with chinking stones, and the inside walls are plastered. But the usual practice is to com- plete the house at once, finishing it inside and out with the rccpiisite mortar. In some instances the outside walls are coated, completely 138 riTERLO ARCHITECTURE. covering the masonry, but this is not done in many of tlio houses, as may l)e seen by reference to the preceding illustrations of the Tusayan villages. At Zuni, on the other hand, a liberal and frequently renewed coating of nuul is applied to the walls. Only one piece of masonry was seen in the entire village that did not have traces of this coating of mud, viz, that portion of the second story wall of house No. 2 described as possibly belonging to the ancient nucleus pueblo of Halonaand illus- trated in I'l. LViii. Even the rough masonry of the kivas is partly surfaced with this medium, though many jagged stones are stUl visible. As a result of this practice it is now in many cases inipossible to deter- mine from mere superficial inspection whether the nndcrlying masonry has been constructed of stone or of adobe ; a difficulty that may be realized from an examination of the views of Znhi in Clia])ter in. Where the fall of water, such as the discharge from a roof-drain, has removed the outer coating of nuid that covers stonework and adobe alike, a large proportion of these exposures reveal stone masonry, so that it is clearly apparent that Zuni is essentially a stone village. The extensive use of sun-dried bricks of adobe has grown up within quite recent times. It is apparent, however, that the Zuni builders preferred to use stone; and even at the present time they frequently eke out with stonework portions of a house when the supply of adobe has fallen short. An early instance of such supplementaiy use of stone masonry still sur- vives in tlie church building, where the old Spanish adobe has been repaired and filled in with the tyi»ical tabular aboriginal masonry, con- sisting of small stones carefully laid, with very little intervening mortar showing on the face. Such reversion to aboriginal methods probably took place on every opportunity, though it is rcnmrkable that the Indians should have been allowed to einjjloy their own methods in this instance. Although this (ihurch building has for many generations furnished a conspicuous example of tyjiical adobe construction to the Zuni, he has never taken the lesson sufiiciently to heart to closely imi- tate the Spanish methods either in the preparation of the material or in the manner of its use. The adobe bricks of the church are of large and uniform size, and the mud froiu which they were made had a liberal admixture of straw. This binding material does not appear in Zuiii in any other example of adobe that has been examined, nor does it seem to have been utilized in any of the native pueblo work either at this jilaceor at Tusayan. Where molded adobe bricks have been used by the Zuni in housebuilding they have been made from the raw material just as it was taken from the fields. As a result these bricks have little of the durability of the Spanish work. PI. xcvi illustrates an adobe wall of Zuni, part of an unroofed house. The old adobe church at Hawikuh (PI. xlviii), abandoned for two centuries, has withstood the wear o? time and weather better than any of the stonework of the surrounding houses. On the right-hand side of the street that shows in the foreground of PI. lxxviii is an illustration of the construction MDJUELEFF.] USE OF ADOBE IN ZUNI. 139 of a wall with adobe bricks. This example is very recent, as it has not yet been roofed over. The top of the wall, however, is temporarily pro- tected by the usual series of thin sandstone slabs used in the finishing of wall copings. The very rapid disintegration of native-made adobe walls has brought about the use in Zufii of many x)rotec- tive devices, some of which will be noticed in connection with the discussion of roof drains and wall copings. Figs. 32 and 33 illustrate a curious employment of jjottery frag- ments on a mud-plastered wall and on the base of a chinniey to protect the adobe coating against rapid erosion by the rains. These pieces, usually FlO. 32. A ZuDi chimnej^ showing pottery fragments fragments from large vessels, embedded in ita a^lobo base. are embedded in the adobe with the convex side out, forming an armor of i)ottery scales well adapted to resist disintegration by the elements. Fig. 33. A Ziiui oven with pottery aciles embedded in its surface. The introduction of the use of adobe in Zufu should probably be attributed to foreign influence, but the i)osition of the village in the open plain at a distance of several miles from the nearest outcrop of suitable building stone naturally led the builders to use stone more sparingly when an available substitute was found close at hand. The thin slabs of stone, which had to be brought from a great distance, came to be used only for the more exj)osed portions of buildings, such as copings on walls and borders around roof openings. Still, the pueblo 140 PUEBLO ARCHITECTURE. builders never attained to a full appreeiation of the advantages and requirements of this mediuni as compared with stone. The adobe walls are built only as thick as is absolutely necessary, few of them being more than a foot in thickness. The walls are thus, in jiroportion to height and weight, sustained, thinner than the crude brick construction of other peoples, and require protection and constant repairs to insui'e durability. As to thickness, they are evidently modeled directly after the walls of stone masonry, which had already, in both Tusayan and Cibola, been pushed to the limit of thinness. In tiict, since the date of the survey of Zuni, on which the published plan is based, the walls of several roOms over the court passageway in the house, ilhistrated in PI. LXXXii, have entirely fiillen in, demonstrating the insufficiency of the thin walls to sustain the weight of several stories. The climate of the pueblo region is not wholly suited to the employ- ment of adobe construction, as it is there practiced. For several months in the year (the rainy season) scarcely a day passes without violent storms wliich play havoc with the earth-covered houses, necessitating constant vigilanct^ and frequent repairs on the ]>art of the occupants. Though the practice of luud-coating all walls has iu Cibola un- doubtedly led to greater carelessness and a less rigid adherence to ancient methods of construction, the stone masonry may still be seen to retain some of the peculiarities that characterize ancient examples. Featiires of this class are still more apparent at Tusayan, and notwith- standing the rudeness of much of the modern stone masonry of this province, the fact that the builders are familiar with the superior methods of the ancient builders, is clearly shown in the masonry of the i)resent villages. Perhaps the most noteworthy characteristic of pueblo masonry, and one which is more or less present in both ancient and modern examples, is the use of small chinking stones for bringing the masonry to an even face after the larger stones forming the body of the wall have been laid in place. This method of construction has, in the case of some of the best built ancient pueblos, such as those on the Chaco in ^Tew ^Mexico, resulted in the production of marvelously finished stone walls, in which the mosaic-like bits are so closely laid as to show none but the finest joints on the face of the wall with but little trace o{ mortar. The chink- ing wedges necessarily varied greatly in dimensions to suit the sizes of the interstices between the larger stones of the wall. The use of stone in this manner no doubt suggested the ])andcd walls that form so strik- ing a feature in some of the Chaco houses. This arrangement was likely to be brought about by the occurrence in the cliffs of seams of stone of two degrees of thickness, suggesting to the builders the use of stones of similar thickness in continuous bands. The ornamental effect of this device was originally an accidental result of adopting the most convenient method of using the material at hand. Though the masonry of the modern pueblos does not aftbrd examples of distinct bauds, the 1 ^•# 1 MnrDEI-EFF.] ZUNI MASONRY. 141 introdiiction of the small chinking spalls often follows horizontal lines of considerable length. Even in mud-plastered Znni, many ()utero])s of these thin, tabular wedges protrude from the partly eroded niud- coatiug of a wall and indicate the presence of this kind of stone masonry. An example is illustrated in Fig. 34, a tower-like projection at the northeast corner of house No. 2. Fig. 34. Stone wedges of Zuui masonry exposed in rain-washed wall. In the Tusayan hoiTse illustrated in PI. Lxxxiv, the construction of which was observed at Oraibi, the interstices between the large stones that formed the body of the wall, containing but small quantities of 142 PUEBLO AECHITECTUEE. mud mortar, were filled in or plugged with small fragments of stone, whicli, after being ])artly embedded in the mud of the joint, were driven in \yith uiihafted stone hammers, pn)ducing a fairly even face ofmasonry, afterward gone over with mud plastering of the cousisteney of model- ing clay, applied a handful at a time. Piled iip on the ground near the new house at convenient points for the builders may be seen exam- liles of tlic larger wall stones, indicating the marked tabular character of the pueblo masons' material. The narrow edges of similar stones are visible in the uuplastered portions of the house wall, whii-h also illus- trates tlie relative proportion of chinking stones. Tliis latter, however, is a varial)le feature. PI. xv affords a clear illustration of the jyropor- tiou of tliese small stones in the old masonry of Payupki; while in PI. XI, illustrating a ])ortion of the outer wall of the Fire House,' the tablets are fewer in number and thinner, their use i)redoniiuating in the hori- zontal joints, as in the best of the old examples, but not to the same extent. Fig. 35 illustrates the inner face of an unjilastered wall of a Fig. 35. An uniilastind holisi' wall iii Oj.i ('alii-utu. small house at Ojo Caliente, in which the modern method of using the chinking stones is shown. This example bears a strong resemblance to the Payupki masonry illustrated in PI. xv in the irregularity with which the chinking stones are distributed in the Joints of the wall. The same room affords an illustration of a cellar-like feature having the aj)pearance of an intentional excavation to attain a depth for this room BUREAU OF ETHNOLOGY © » #'.^i^ -:r\J^ 'A^ euacAu Of tTMNOif^ tiOMTH UMMIHMMT ty UMI ,..-0 \>,.^' tfft^ W" MINDELEFF.] ZUNI MASONRY. 143 corresponding to the adjoining floor level, but this effect is due simply to a clever adaptation of the house wall to an existing ledge of sand- stone. The latter has had scarcely any artificial treatment beyond the partial smoothing of the rock in a few places and the cutting out of a small niche from the rocky wall. This niche occupies about the same position in this room that it does in the ordinary x)ueblo house. It is remarkable that the pueblo builders did not to a greater extent utilize their skill in working stone in the preparation of some of the irregular rocky sites that they have at times occupied for the more convenient reception of their wall fouiulations; but in nearly all such cases the buildings have been modified to suit the ground. An example of this practice is illustrated in PI. xxiii, from the west side of Walpi. In some of the ancient examples the labor required to so prepare the sites would not have exceeded that expended on the massive masonry com- posed of numberless small stones. Many of the older works testify to the remarkable patience and industry of the builders in amassing and carefully adjusting vast quantities of building materials, and the mod- ern Indians of Tusayan and Cibola have inherited much of this ancient spirit yet this industry was rarely diverted to the excavation of room ; or village sites, except in the case of the kivas, in which special motives led to the practice. In some of the Chaco pueblos, as now seen, the floors of outer marginal rooms seem to be depressed below the general level of the surrounding soil; but it is now difficult to determine wliether such was the original arrangement, as much sand and soil have drifted against the outer walls, raising the surface. In none of the pueblos within the limits of the provinces under discussion has there been found any evidence of the existence of underground cellars; the rooms that answer such purpose are built on the level of the ground. At Tusayan the ancient iiractice of using the ground-floor rooms for storage still prevails. In these are kept the dried fruit, vegetables, and meats that constitute the ijrincipal wiuter food of the Tusayan. Throughout Tu- sayau the walls of the first terrace rooms are not finished with as much care as those above that face the open courts. A quite smoothly fin- ished coat of adobe is often seen in the upper stories, but is much more rarely applied to the rough masonry of the ground-floor rooms. At Zufii no such dift'erence of treatment is to be seen, a result of the lecent departure from their original defensive use. At the present day most of the rooms that are built on the ground have external doors, often of large size, and are regarded by the Zuili as x^referable to the upper terraces as homes. This indicates that the idea of convenience has already largely overcome the traditional defensive requirements of pueblo arrangement. The general finish and quality of the masonry, too, does not vary noticeably in diti'erent portions of the village. An occasional wall may be seen in which underlying stones may be traced through the thin adobe covering, as in one of the walls of the court illustrated iu PL Lxxxii, but most of the walls have a fairly smooth 144 PUEBLO ARCHITECTURE. finish. The occasional examples of rougher masonry do not seem to be confined to any particular portion of tlie village. At Tusayan, on the other hand, there is a noticeable ditt'erence in the extent to which the finishing coat of adobe has been used in the masonry. The villages of the first mesa, whose occupants have come in frequent contact with the eastern pueblo Indians and with outsiders generally, show the eft'ect in the adoption of several devices still unkno\A'n to theii' western neighbors, as is shown in the discussion of the distribution of roof openings in tl>ese villages, pp. 201-208. The builders of the first mesa seem also to have imitated their eastern brethren in the free use of the adobe coat- ing over their masonry, while at the villages of the middle mesa, and par- ticularly at Oraibi, the practice has been comparatively rare, impart- ing an appearance of ruggedness and antitpiity to the architecture. The stonework of this village, perhaps approaches the ancient types more closely than that of the others, some of the walls being noticeable for the frequent use of long bond stones. The execution of the masonry at the corners of some of the houses enforces this resemblance and indi- cates a knowledge of the principles of good construction in the proper alternation of the long stones. A conqjarison with the Kin-tiel masonry (PI. Lxxxix) will show this resemblance. As a rule in pueblo masonry an upper house wall was supported along its whole length by a wall of a lower story, but occasional exceptions occur in both ancient and modern work, where the builders have dared to trust the weight of upper walls to wooden beams or girders, supported along part of their length by buttresses from the walls at their ends or by large, clumsy pieces of masonry, as was seen in the house of Sichumovi. In an upi)er story of Walpi also, partitions occur that are not built immediately over the lower walls, but on large beams supported on masonry piei-s. In the much higher terraces of Zuni, the strength of many of the inner ground walls must be seriously taxed to withstand the superincumbent weight, as such walls are doubtless of only the average thickness and strength of ground walls. The dense clustering of this village has certainly in some instances thrown the weight of two, three, or even four additional stories upon walls in which no provision was made for the unusual strain. The few siqqjorting walls that were accessible to insi>ectiou did not indicate any provision in their thickness for the suj)- port of additional weight; in fact, the builders of the original walls could have no knowledge of their future requirements in this respect. I u the pueblos of the Chaco upper ])artition walls were, in a few instances, supported directly on double girders, two posts of 12 or 14 inches in diameter placed side by side, without reinforcement by stone piers or buttresses, the room below being left wholly unobstructed. This con- struction was practicable for the careful builders of the Chaco, but an attempt by the Tusayan to achieve the same result would probably end in disaster. It was quite common among the ancient builders to divide the ground or storage fioor into smaller nxmis than the floor above, still Xn-eserviug the vertical alignment of the walls. I ft ,/ ) S 5?- -1 •* i: ). i i MINDELKFK. INSIDE WALL FINISH. 145 The fluish of i)ueblo masonry rarely went far beyond the two leading forms, to which attention has been called, the free nse of adobe on the one hand and the banded arrangement of ancient masonry on the other. These types appear to present development along divergent lines. The banded featnre doubtless reached snch a point of development in the Chaco pueblos that its decorative value began to be appreciated, for it is apparent that its elaboration has extended far beyond the requii-e- ments of mere utility. This point would never have been reached had the practice prevailed of covering the walls with a coating of mud. The cruder examples of banded construction, however—those that still kept well within constructional expediency—were dcmbtless covered with a coating ofplaster where they occurred inside ofthe rooms. At Tusayan and Cibola, on the otlier liand, the tendency has been rather to elaborate the plastic element of the masonry. The nearly universal use of adobe is undoubtedly largely responsible for the more slovenly methods of building now in vogue, as it effectually conceals careless construction. It is not to be expecteil that walls would be carefully constructed of banded stonework when they were to be subsequently covered with mud. The elaboration of tlu^ use of adobe and its employment as a periodical coating for the dwellings, probably developed gradually into the use of a whitewash for the hcnise walls, resulting finally in crude attempts at wall decoration. Many of the interiors in Zuni are washed with a coating of white, clayey gypsum, used in the form of a solution made by dissolving in hot water the lum])s of the raw material, found in many localities. The mixture is applied to the walls while hot, and is spread by means of a rude glove-like sack, made of sheep or goat skin, witli the hair side out. With this primitive brush the Zuiii housemves succeed in laying on a smooth and uniform coating over the plaster. An example of tliis class of work was observed in a room of house No. -J. It is ditticult to de- termine to what extent this idea is aboriginal; as now employed it has doubtless been aft'ected by the methods of the neighboring Spanish population, among whom the practice of white-c(«iting the adobe houses inside and out is (jnite common. Several traces of whitewashing have been found among the dilf-dwellings of Canyon de Chelly, notably at the rnin known as Casa Blaiica, but as some of these ruins contained evi- dences of post-Spanish occupation, tlie occurrence there of the white- wash does not necessarily imply any great antiquity for the practice. External nse of this material is much rarer, particnlaily in Zuiii, where only a few walls of ui)per stories are whitened. Where it is not protected from the rains by an overhanging co|>ing or other feature, the finish is not durable. Occasionally where a doorway or other opening has been repaired the evidences of patchwork are o1)literated by a sur- rcninding band of fi-esh jjlastcring, varying in width fiom 4 inches to a foot or more. Usually this band is laid on as a thick wasli of adobe, but in some instances a decorative effect is attained by using white. It -S ETH 10 146 Pl'EBLO ARCHITECTURE. is curious to find that at Tusayan the decorative treatment of the finish- ing wasli has been carried fartlier than at Zufii. The use of a darker band of color about the base of a wliitewaslied room luis already been noticed in the description of a Tusayan interior. On many of the outer walls of upiwr stories the whitewash has been stopped within a foot of the coping, the unwhitened portion of the walls at the top hav- ing tht; eft'ect of a frieze. In a second story house of ^Nfashongnavi, that had been carefully whitewashed, additional decorative effect was pro- duced by tintiug a broad band about the base of the wall with an appli- cation of bright ])inkish day, which was also carried around the door- way as an enframing band, as in the case of tbe Zuni door above de- scribed. The angles on each side, at the junction of tlie broad base band with the narrower doorway border, were ftlled in with a design of alternating |)ink and wliite sipiares. This doorway is illustrated in Fig. 30. Farther nortli, on the same terrace, the jamb of a whitewashed V I Fig. 36. Wall decorations in Masbou^uiivi uxcLuifil i . ou :i white ground. doorway was decorated with the design shown on the right hand side of Fig. 36, executed also iii pink clay. This design closely resembles a ])atfern tlint is commonly embroid('ic(l ui)on the large white ''kachina," or ceremonial blankets. It is not known whctlicr the device is here regardoti as having any special significance. The ])iuk clay in which these designs have been executed has in Sichumovi l)een used for the coating of an entire house front. In addition to the al)ove-mentioned uses of stone and earth in the masonry of house walls, the ])ueblo builders have employed both these materials in a more jyrimitive manner in building the walls of corrals and gardens, and for other purposes. The small terraced gardens of Zuni, located on the borders of the village on the southwest and southeast sides, close to the river bank, are each surrounded by walls li.J or 3 feet high, of very light construction, the average tliickness not exceeding (i or 8 inches. Tliese rude walls are built of small, irregu- larly rounded lumps of adobe, formed by hand, and coarsely plastered with mud. When the crops are gathered in the fall the walls are broken down in places to facilitate access to the inclosures, sotliatthcy re(iuire repairing at each planting season. Aside from this they are so frail as to rei|niie fi'e(|uent re])aiis throughout tlu^ ])eiiod of their use. This method of building walls was adopted because it was the readiest and BUREAU OF ETHNOLOGY EIGHTH ANNUAL REPORT PL. LXIX BUHtAU OF ETMNOLOOr ^'O*^'^" AHNUAl REKMT PI. tXDC Ptt^ MINDEl.KFK.J USE OF LARGE SLABS OF STONE. 147 least laborious means of iuclosiug the required space. The character of tliese garden walls is illustrated iu PI. xc, and their construction with rough luinps of crude adobe shows also the contrast between the weak appearance of this work and the more sutjstautial ett'ect of the masonry of the adjoining uuflnisheil house. At the Cibolaa farming pueblos inclosing walls were usually made of stone, as were also those of Tusayan. PI. LXX indicates the manner in which the material has been used in the corrals of Pescado, located within the village. The stone walls are used in combination with stakes, such as are employed at the main pueblo. Small inclosed gardens, like those of Zuni, occur at several points in Tusayan. The thin walls are made of dry masonry, quite as rude in character as those inclosing the Zuni gardens. The smaller clusters are usually located in the midst of large areas of broken stone that has fallen from the mesa above. In the foreground of PI. xxii may be seen a number of examples of such work. PI. xci illustrates a group of cor- rals at Orailji whose walls are laid up without the use of mud mortar. Where exceptionally large blocks of stone are available they have been utilized in an upright position, and occur at greater or less inter- vals along the thin walls of dry masonry. An example of this use was seen in a garden wall on the west side of Walpi, where the stones had been set on end in the yielding surface of a sandy slo^ie among the foot- hills. A similar arrangement, occurring close to the houses at Ojo Cal- iente, is illustrated in PL xcil Large, upright slabs of stone have been used by the pueblo builders in many ways, sometimes incori)orated into the architectui-e of the houses, and again in detached positions at some distance from the villages. Pis. xciii and xciv, drawn from the ])hotographs of Mr. W. H. Jackson, afford illustrations of this usage in the ancient ruins of Montezuma Canyon. In the first of these cases the stones were utilized, apparently, in house masonry. Among the ruins ill the valley of the San Juan and its tributaries, as described by Messrs. W. H. Holmes and W. H. Jackson, varied arrangements of upright slabs of stone are of frequent occurrence. The rows of stones are some- times arranged iu squares, sometimes in circles, and occasionally are incorporated into the walls of ordinarj' masonry, as in the example illus- trated. Isolated slabs are also met with among the ruins. At K'ia- kima, at a point near the margin of the ruin, occurs a series of very large, upright slabs, which occupy the positions of headstones to a number of small inclosures, thought to be mortuary, outlined upon the ground. These have been already described in connection with the ground ])lan of this village. The emploJ^^lent of upright slabs of stone to mark graves probably prevailed to some extent in ancient practice, but other uses suggest themselves. Occupying a conspicuous point in the village of Kin-tiel (PI. lxiii) is an ujiright slab of sandstone which seems to stand iu its original position undisturbed, though the walls of the adjoining rooms 148 ITKHI.O ARCHITECTURE. ait' ill luiiis. A similar feature was seen at Petiasco Blanco, on the east side of tlie village and a short distance without the inelosing wall. Both these rude ])illarsare, in charaeter and injiosition, very similar to an upright stone nf known use at Zuui. A hundred and fifty feet from this pueblo is a large upright hlock of sandstone, which is said to be used as a datum i)oint in the ol)servations of the sun made by a priest of Ziini for the regulation of the time for planting and harvesting, for determining the new year, and for fixing the dates of certain other ceremonial obserAain^es. By the aid ot such devices as the native priests hav(^ at their command they are enabled to fix the date of the winter solstice with a fair degree of accuracy. Such rude determination of time was probably an aboriginal invention, and may have furnished the motive in other cases for placing stone jiillars in such nnusual ])osi- tions. The explanation of the g(jvernor of Zuui for a sun symbol seen on an upright stime at Mat.saki has been given in the description of that place. Single slabs are also used, as seen in the easternmost room groui) of Tiiaaiyalana, and in the southwestern cluster on the same mesa, in the building of shrines for the deposit of ])lume sticks and other ceremonial objects. An uuusual employment of small stones in an upright position occurs at Zuiii. The inclosing wall of the church yard, still used as a burial l)lace, is provided at intervals along its top with upright pieces of stone set into the joints of a regular coping course that caps the wall. This feature may have some connection with the idea of vertical grave stones, noted at K'iakima. It is difficult to surmise what practical pur- pose could have been subserved by these small ujiright stones. Notwithstanding the use of large stones for special purposes the pueblo builders rarely appreciated the advantages that might be obtained by the proper use of such material. Pueblo masonry is essentially made up of small, often minute, constructional units. This restriction doubt- less resulted in a higher degree of mural tinish than would otherwise have been attained, but it also imposes certain limitations upon their architectural achievement. Some of tliese are noted in the discussion of openings and of other details of construction. PI. XLV, an illustration of a Mormon mill building at Moen-kopi, already referred to in the description of that village, is introduced for the purpose of comparing the methods adopted by the natives and by the whites in the treatment of the same class of material. Perhaps the most noteworthy contrast is seen in the sills and lintels of the openings. HOOFS A.\"0 1 LOOKS. In the i)uel)lo system of tmilding, roof and floor is one; for all the floors, excei)t such as are foinied immediately on the surface of the ground, are at the same time the roofs and ceilings of lower rooms. The jiueblo jjlan of to-day readily admits of additions at any time and aluiost at any point of the basal coustruction. The addition of rooms MINDELEFF.] ZUNI ROOF CONSTEUCTION. 149 above converts a roof into the floor of the new room, so that there can be no distinction in method of construction between floors and roofs, excejjt the floors are occasionally covered with a complete paving of thin stone slabs, a device that in external roofs is confined to the cop- ings that cap the walls and enframe openings. The methods of roofing their liouses practiced by the pueblo build- ers varied but little, and followed the general order of construction that has been outlined in describing Tusayan house building. Tlie diagram shown in Fig. 37, an isometric projection illustrating roof Fig. 37. Diagram of Zufii roof construction. construction, is taken from a Zuni example, the building of which was observed by the writer. The roof is built by first a series of principal beams or rafters. These are usually straight, round poles of 6 or 8 inches in diameter, with all bark and projecting knots removed. Squared beams are of very rare occurrence; the only ones seen were those of the Tusayan kivas, of Spanish manufacture. In recently constructed houses the prnicipal beams are often of large size and are very neatly squared oft' at the ends. Similar scpuire ended beams of large size are met with in the ancient work of the Chaco pueblos, but there the enor- mous labor involved in producing the result with only the aid of stone implements is in keeping with tlie highly finished character of the masonry and the general massiveness of the construction. The same treatment was adopted in Kin-tiel, as may be seen in PI. xcv, which illustrates a beam resting upon a letlge or ottset of the inner walls. The recent introduction of improved mechanical aids has exerted a strong influence on the character of the construction in greatly facilitating execution. The use of the American ax made it a much easier task to cut large timbers, and the introduction of the "burro" and ox greatly facilitated their transportation. In the case of the modern pueblos, such as Zuiii, the dwelling rooms that were built by families so poor as not to have these aids would to some extent indicate the fact by their more jjrimitive construction, and particularly by their small size, iu 150 PUERLO ARCHITECTURE. this respect more closely resenibliiij;- the rooms of the ancient pneblos. As a result the i)()orer classes would he uiore likely to perpetuate i>rimi- tive devices, through the necessity for practicing methods that to the wealthier memhers of the tribe were becoming a matter of tradition only. In such a sedentary tribe as the present Zuni, these ditterences of wealth and station are more marked than one would expect to find among a peoi)le practicing a style of architecture so evidently influenced by the communal ])rinciple, and the architecture of to-day shows the effect of such distinctions. In the house of the governor of Zuni a new room has been recently built, in which the second series of the roof, that aijplied over the principal beams, consisted of pine shakes or shingles, and these supported the final earth covering without any in- tervening material. In the typical arrangement, however, illustrated in the figure, the first series, or principal beams, are covered by another series of snuill poles, about an inch and a half or two inches in diameter, at right angles to the first, and usually laid (piite close together. Tlie ends of these small poles are partially eiabedded in the masonry of tlie walls. In an example of the more careful and laborious work of the ancient builders seen at Pefiasco Blanco, on the Chaco, the principal beams were covered with narrow boards, from 2 to 4 inches wide and about 1 inch thick, over which was put the usual covering of earth. The boards had the apjiearance of having been sjilit out with wedges, the edges and faces having the characteristic fibrous appearance of torn or split wood. At Zuiii an instance occurs where split poles have been iised for the second series of a roof extending through the whole thickness of the wall and i)rojecting outside, as is commonly the case with the first series. A similar arrangement was seen in a ruined tower in the vicinity of Fort Wingate, New Mexico. In the typical roof con- struction illustrated the second series is covered with small twigs or brush, laid in close contact and at right angles to the underlying series, or parallel with the main beams. PI. xcvi, illustrating an unroofed adobe house in ZuTii, shows several bundles of this material on an adjoin- ing roof Tliis series is in turn coxcred with a layer of grass and small brush, again at right angles, wliicli ]ireiinr('s the frame for the receiition of the final earth covfriiig, this latter being tlie fiftli ai>i)lication to the roof. In the example illustrated the entire eartli covering of the roof was finished in a single application of the material. It has been seen that at Tnsayan a layer of moistened earth is applied, followed by a thicker layer of the dry soil. In ancient construction, the method of arranging the material varied somewhat. In some cases series 3 was very carefully constructed of straight willow v\aiids laid side by side in contact. This gave a very neat ai)i)earance to the ceiling v\itliin the room. Examples were seen in Canyon de Ohelly, at Mummy Cave, and at Huugo Pa\ie and Pueblo Bonito on the Chaco. Again exam])les occur where series 2 is comjiosed of L'-inch poles in contact and the joints are chinked on the npper side with small MiNDEi.EFF] WALL COPINGS AND ROOF DRAINS. 151 stones to prevent the earth from sifting tlirough. This arriingenient was seen in a small cluster on the eanyou bottom on the lie (Jhelly. The small size of available roottng rafters has at Tusayau brought about a construction of clumsy piers of masonry in a few of the larger rooms, which support the ends of two sets of main girders, and these in turn carry series 1, or the main ceiling beams of the roof. Tlie girders are generally double, an arrangement that has been often employed in ancient times, as many examples occur among the ruins. The purpose of such arrangement may have been to admit of the abutment of the ends of series 1, when th(> members of the latter were laid in contact. In the absence of squared beams, which seem never to have been used in the old work, this aliutment could only be securely accomplished by the /^ 1 1,1 • 1 1 m• ^'° - Si^- Showing a'oiitment ofdouble smaller roofuse of gn-ders, as suggesti."ed ,,^^,„, 7,„.,, ^„„^j ^^^^^ the following diagram. Fig. ;5S. The tinal roof covering, composed of clay, is usually laid on very care- fully and firmly, and, when the surface is unbroken, answers fairly well as a watershed. A slight slope or fall is given to the roof. This roof subserves every purpose of a front yard to the rooms that o^ien upon it, and seems to be used exactly like the ground itself. Sheepskins are stretched and pegged out upon it for tanning or drying, and tlie char- acteristic Zuui dome-shaped oven is frequently built upon it. In Zufu generally upper rooms are provided only with a mud floor, although occasionally the method of paving with large thiu slabs of stone is adopted. These are often somewhat irregular in form, the object being to have them as large as possible, so that considerable ingenuity is often displayed in selecting the pieces and in joining the irregular edges. This arrangement, similar to that of the kiva floors of Tusayan, is oc- casionally met with in the kivas. In making excavations at Kin-tiel, the floor of the ground room in which the circular door illustrated iu PI. c, was found was paved with large, irregidar fragments of stone, the thickness of which did not aver- age more than an inch. Its floor, whose paving was all in place, was strewn with broken, irregular fragments similar in (character, which nnist have been iised as the flooring of an upper chamber. WALL COPIXUS AXI> KOOF llliAINS. In the construction of the typical pueblo house the walls are carried up to the height of the roof surface, and are then capped with a contin- uous protecting coping of thin flat stones, laid in close contact, their outer edges flush with the face of the wall. This arrangement is still the prevailing one at Tusayan, though there is an occasional exami)le of the projecting coping that practically forms a cornice. This latter is the more usual form at Zufu, though in the farming pueblos of (Jibola 152 PUEBLO AKCHITECTURE. it rotected at all vulnerable points with slabs of stone in order quickly to divert the water and pre- serve the roofs and walls from destruction. LADDERS AND STEPS. In the inclosed court of the old fin-tress pueblos the first terrace was reached only by means of ladders, but the tc^rrai^es or rooms above this were reached both by ladders and steps. The removal of the lower tier of ladders thus gave security against intrusion and attack. The build ers of Tusayan liave preserved this primitive arrangement in mucli greater purity than those of Cibola. In Zuili numerous ladders are seen on every terrace, but the purpose of these, on the highest terraces, is not to provide access to the rooms of tlie upper story, which always have external doors opening on the terraces, but to facilitate repairs of the roofs. At Tusayan, on tlie nilNDELEFF.] LADDERS. 157 other haml, liicklers are of rare occurrence above the first terrace, their phice heiiiji' supplied hy Hights of stone steps. The rehvtive scarcity of stone at Zufii, suitable for building- niaterial, and its great abimilaiice at Tusayan, undoubtedly account for this differeiu'C of usag(\ especially as the proximity of the timber supply of the Zuiii mountains to the Ibriuer facilitates the substitution of wood for steps of luasomy. Fid. 4S. A inoili-ni iioli'lii-cl lailclc-r in Oraibi. Fio. 46. I'u.sHyaii uotched ladders from Masbongnavi. The earliest form of ladder among the puebh)s was i)robably a notched log, a form still occasionally used. Figures 45 aud 4(j illu.strate exam- ples of this type of ladder from Tusayan. 158 PUEBI.O ARCHITECTURE. A notched liiddcr from Orailu, made with a modern axo, is sliowu. This sx't'ciuit'u has a squareness of outline and an evenness of surface not observed in the ancient examples. The hidder from Mashongnavi, illustratecl on the h-ft of Fit;. 40, closely resembk^s the Oraibi s]M'cim<'n, though the workmanship is somewhat ruder. The example illustrated on the right of the same figure is from < )raibi. This ladder is very old, and its ])resent rough and weatherbeaten surface aft'ords but little evi- dence of the character of the implement used in making it. The ladder having two poles connected by cross rungs is undoubtedly a native invention, and was ])robably develo])ed through a series of im- provements on the jirimitive iiotclicd tyi)e. It is described in detail in the earliest Spanish accounts. Fig. 47 illustrates on the left the notched Fig. 47. Aboriginal American forms of ladder. ladder, and on the right a typical two-juile ladder in its most primitive form. In this case the rungs are simi)ly lashed to the uprights. The center ladder of the diagram is a Mandan device illustrated by Mr. Lewis H. Morgan.' As used by the Maiidans this ladder is i)laced witli its forked end on the ground, the rever.se of the Pueblo practice. It will readily be seen, on comparing these examples, that an elongation of the fork which occurs as a ccmstant accom])aiiimeiit of the notched ladilcr might eventually .suggest a construction .similar to that of the Mandaii ladder rever.sed. The function of the fork on the notched ladder in .steadying it when jdaccd against the wall would be more eU'ectually performed by enlarging this feature. ' Cont. to N. A. Ethn., vol. i. Houses and House Life, pp. 129-131. BUREAU OF ETHNOLOGY Eighth annual report Pl. lxxvii OUTLINE PLAN OF ZUNI, SHOWING DISTRIBUTION OF OBLIQUE OPENINGS. MINDELEFF.I LADDERS. 159 At one stage in the (levelopineiit oftlie foiin of ladder in eoininon nse to-day the rungs were hiid in depressions or notches of the vertical i)oles, resembling the larger notclies of the single ladder, and then lashed on with thongs of rawhide or with other materials. Later, when the use of iron became known, holes were burned through the side poles. This is tiie nearly universal i)ractice to-day, though some of the more skillful pueblo carpenters manage to chisel out rectangular lioles. The i)iercing of the side poles, particulary prevalent in Zuni, has brought about a curions departTire from the ancient practice of removing" the ladder in times of threatened danger. Long rungs are loosely slippeil into the holes in the siile pieces, and the security formerly gained by taking up the entire ladder is now obtained, partially at least, by the removal of the rungs. The boring of the side pieces and the employment of loose rungs seriously interferes with the stabihty of the structure, as means nuist be proNnded to prevent the si)rea(ling apart of the side pieces. Tlie Zuni architect has met this difdculty by prolonging the poles of the lad- der and attaching a cross piece near their upper ends to hold them to- gether. As a rule this cross piece is provided with a hole near <>ach end into which the tapering extremities of the poles are inserted. From their high position near the extremities of the ladders, seen in silhouette against the sky, they form peculiarly striking features of Zuni. They arc frequently decorated with rude carvings of terraced notches. Exam- ples of this device may be seen in the views of Zuni, and several tyjjical specimens are illustriited in detail in PI. XCVIII. The use of cross jiieces on ladders emerging from roof openings is not so conmion as on external ones, as there is not the same necessity for holding together the poles, the sides of the opening performing that office. There are two places in Zuni, i)ortions of the densest house cluster, where the needs of unusual traffic have been met by the em])loyment of double ladders, made of three vertical poles, which accommodate two tiers of rungs. The sticks forming the rungs are inserted in continuous lengths through all three poles, and the cross pieces at the top are also continuous, being formed of a single flat piece of wood perforated by three holes for the reception of the tips of the poles. In additional to the usual cross pieces pierced for the reception of the side poles and rudely carved into ornamental forms, many temixnary cross pieces are added during the harvest season in the early autumn to support the strips of meat and melons, strings of red pe])pers, and other articles dried in the open air prior to storage for winter use. At this season every device tliat will serve this purpose is employed. Occasionally poles are seen extending across the reentering angles of a house or are supported on the coping and rafters. The pid|ecting roof beams also are similarly utilized at this season. Zuni ladders are usually provided with about eight rungs, but a few have as many as twelve. The women a.scend these ladders carrying oUas of water on their heads, children i)lay upon them, and a few of the IGO PUP:1?L() ARt'HITf:CTURE. most t'xiK'rt of tlio numerous dogs tliat infest the vilhige can eluinsily make tlieir way up aud down them. As described in a j)revious seetion all houses built during the year are consecrated at a certain season, and among other details of the ceremonial, certain rites, intended to prevent accidents to cliildren, etc., are performed at the foot of the lad- ders. In Tusayan, where sti)ue is abundant, the ladder lias not reached the elaborate development seen in Zuni. The perforated cross piece is rarely seen, as there is little necessity for its adojition. The side jioles are held together by the top and bottom rungs, which pass entirely through the side pieces and ai'e securely fixed, while the ends of the others are only ]iartly embedded in the side pieces. In other cases (IM. XXXII) the ])oles are rigidly held in place by ropes. or rawhide lashings. Short ladders whose side poles are but little prolonged beyond the tiOi) rung are of common occurrence, i)articularly in Oraibi. Three such ladders are shown in PI. Lxxxiv. A similar example may be seen in PI. cvii, in connection with a large opening closed with rough masonry. In these cases the rungs are made to occupy slight notches or depres- sions in the uj)right poles and are then firmly lashed with rawhide, form- ing a fairly rigid structure. This type of ladder is ])robably a survival of the earliest form of the pueblo ladtler. In addition to the high cross j)iece whose function is to retain in place the vertical poles, the kiva ladders are usually provided, both in Zufii and Tusayan, with a cross piece consisting of a round stick tied to the uprights and placed at a uniform height above the kiva. roof. This stick affords a handhold for the masked dancers who are often encumbered with ceremonial paraphernalia as they enter the kiva. In the case of the Oraibi kiva occup.ving the foreground of PI. xxxviii, it may be seen that this handhold cross \neve is inserted into holes in the side poles, an exception to the general practice. In PI. lxxxvii, illustrating kivas, the position of this feature will be seen. The exceptional mode of access to Tusayan ki\a hatchways by means of short flights of stoue steps has already been noticed. In several instaiu'cs the top steps of these short flights cover the thickness of the wall. The remains of a similar stairway were observed in Pueblo Bonito, where it evidently reached directly fi-om the ground to an external doorway. Access by such means, however, is a departure from the original defensive idea. Modern practic(» in Zuni has depaited nioi'c widely from tlie jirimitive system than at Tusayan. In the former puel)lo short Hights of .stone steps giving access to doors raised but a short distance above the ground are very commonly seen. Even in the small farming ])ueblo of Pescado two examples of this arrangement are met with. PL xcix illustrates one of these found on the north outside wall. In the general views of the Tusayan villages the closer adherence to ])riniitivc methods is MlKDELEFF.l STONE STEPS. IGl clearly iudicated, although the modein compare very vinfavorably with the aucieut examples iu precision of execution. PI. xxxii illustrates Fig. 48. Stuue slt^ps at Oraibi, with platform at corner. two flights of stone stejis of Sluipaulovi. Iu many cases the work- manship of these stone steps does not surpass that seen in the Walpi trail, illustrated in PI. xxv. Fig. 49. Stone steiis, with platform at rhimnf.v, iu Oraibi, S ETU H 162 PUEBLO AKCHITECTUKE. Perhaps in no one detail of i)nel)lo construction are the careless and shiftless modern methods so conspicuous as iu the stone steps of the upper terraces of Tusayan. Here are seen many awkward makeshifts by means of which the buihh'rs lia\e tried to compensate for their lack of foresight in planning. The absence of a definite lilan for a house cluster of many rooms, already noted iu the discussion of dwelling- house constructu)n, is rendered conspicuous by the manner in which the stone stairways are used. Figs. 48 and 49 illustrate stone stei)s on upper terraces iu Oi-aibi. In both cases the steps have been added long after the rooms against wliicli they abut were built. In order to conform to the fixed reiptiicmeut of ](lacing such means of access at the corners of the upjier moms, the builders constrru-ted a clumsy jdatform to afford passage aiound the itreviously built chimney. Fig. 50 shows the result of a similar lack of foresight. The upj)er portion of r7£*y-rr'-^;,-i •-'-», Fig. 5U. Stune stejis iu SUumopavi. the flight, consisting of three steps, has been abruptly turned at right angles to the nuiin flight, and is sup^iorted upon rude poles and beams. The restriction of this feature to the corners of upper rooms where they were most likely to conflict with chimneys is undoul)tedly a survival of ancient practice, and due to the necessary vertical alignment of walls and masonry iu this primitive construction. COOKING PITS AND OVENS. Most of the cooking of the ancient Pueblos was probably done out of doors, as amoug the ruins vestiges of cooking pits, almost identical in %f MINDELEFF.l COOKING PITS. 163 character with those .still found in Tusayan, are frequently seen. In Cibola the large donie-shapeil ovens, eouinion to the Puel>los of the llio Grande and to their Mexican neighbors are in general use. In Tu- sayan a few examples of tliis f(n'm of oven occur upon the roofs of the terraces, while the cooking pit in a variety of forms is still extensively used. The distribution of the dome-shaped ovens in Cibola and in Tnsayan may be seen on the ground ])lans in Chapters ill and iv. The simplest form of cooking pit, still connuonly used in Tusayan, ccmsists of a de- pression in the gr(niud, lined with a coating of mud. The pit is usually of small size and is commonly placed at some little distance from the house; in a few cases it is located in a sheltered corner of the building. Fig. 51 illustrates a series of three such primitive ovens built against a house wall, in a low bench or ledge of masonry raised 6 inches above the g '!!Bli||llliUiiSaB4(IB ground; the holes measure and Fig. 51. A series of cooking pits in Maahongnavi.about a foot across are about 18 or 20 inches deep. Many similar jiits occur in the Tusayan villages ; some of them are walled in with upright stone slabs, whose rough edges project 6 or 8 inches above the ground, the result closely resembling the ancient foriu of in-door lirej)lace, such as that seen in a room of Kin-tiel. (PI. c. l-'vi .')•_'. Pi-gumnii ovens of Mashongnavi. In its perfected form the cooking pit in Tusayan takes the place of the more elaborate oven used in Zuui. Figs. 52 and 53 show two speci- Fig. 53. Cross sections ofpi-gummi ovens of ilashongnavi. mens of pits used for the jireparatiou of pi-gummi, a kind of baked mush. 164 PUEBLO ARCHITECTURE. These occur on the east side of Mashoiigna\i. They project 6 or 8 inches above the oTonnrt, and have a dejjth of fi-oin 18 to 24 inches. Tlie debris scattered about the pits indicates the niaiiuer inwliich tliey are covered with shibs of stone and sealed with mud when in use. In all the oven devices of tlie pueblos the interior istirst thorou};hly heated by a long' continiuHl tire within the structure. When the temperature is sufliciently high the ashes and dirt are cknined out, the artich^s to be cooked inserted, and the orifices sealed. The food is often left in these heated rece])tack^s for 12 hours or more, and on lemoval it is gen- erally found to be very nicely cooked. Each of the pigummi ovens illustrated above is jjrovided with a tube-like oritice 3 or 4 inches in diameter, descending obliquely from the ground level into the cavity. Through this o])ening the fire is arranged and kept in ()rder, and in this respect it seems to be the counterpart of the smaller hole of the Zuiii dome-shaped ovens. When the principal opening, by which the vessel containing the pigummi or other articles is introduced, has been covered with a slab of stctne and sealed with mud, the eftect is similar to that of the dome-shaped oven when the ground-opening or doorway is hermetically closed. No example of the dome-shaped oven of ])re-Oolumbian origin has been found among the pueblo ruins, although its prototyjjc probaljly existed in ancient times, ]iossibly in the form of a kiln for baking a fine quality of pottery formerly manufactured. Ilowever, the cooking ])it alone, developed to the point of the pigummi oven of Tusayan, may have been the stem upon whicli the foreign idea was engrafted. In- stances of the complete adoption by these conservative ])eople of a wholly foreign idea or feature of construction are not likely to be found, as improvements are almost universally confined to the mere modifica- ticm of existing devices. In the few instances in which more radical changes are attemjjted the resulting forms bear evidence of the fact. "^ 7 Fig. 54. Diagram showing foundation atones of a Zuui oven. In Cibola the construction of a dome-shaped oven is begun by laying out roughly a circle of flat stones as a foundation. Ui)on these the MINUELEFP-T OVENS. 165 upper strnctnre is rudely built of stones laid in the mud and approxi- mately in tlie courses, tliouj,''li often (luring construction one side will be carried considerably liigiier tlian anntlier. The walls curve inward to an apparently unsafe degree, but the nuid mortar is often allowed to partly dry before carrying the overhanging ])ortion so far as to endanger the structure, and accidents rarely happen. The oven illustrated in PI. xcvii shows near its broken doorway the arrangement of foundation stones referred to. Typical examples of the dome oven occur in the foreground of the general view of Zuni shown in PI. Lxxviii. The dome ovens of Cibola are generally snioothly plastered, inside and out, but a few examples are seen in which the stones of the masonry are exposed. In PI. xrix may be seen two ovens diftering in size, one of wliich shows the manner in which the opening is blocked u|i with stone to keep out stray dogs during periods of disuse. Fig. 55 illustrates a nnul- jilastered oven at Pescado, which is elevated about a foot above the ground on a base or plinth of masonry. The 0])ening of this oven is on the side toward the houses. This form is cjuite exceptional in Cibola, ks! l^"-) «M Fig. 55. Dome-shaped ovei] on a plinth of masonry. tlionfi;ii of frequent oceurrence aimm^ the liio (IraiKl*^ imeblos. Avery huge iiud carefully linished exainjtle was exaiuiiieil at Jeiiiez. 166 PUEBLO ARCHITECTURE. Figs.Sfi and 57 illustrate two spcciinens of rouj^li masonryovens soon at Pesf-ado. In oiieof these a decided liurizuntal arraiiaenient of the stones ..f^ Fig. 51). Oveu iu Pfscado es]>osiug stoues of maaoiiry. ill tlic masonry prevails. The specimen at tlie right is small and rndely constructed, showing but little care in the use of the building material. The few specimens of donu> ovens seen in Tusayan are characterized by the same rudeness of construction noticed in their house nuisonry. The rarit\' of this oven at Tusayan, wliere so man v of the constructions have e^k. Fig. 57. Oven in Pescatlo exposing stoues of ma.sonry. retained a degree of priniitiveness not seen elsewhere, is perhaps an ad- ditional evidence of its foreign origin. UINDBLEFF.] OVENS, ETC. 1(57 r-SHAI'ED STRl'CTURKS. In Tusayan, there arc otlier structure.s, of rude doine-sliape, lilcely to be mistaken for some form of cooking device. Fig. .58 illustrates two specimens of shrines that occur in courts of Mashouguavi. These are Fig. 58. Shrines in M.ishongnavi. receptacles for plume sticks (bahos) and other votive offerings used at certain festivals, which, after being so used, are sealed up with stone slabs and adobe. These shrines occur at several of the villages, as noted in the discnssion of the plans in Chapter iii. In the foreground of PI. XXXVIII may l>e seen an Oraibi specimen somewhat resembling those seen at Mashongnavi. Fig. 59. A ponltry house in Sichumovi resembling an oven. Fig. 59 illustrates a very rude structure of stones in Sichumovi, re- sembling in form a dome oven, which is used as a poultry house. Sev- eral of these are seen in the Tusayan villages. FIRKPLACES AND CHIMNEYS. The original fireplace of the ancient pueblo builders was probably the simple cooking pit transferred to a position within the dwelling room, and employed for the lighter cooking of the family as well as for warm- 1G8 PUEBLO ARCHITF.f'TT'RE. ing' tlie dwelling'. It was jjlacfd in tlie coiiter of tlic floor in oider that the occupants of the house might conveniently g;ither around it. One of the flr.st iniprovenieuts made in this shallow indoor cooking pit must have consisted in surrounding it with a wall of sufficient height to pro- tect the lire against drafts, as seen in the outdoor jiits of Tusayan. In exciivating a room in the ancient ])ueblo of Kin-tiel, a completely pre- served fireplace, about a foot dee]), and walled in with thin slabs of stone set on edge, was brought to liglit. The dejiression had been hollowed out of the solid rock. This tireplace, together with the room in whicli it was found, is illus- trated iu PI. c and Fig. 60. It is of rectangular form, but other ex- Flc. 6U, (JiuuiJil plan of ;iu excavatfd rnom in Kiu-tiel. amples have been found which are circular. Mr. W. 11. Jackson de- scribes a fireplace in a cliff dwelling iu " Echo Cave" that consisted of a circular, basin-like deju-ession 30 inches across and 10 inches deep. Rooms furnishing evidence that fires were made in the corners against the walls are found in many cliff dwellings; the smoke escaped over- head, and the l)lackeued walls aflbrd no trace of a chimney or flue of any kind. The i>ueblo chimney is undoubtedly a post-Spanish feature, and the best forms in use at the present time are jn'obably of very recent origin, thcmgh they are still associated with fireplaces that have departed little from the aboriginal form seen at Kin-tiel and elsewhere. It is interest- ing to note, in this connection, that the ceremony consecrating the house is perfornu'd iu Tusayan before the chimney is added, suggesting that the latter feature did uot form a part of the aboriginal dwelling. II snuDELEFF.] CHIMNEYS. lb'9 In Cibola a few distinct forms of chimney are nsecl at the present time, but in the more remote Tnsayan the cliimney seems to be still in the experimental stage. Numbers of a^\kwar(l constrnetions, varj-iug from the ordinary cooking i)it to the more elaborate hooded strnctnres, testify to the chaotic condition of the cliimney-bnildiug art in the latter province. Before the invention of a chimney hood, and while the primitive fire- place occupied a central position in the floor of tlie room, the smoke probably escaped through tlie door and window openings. Later a hole in the roof provided an exit, as in the kivas of to-day, where cere- monial use has perpetuated an arrangement long since superseded in dwelling-house construction. The comfort of a dwelling room i)ro\ided with this featui'c is sufficiently attested by the jjopularity of the modern kivas as a resort for the men. The idea of a rude hood or flue to facil- itate the egress of the smoke would not be suggested until the fireplace was transferred from the center of a room to a corner, and in the first adopti(m of this de\'ice the builders would rely upon the adjacent walls for the needed support of the constmctional members. Practically all of the chimneys of Tnsayan are placed in corners at the present time, though the Zuiii builders have developed suflicient skill to construct a rigid hood and flue in the center of a side wall, as may be seen in the view of a Zuni interior, PI. lxxxvi. Although the pueblo chimney owes its existence to foreign sugges- tion it has evidently reached its present form through a series of timid experiments, and the proper principles of its construction seem to have been but feebly apprehended by the native builders, particularly in Tusayan. The early form of hood, shown in Fig. 66, was made by plac- ing a short supporting ])ole across the corner of a roimi at a sufficient distance from the fioor and upon it arranging sticks to form the frame work of a contracting hood or flue. The whole construction was finally covered with a thick coating of nmd. This primitive wooden construc- tion has probably been in use for a long time, although it was modified in special cases so as to extend across the entire width of narrow rooms to accommodate " piki " stones or other cumbersome cooking devices. It embodies the principle of roof construction that nuist have been em- l)loyed in the primitive house from which the jjueblo was developed, and practically constitutes a miniatme conical roof suspended over the fireplace and depending upon the walls of the room for support. On account of the carctiil and economical use of fuel l)y these peo])le the light and infiannnable material of which the chimney is constructed does not involve the danger of combustion that would be expected. The l)erfect feasibility of such use of wood is well illustrated in some of the old log-cabin chimneys in the Southern States, where, however, the ar- rangement of the pieces is horizontal, not vertical. These latter curi- ously exem])lify also the use of a miniature section of house constiiiction to form a conduit for the smoke, placed at a sufficient heiglit to admit of access to the fii'e. 170 PUEBLO ARCHITECTURE. A further improvement in tlie chimney was the oonstruction of a corner hood support hy means of two short poles instead of a single piece, thus forming a rectang- ular smoke hood of enlarged capacity. This latter is the most common form in use at the present time in both jiro- vinces, but its arrangement in Tusayan, where it represents the highest achievement of the natives in chimney con- stniction, is much more varied than in Cibola. In'the latter provhice the same form is occasionally' executed in stone. Fig. 01 illustrates a corner hood, in which the crossed ends of the supporting poles ^-. ,, are exposed to view. The W outer end of the lower pole is ''" \ supported from the roof beams or the latter Fig. 61. A cunitrcliimney hou.l withtwusupportiDg "^.V '^ Cord ropC, poks (TiLsayuu). being embedded in the mud plastering with which the hood is finished. The A-ertically ridged character of the suil'ace reveals the underlying construction, in which light sticks have been used as a base for the plaster. The Tusayans say that large sunflower stalks are i)re- ferred for this purpose on account of their lightness. Figs. 63 and 64 show another Tusayan hood of the t}'\^e de- scribed, and in Fig. 69 a large hood of the same general form, sus])endedover a piki-stone, is noticeable for the frank treatment of the sus[)ending cords, which are clearly exposed to view for nearly their entire length. In a chinuiey in a Mashongnavi house, illustrated in Fig. 62, a simple, sharply curved piece of wood has been used for the lower rim of this hood, thus obtaining all the capacity of the two-poled form. The vertical sticks in this example are barely discernible thr(mgh the jHastering, which has been applied with more than the usual de- FlG. 62. A curved cbimney lioud ui* Ma- "(^i-pp of C'lreahouguavi. MrHDELEFF." CHIMNEYS. 171 A curious example illustrating a rudimentary form of two-poled hood is shown in Fig. 63. A straight pole of unusual length is built into the Fia. 03, A Mashongnavi chimDey liood auit walled up lireiilace. walls across the corner of a room, and its insertion into the wall is much farther from the corner on one side than the other. From the longer stretch of inclosed wall ])rotrudes a short pole that joins the prin- cipal one and serves as a support for one side of the chimney-hood. In this case the builder appears to have been too timid to venture on the bolder construction recpiired in the i)erfecte(l two-poled hood. This example probably represents a stage in the development of the higher form. In some instances the rectangular corner hood is not suspended from the ceiling, but is supported from beneath by a stone slab or a piece of wood. Such a chimney hood seen in a house of Shni)aulovi measures nearly i by 5 feet. The short side is su]iported by two stone slabs built into the wall and extending from the hood to the floor. Upon the upper stone rests one end of the wooden lintel supporting the long side, while the other end, near the corner of the room, is held in ])osition by a light crotch of wood. Fig. Gi illustrates this hood ; the plan indicating the relation of the stones and the forked stick to the corner of the room. Fig. 71, illustrating a terrace fireplace and chimney of Shumopavi, shows the employment of similar supports. Corner chimney lioods in Zuiii do not (lifter essentially from the more symmetrical of the Tusayan sijecimens, but they are distinguished by 172 PUEBLO ARCHITECTURE. better tinisli aud by less exposure of tlie framework, having been, like the ordinary masonry, subjected to an unusually free application of adobe. Fig. 64. A chiinuey hootl i»t' Slmpanhivi. The linililers of Tusayau apjiear to have been afraid to add the neces- sary weight of mud mortar to produce this finished effect, the hoods usually showing a vertically ridged or cre- uateil surface, caused by the sticks of the framework showing through the thin nuid i-oat. Stone also is often employed in their construction, aud its use has develoj^ed a large, square-headed tjiie of chimney unknown at Tu.sayan. Tliis is illustrated in Fig. (!.">. This form of hood, projecting some distance beyond its flue, affords s])ace that may be used as a mantel-shelf, an advantage gained only to a very small degree by the forms discussed above. This chimney, as before stated, is built Fig . 65. A seini-dctacbt^d square against one of the walls of a room, and near the chimney hood of ZuQi. middle. AU the joints of these hood.s, and even the juaterial used, arc gener- ally concealed from view by a carefully applied coating of plaster, sup- plemented by a g,\i)sum wash, and usually there is no visible evidence of the manner iu which they are built, liut the construction is little sujierior to that of the simjjle corner hoods. The method of fi'aming the various ty])es of hoods is illustrated in Fig. 06. Tlie exam])le ou th(^ left shows an unplastered wooden hood skeleton. The arrange- ment of the parts iu projecting rectangular stone hoods is illustrated iu the right-hand diagram of the figure. In con.structing such a chimney a thin buttre^is is first built against the wall of sullicieut width aud I o H O MmDELEFK.) CHIMNEYS. 173 height to sui)])ort oiio side of the hood. The opposite side of the hood is supported by a flat stone, firmly set ou edge into the masonry of the Fig. G6. Unplastered Ztifii chimney hoods, illustrating coustructiou. wall. The front of tlie hood is supported by a second flat stone which rests at one end on a rude shoulder in the projecting slab, and at the other end upon the front edge of tlic buttress. It would be quite practi- cable for the pueblo builders to form a notch in the lower corner of the supported stone to rest firmly upon a projection of the supporting stone, but in the few cases in which the construction could be observed no such treatment was seen, for they depended mainly ou the interlocking of the ragged ends of the stones. This structure serves to support the body of the flue, usually with an intervening stone-covered si)ace form- ing a shelf. At the present period the flue is usually built of thin sandstone slabs, rudely adjusted to afford mutual support. The whole structure is bound together and smoothed over with mud plastering, and is finally finished with the gypsum wash, applied also to the rest of the I'oom. Mr. A. F. Bandelier describes "a regular chimney, with mantel and shelf, built of stone slabs," which he found "in the caves of the Eito de los Frijoles, as well as in the clift' dwellings of the regular detached family Iiouse tyi)e,"' which, from the description, must have tilosely resembled the Zuiii chimney described above. Hcnises contain- ing such de\'ices may be quite old, but if so they were certainly reoccu- pied in post-Spanish times. Siu'h dwellings are likely to have been used as places of refuge in times of danger up to a comparatively recent date. Among the many forms of chimneys and fireplaces seen in Tusayan a curious ai)proachto our own arrangement of fireplace and mantel was noticed in a house in Sichumovi. In addition to the principal mantel ledge, a light wooden shelf was arranged against the wall on one side of the flue, one of its ends being supx)orted by an ui)right ])iece of wood with a cap, and the other resting on a peg driven into the wall. This fireplace and mantel is illustrated in Fig. G7. Aside from the iieculiar"guyave" or " piki" baking oven, there is but little variation in the form of indoor fireplaces in Cibola, while in Tu- sayan it appears to have been subjected to about the same mutations ' Fifth Ann. Kept. Arch. Inst. Am., p. 74. 174 PUEBLO ARCHITECTURE. already noted in the ontdoor cooking pits. A serious problem was en- conntered by the Tusayan bnihler when he was called n^ion to con- struct cooking-pit lireplaces, a foot or more deep, in a room of an upper Fio. 67. A fireplace and mantel in Sichumovi. terrace. As it was impracticable to sink the pit into the floor, the nec- essary depth was obtained by walling up the sides, as is shown in Fig. Fig. 68. A secoml-story fireplace in MasliongnaTi. 68, which illustrates a second-story fire])lace in Mashongnavi. Other ex- amples may be seen in the outdoor chimneys shown in Figs. 7;i and 73. MINDELEFF.) CHIMNEYS. 175 A modification of the interior fireplace designed for cooking the thin, paper like bread, known to the Spanish-speaking peoples of tlus region as "gnyave," and l)y the Tnsayan as "piki," is common to both Oibola and Tnsayan, though in the former province the contrivance is more carefully constructed than in the latter, and the surface of the baking stone itself is more highly finished. In the gnyave oven a tablet of carefully prepared sandstone is supported in a horizontal position by two slal)s set on edge and firmly imbedded in the floor. A horizontal flue is thus tbrmed in which the fire is built. The upper stone, whose surface is to receive the thin guyave batter, undergoes during its orig- inal pieparatioTi a certain treatment with fire and piSou gum, and per- haps other ingredients, which imparts to it a highly polished black finish. This operation is usually performed away from the pueblo, near a point where suitable stone is found, and is accompanied by a ceremo- nial, which is intended to prevent the stone from breaking on exposure to the fire when first used. During one stage of these rites the strictest silence is enjoined, as, according to the native account, a single word spoken at such a time would crack the tablet. When the long gnyave stone is in position upon the edges of the back and front stones the fire must be so applied as to maintain the stone at a uniform temperature. This is done by frequent feeding with small bits of sage brush or other fuel. Tlie necessity for such economy in the use of fuel has to a certain extent affected the forms of all the heating and cooking devices. Pig. 69 illustrates a Sichumovi piki Fig. 69. Piki stone and chlnmey hood in Sichnmovi. stone, and Fig. 70 shows the use of the oven in connection with a cooking fireplace, a combination that is not uncommon. The latter ex- 176 PUEBLO ARCHITECTURE. ample if=i from Shumopavi. The ilhisti'ation shoM-s an interesting feature in the use of a primitive andiron or boss to support the cooking pot in Flo. 70. Pilii stone and primitive andirou in Shumopavi. ]iositi()n above the fire. This boss is mo(h'kHl from the same clay as the fireplace floor and is attached to it and forms a ])art of it. Mr. Stephen has collected free specimens of these primitive i)rops which had never been attaclicd to the floor. These were of the rudely coni- cal form illustrated in the figure, and were made of a coarsely mixed clay thoroughly baked to a stouy hardness. Chimneys and firei>laces are often found in Tusayan in the small, re- cessed, balcony-like rooms oi the scc^ond terrace. When a deep cooking- pit is retjuired in such a position, it is obtained by building up the sides, as in th(» indoor firejilaces of upper rooms. Such a flrejilace is illustrated in Fig. 71. A roofed recess which usually occurs at one end of the first terrace, called "tupulii," takes its name from the flat piki oven, the variety of fireplace generally built in these alcoves. The transfer of the flrei)laee from the second-story room to the corner of sui'h a roofed-ter- race alcove was easily accomi)lislicd, and prol)ably led to the occasional use of the cooking-pit, with protecting chimney hood on the open and unsheltered roof. Fig. 7li illustrates a deep cooking-pit on an upper UlNDElXFr.] CHIMNEYS AND FIREPLACES. 177 terrace of Walpi. In tliis instance the cooking pit is very massively built, and in the absence of a sheltering "tupubi" corner is effectually Fig. 71. A terrace fireplace and cliimuey of Shumopavi. protected on three sides by mud-plastered stone work, the whole being capped with the usual chimney pot. The contrivance is placed con- veniently near the roof hatchway of a dwelling room. Fig. 711. A terrace cooking-pit and chimney "f Walpi. The outdoor use of the above-described fireplaces on ui)pei' terraces has apparently suggested the i::iprovemeut of the ground cooking pit 8 ETH- -Iti 178 PUEBLO ARCHITECTURE. in a similar iiiaiiiicr. Several specimens were seen in which the cooking ])it of the ordinary depressed t,vi)e, excavated near an inner (uirner of a house wall, was provided with sheltering masonry and a chimney cap; but such an arrangement is by no means of frequent occurrence. Fig. 73 illustrates an example that was seen on the east side of Slinmojiavi. it will be noticed that in the use of this arrangement on the ground—au Fig. 73 A grouinl tooking-pit of Slininopavi covered with a chiiiiney, arrangement that evidently originated on the terraces—the builders have reveited to the earlier form of excavated pit. In other respects the example illustrated is not distingiiishable from tlie terrace forms above described. In the discussion of the details of kiva arrangement in Tusayan (p. 121) it was shown that the chimney is not used in any form in these cere- monial chand)ers; but the sim|)le roofopciiing forming the hatchway serves as a smoke vent, without the addition of either an internal hood or an external shaft. In the Zufii kivas the smoke also finds vent througli the o])ening that gives access to the chamber, l)ut in the fram- ing of the roof, as is shown elsewhere, some distinction between door and chimney is observed. The roof-hole is made double, one portion accommodating the ingress ladder and the other intended to serve for the egiess of the smoke. The external chimney of the pueblos is a simple structure, and exhibits but few variations from the t^^JC. The original form was undoulitedly a mere hole in the roof ; its use is peri)etTuited in the kivas. Thisi>rim- itive form was gradually improved by raising its sides above the roof, forming a rudimentary shaft. The earlier forms are likely to have been rectangular, the nmnd following and developing later short masonry shafts M'hich were finally given height by the aildition of chimney pots. In Zufii the chimney has occasionally developed into a rather tall shaft, projecting sometimes to a height of 4 or 5 feet above the roof. This is particularly noticeable on the lower terraces of Zufii, the chimneys of M^M'^ I .*., IT It^'W) MINDELEFF.] CHIMNEY-TOPS. 179 the higher rooms being more frequently of the short types jirevalent in the farming pueblos of Cibola and in Tusayan. The tall chimneys found in Zuni proper, and consisting often of four or five chimney pots on a substructure of masonry, are undoubtedly due to the same conditions that have so much influenced other constructional details ; that is, the exceptional height of the clusters and crowding of the rooms. As a result of this the chimney is a more conspicuous feature in Zufii than elsewhere, as will be shown by a comparison of the views of the villages given in Chapters iii and iv. In Tusayan many of the chimneys are quite low, a single pot sur- mounting a masonry substructure not more than (i inches high being quite common. As a rule, however, the builders preferred to use a series of pots. Two typical Tusayan chimneys are illustrated in Fig. 74. Most of the substructures for chimneys in this province are rudely Fig. 74. Tusayan chimneys. rectangular in form, and clearly expose the rough stonework of the masonry, while in Zuni the use of adobe generally obliterates all traces of construction. In both lu'ovinces chimneys are seen without the chimney pot. These usually occur in clusters, simply because the builder of a room or group of rooms preferred that form of chimney. PI. CI illustrates a portion (jf the upper terraces of Zuni where a num- ber of masonry chimneys are grouped together. Those on the highest roof are principally of the rectangular form, being probably a direct development from the square roof hole. The latter is still sometimes seen with a rim rising several inches above the roof surface and formed of slabs set on edge or of ordinary masonry. These upper chimneys are often closed or covered with thin slabs of sandstone laid over them in the same inanner as the roof holes that they resemble. Tlie flrejilaces to which some of them belong appear to be used for heating the rooms rather than for cooking, as they are often disused for long periods dur- ing the summer season. 180 ITEIILO ARCHITECTURE. n. CI also illustrates chimneys iu wliicli pots have been used in cou- uectioii witli masonry bases, and also a round masonry cliimncy. The la tterisimmediately behind the single pot chimney seen in the foreground. On the extreme left of the figure is shown a chimney iuto which fire ])ots have been incorporated, the lower ones being almost concealed from view by the coating of adobe. A similar ert'ect may })e seen in the small chimney on the highest roof shown in I'l. LViii. I'l. Lxxxii shows various methods of using the chimney ])ots. In one case the <-liimney is capped with a reversed large-mouthed jar, the broken l)ottom serv- ing as an outlet for the smoke. The vessel usually employed, for this pnrjiose is an ordinary black cooking i)ot, the bottom being burned out, or otherwise rendeied unfit for houseliold use. Other vessejs are occa- sionally used. n. Lxxxiii shows the use, as the crowning member of the chimney, of an ordinary water jar, with dark decorations on a white ground. A vessel very badly broken is often made to serve in chimney building by skillful use of mud and mortar. To facilitate smoke exit the upper pot is made to overlap the neck of the one below by break- ing out the bottom sufficiently. The joining is not often visil)le, as it is usually coated with adobe. The lower pots of a series are in many cases entirely embedded iu the adobe. The pueblo builder has never been able to construct a detached chim- ney a full story in height, either with or without the aid of chimney jiots; where it is necessary to build such shafts to obtain tlie jiroper diaft he is compelled to rely on the support of adjoining walls, and usually seeks a corner. PI. ci shows a chimney of this kind that has been built of masonry to the full height of a story. A similar exam})le is shown in the foreground of PL lxxviii. In PI. xxii may be seen a chimney of the full height of the adjoining story, but in this instance it is con- structed wholly of i)ots. PI. Lxxxv illustrates a sinular case indoors. The external chimney probably developed gradually from the simple roof opening, as ju'cviously noted. The raised combing about trap- doois or roof holes afforded the first suggestion in this direction. From this develoi)ed the square chimney, and finally the tall round shaft, crowned with a series of pots. The whole chimney, both internal and external, excluding only the primitive fireplace, is ])robably of compar- atively recent (U'igiu, and based on the foreign (Sjianish) suggestion. (iATEWAYS AN1> COVEREl> PASSAGES. Gateways, arranged for defense, occur in many of the more compactly- biult ancient pueblos. Some of the passageways in the modern villages of Tusayan and Cibola resemble these older examples, but most of the narrow passages, giving access to the inner courts of the inhabited villages, are not the result of the defensive idea, but are formed by the crowding together of the dwellings. They occur, as a rule, within the l)Ueblo and not u])on its jx-riphery. Alanyof the terraces now face out- ward and are readied from the outside of tlie pueblo, Ix'ing in marked contrast to the early arraiigement, in which narrow passages to inclose I I? if ^ *'*;!, ;.# '' iF'' * . ' ''^M'i'»' 'IF ''Si n^4* MiNDSLEFF.l GATEWAYS AND COVERED PASSAGES. ISl courts wen! exclusively used for access. In tlie grouud i)laiis nf s(^vei;il villages occupie(l within historic times, but now ruined, vestiges of ojienings arranged on the original defensive plan may be traced. About nudway on the noi'theast side of Awatubi fragments of a stand- ing wall were seen, api)arently the two sides of a passageway to tlie inclosed court of the ]uu'blo. The masonry isnuich broken down, how ever, and no iiulit^ation is aftbrded of the treatment adopted, nor do the remains indicati^ wliether this entrance was originally covered or not. It is illustrated in IM. cii. Other examples of this feature may be seen in the ground plans of Tebugkihu, Chukubi, and Payu])ki (Fig. 7, and I'ls. xii and xiii). In the first of thesis thc^ deep Jaml)s of the opening are cleaiiy de- fined, but in the other two only low mojinds of ddbris suggest the gate- way. In the ancient ('il)olan iiiu'blos, including those on the nu^sa of Tiiaaiyalana, no remains of external gateways have been found; the plans suggest that the disposition of the variou.s clusters approximated somewhat the irregular arrangement of the present day. There are only occasional traces, as of a. continuous defensive outer wall, su(di as those seen at Nutria and Pescado. In the pueblos of the Cibola grtmp, ancient and modern, access to the inner i)ortion of the pueblo was usually attbrded at a numberof points. In the ])uelilo of Kin tiel, however, occurs an excellent example of the defensive gateway. The jambs and cor- ners of the opening are finished with great neatness, as may be seen in the illustration (PI. oiii). This gateway or passage wa,s roofed over, and the rectangular dejiressions for tlu^ recejition of cross-beams still contain short stumps, protected from destruction by the masonry. The masonry over the passageway in falling carried away part of the masonry above the Jamb corner, thus indicating continuity of bond. The ground plan of this ruin (PI. LXiii) indicates clearly the various points at which access to the inner courts was obtained. On the east side a noticeable feature is tiie overlapping of the bimndary wall of the south wing, forming an indirect entranceway. The remains do not indi- cate that this passage, like the one just described, was roofed over. In some cases the modern passageways, as they follow the jogs and angles of adjoining rows of houses, dis]ilay similar changes of direction. In i Shupaulovi, which preserves most distinctly in its plan the idea of theinclosed court, the passageway at the south end of the village changes its direction at a right angle before emerging into the court (PI. xxx). This arrangement was undoubtedly determined by the i)ositioii of the terraces long before the i)assageway was roofed over and built ui)on. PI. XXII shows the south passageway of Waliii; the entrances are made narrower than the rest of the i)assage by building buttresses of masonry at the sides. This was probably done to secure the necessary sui)|)ort for the north and south walls of the upper story. One of the walls, as maybe seen in the illustration, rests directly upon a cross beam, strength- ened in this manner. 182 PUEBLO ARCHITECTURE. One of the smaller inclosed cimits of Zufii, illustrated in PI. Lxxxil, is reaelied by means of two covered passages, bearing some general resemblance to tlie ancient defensive entrances, but these liouses, reached from within the court, have also terraces without. The low passage shown in the figure has gradually been surmounted by rooms, reaching in some cases a height of three terraces above the openings; but the accumulated weight finally jjroved too much for the beams and sustain- ing walls—probably never intended by the builders to withstand the severe test afterwards put ujKtn tliem—andfolhnving an unusually pro- tracted period of wet weather, the entire section of rooms above fell to the ground. This occurred since the sirrveying and ])hotogra]diing. It is rather remarkable that the frail adobe walls withstood-so long the unusual strain, or even that they sustained the addition of a top story at all. In the preceding examples the passageway was covered throughout its length by rooms, but cases occur in both Tusayan and Cibohi in which only iiortious of the roof form the floor of superstructures. PL CIV shows a passage roofed over beyond the two-story portion of the l)uilding for a sufficient distance to form a small terrace, upon which a ladder stands. PI. xxiii illustrates a similar arrangement on the west side of Walpi. The outer edges of these terraces are covered with coping stones and treated in the same manner as outer walls of lower rooms. In Zuni an exami)]c of this form of passage roof occurs be- tween two of the eastern house rows, where the rooms have not been subjected to the close crowding characteristic of the western clusters of the pueblo. DOORS. In Zuiii many rooms of the ground story, which in early times must have been used largely for st(n-age, have been converted into well- lighted, habitable apartments by the addition of external doors. In Tusayan this uiodiflcation has not taken place to an equal extent, the distinctly defensive character of the first terrace reached by removable ladders being still preserved. In this province a doorway on the ground is always provided in building a house, but .originally this space was not designed to be permanent; it was left merely for ccmvenience of passing in and out during the construction, and was built up before the walls were completed. Of late years, however, such doorways are often preserved, and additional small openings are constructed for windows. In ani'ient times the larger doorways of tln^ upjter terraces were probably never closed, except by means of blankets tty rabbit-skin robes hung over them in cold weather. Examples have been seen that seem to have been constructed with this object in view, for a slight pole, of the same kind as those used in the lintels, is built into the masonry of the jambs a few inches below the lintel ijroper. Openings imperfectly closed against the cold and wind were naturally placed in the lee walls to avoid the prevailing southwest winds, and the ground plans of the exposed mesa villages were undoubtedly influenced by this circumstance, MINDELEFF] DOORS. 183 the tendency being to change them from the early inclosed court type and to place the houses in longitudinal rows tacing eastward. This is noticeable in the plans given in Chapter li. Doorways closed with masonry are seen in many ruins. Possibly these are an indication of the temporary absence of the owner, as in the harvest season, or at the time of the destruction or abandonment of the village; but they may have been closed for the ])urpose of economi/.ing warmth and fuel during the winter season. No provision was uuule for closing them with movable doors. The practice of fastening up the doors during the harvesting season prevails at the present time among the Zuni, but the result is attained witliout great difficulty by means of rude cross bars, now that they have framed wooden doors. One of these is illustrated in Fig. 7.1. These doors are usually opened by a latch- string, which, when not hung outside, is reached by means of a small round hole through the wall at the side of the door. Through this hole the owner of the house, on leaving it, secures the door by props and braces on the inside of the room, the hole being sealed up and plastered in the same manner "that other openings are treated. This curious arrangement attbrds another illustration of the survival of ancient methods in modified forms. It is not employed, however, in closing the doors of the first terrace; these are tasteued by barring from the inside, the exit being made by means of internal ladders to the ter- race above, the upper doors only being fastened in the manner illus- Fio. 75. A barred Zuni door. trated. In PI. lxxix may be seen good examples of the side hole. Fig. 75 shows a barred door. The plastering or sealing of the small side 184 PUEBLO ARCHITECTURE. hole instead of the entire oi)eiiing was brought about by tlie introduc- tion of the "wooden door, which in its i>resent paneled form is of foreign introduction, but in this, as in so many other cases, some analogous feature which facilitated tlit> adoption of the idea probably already ex- isted. Tradition points to the early use of a small door, made of a single slab of wood, that closed the small rectangular wall niches, in which valuables, such as tur(|Uoise, shell, etc., were ke])t. This slab, it is said, was reduced and smoi>tlic(l by rubbing with a piece of sandstone. A number of beams, rafters, and roofing planks, seen in the Chaco pueblos, were ])robably S(iuare(l and finished in this way. The latter examples show a degree of familiarity with this treatment of wood that would enable the builders to construct such doors with ease. As yet, however, no examjdes of wooden doors have been seen in any of the pre-Columbian ruins. The jiueblo tyj)e of paneled door is much more frequently seen in Cibola than in Tusayan, and in the latter province it does not assume the variety of treatment seen in Zuni, nor is the work so neatly executed. The views of the modern pueblos, given in Chai>ters ill and IV, will indicate the extent to which this fea- ture occuis in the two groups. In the construction of a paneled door the vertical stile on one side is prolonged at the top and bottom into a rounded pivot, which works into cup-like sockets in the lintel and sill, as illustrated in Fig. 70. The hinge is thus produced in the wood itself without the aid of any external appliances. It is difficult to trace the origin of this device Fio. 76. Wooden pivot amoug tlic pucblos. It clo-sely resemblcs the i>i vot hinges of .1 Zuoi door. liiugcs sonu'tiiues Used in niedia'val Eurojie in counection with massive gates for closing masonry jiassages; in such cases the prolonged pivots worked in cavities of stone sills and lintels. The Indians claim to have emjiloyed it in very early times, but no evi- dence on this jioint has been found. It is quite possil)le that the idea was borrowed from some of the earlier Mormon settlers who came into the country, as these people use a number of primitive devices which are undoubtedly survivals of methods of construction once common in the countries from which they came. Vestiges of the use of a pivotal hinge, constructed on a much more massive scale than any of the pueblo examples, were seen at an old fortress-like, stone storehouse of the Mormons, built near flic site of Moeu-kopi by the first Mormon settlers. The paneled dooi- now in use among the pueblos is rudely made, and consists of a tVamc inclosing a single i)aiie]. This panel, when of large size, is occasionally made of two or more pieces. These doors vary greatly in size. A few leacli the height of o'feet, but the usual height ; MTNDELEFF.] PANELED WOODEN DOORS. 185 is from 3J to 4 feet. As doors are commonly elevated a foot or more above the ground or floor, the use of such openings does not entail the ftill degree of discomfort th;it the small size suggests. Doors of larger size, with sills raised but an inch or two above the floor or groiuid, have recently been introduced in some of the ground stcn-ies in Zufii but these are very recent, and the idea has been adopted only by the most i^rogressive people. Fig. 77. Paneled wooden doora in Hano. PI. XLi shows a small paneled door, not more than a foot square, used as a blind to close a back window of a dwelling. The smallest exami)les of paneled doors are those emi)loyed for closing the small, square open- 186 PUEBLO ARCHITECTURE. ings iu tlie back walls of house rows, wbich still retain the defensive arrangement so marked in many of the ancient pnehlos. In some instances doors occur in the second stories of uiiterraced walls, their sills being 5 or 6 feet above the ground. In such cases the doors are reached by ladders whose ui)])er ends rest upon the sills. Elevated openings of this kind are closed in the usual manner with a rude, siugle- lianeled door, which is often whitened with a coating of clayey gyi>sum. Carefully worked ])aneled doors are nuicli more common in Zuni than in Tusayan, and within the latter jirovince the villages of the first mesa make more extended use of this type of door, as they have come into more intimate contact with their eastern brethren than other ^^llages of the group. Fig. 77 illustrates a portion of a Hano house in which two wooden doors occur. These specimens indicate the rudeness of Tusayan workmanship. It will be seen that the workman who framed the upper one of these doors met with considerable difficulty in jiroperly joining the two boards of the panel and in connecting these with the frame. The figure shows that at several poiuts the door has been reenlbrced and strengthened by buckskin and rawhide thongs. The same device has been employed in the lower door, both in fastening together the two pieces of the panel and in attaching the latter to the framing. These doors also illustrate the customary manner of barring the door during the absence of the occupant of the house. The doorway is irsually framed at the time the house is built. The sill is generally elevated above the ground outside and the floor inside, and the door openings, with a few exceptions, are thus practically only large windows. In this respect they follow the arrangement character- istic of the ancient pueblos, in which all the larger openings are window-like doorways. These are sometimes seen on the court margin of house rows, and frequently occnr between commuid- cating rooms within the cluster. Theyare usually raised about a foot and a half above the floor, and in some cases are provided with one or two •^ steps. In Zuni, doorways between communicat- ing rooms, though now framed in wood, preserve Framino; of Zuni the same arrangement, as may be seen in PI. (loor-panel. LXXSVI. The side pieces of a paneled pueblo door are mortised, an achieve- ment far beyond the aboriginal art of these people. Fig. 78 illustrates the manner in which the framing is done. All the necessary grooving, and the i)reparatioii of the inojecting tenons is laboriously executed with the most primitive tools, in many cases the whole frame, with all its joints, being cut out with a small knife. Doors are usually tastened by a simple wooden latch, the bar of which turns upon a wooden pin. They are opened from without by lifting the MINDEI.EFF.] DOORWAYS. 187 latch from its wooden catch by means of a string passed thioiiij;!! a snuill hok^ in the door,and lianj^iiiji' outside. Some few doors are, however, pro- vided with a cimibersome wooden lock, ojierated by means of a square, notched stick that serves as a key. These locks are usually fastened to the inner side of the door by thon;.;'s of buckskin or rawhide, ])assed through small holes bored or drilled through the edge of the lock, and through the stile and panel of the door at corresponding points. The entire mechanism consists of wood and strings joined together in the rudest manner. Primitive as this de\'ice is, however, its conception is far in advance of the aboriginal culture of the pueblos, and both it and the string latch must have come from without. The lock was probably a contrivance of the early Mormons, as it is evidently roughly modeled after a metallic lock. Many doors hiiving no permanent means of closure are still in use. These are very common in Tusayan, and occur also in Cibola, i>articu- larly in the farming pueblos. The open front of the "tupubi" or bal- cony-like recess, seen so frequently at the ends of flrst-terrace roofs in Tusayan, is often constructed with a transom-like arrangement in con- nection with the girder supi)orting the edge of the roof, in the same manner in which doorways proper aie treated. PI. xxxii illustrates a balcony in which one bounding side is formed by a flight of stone steps, producing a notched or terraced effect. The supporting girder iu this instance is embedded in the wall and coated over with adobe, obscuring the construction. Fig. 79 shows a rude transom over the sui)porting beam of a balcony roof in the principal house of Hauo. The upper doorway shown in this house has been partly walled in, reducing its size somewhat. It is also provided with a small horizontal ojiening over the main lintel, which, like the doorway, has been partly filled -n-ith masonry. This upper transom often seems to have resulted from carry- ing such openings to the full height of the story. The transom probably originated from the spaces left between the ends of beams resting on the main girder that si>anned the i)rincipal opening (see Fig. 81). Some- what similar balconies are seen in Cibola, both in Zufii and in the farm- ing villages, but they do not assume so much importance as in Tusayan. An example is shown in PI. ci, in which the construction of this feature is clearly visible. In the remains of the ancient pueblos there is no evidence of the use of the half-open terrace rooms described above. If such rooms existed, especially if constructed in the open manner of the Tusayan examples, they must have been among the first to succumb to destruction. The comparative rarity of this feature in Zuiii does not necessarily indicate that it is not of native origin, as owing to the exceptional manner of clustering and to prolonged exposure to foreign influence, this i)ueblo exhibits a wider departure from the ancient type than do any of the Tusayan villages. It is likely that the ancient builders, trusting to the double i)rotection of the inclosed court and the defensive first terrace, 188 PUKIiL( » AIJCHITECTURE. freely adopted this open and convenient arrangement in connection witli the upper roofs. Fig. 79. Kmlr l i:iiitsoiMS nvir 'l'ns;i\ itn Mpi-iiiD^M. The transom-like opening coninionly accompanying the large oj)ening is also seen in many of the inclosed doorways of Tnsayan, but in some of these cases its origin can not be traced to the loof constrnctions, as the ojienings do not approach tlie ceilings of tlie rooms. In early days snch doorways were closed by means of large slabs of stone set on edge, and these weie sometimes siqiidciiiciitcd by a suspended blanket. Tn severe winter Mcatlier many of the openiiig.s were closed witli masonry. At the present time many doorways not provided with |iaiieh'd doors H O O o MINDELEFK. DOOKWAV.S AND TRANSOMS. 189 are closed in such ways. W'licu a diMirway is tliiis treated its transom is l(^fY ojicn for tlie admission of Iij;lit and air. The Indians state tliat in early times this transom was provided for tlie exit of smoke when the Fl(i. 80. A lar])orts a roof-beam; this is shown in the figure. Other examples of transoms may be seen in connection with many of the illustrations of Tusayan doorways. Fiu. si. a (lo.)rwa.v an inches below the main lintel, and fixed into the masonry in a position to serve as a curtain pole. Originally this pole undoubtedly served as a means of suspension for the Idanket or skin rug used in closing the opening, just as such means are now used in the huts of the Navajo, as well as 190 PUEBLO ARCHITECTUKE. occasionally in the houses of Tusayaii. The space above this cross stick answered the same purpose as the transoms of the present time. A most striking feature of doorways is the occasional departure from the quadranfruhir form, seen in some ruined villages and also in someof the modern houses of Tusayan. Fig. 82 illustrates a specimen of this type found in a small cliff ruin in Canyon de Chelly. Ancient examples of this form of opening are distinguished by a symmetrical dispo- sition of the step in the jamb, while the modern doors are seldom so arranged. A modern exami)le from Mashongnavi is shown ""•^iSi^^^, in Fig. 83. Tliis opening also illustrates the Fia. 82. An ancient doorway in double or divided transom. The beam ends Canyon do Clielly cliff ruin. shown in the figure jiroject beyond the face of the wall and support an overhanging coping or cornice. A door- like Avindow, approximating the symmetrical form described, is seen FlQ. 83. A symmetrically notched doorway in Mashongnavi. immediately over the passage-way shown in PL xxii. This form is evidently the result of the partial closing of a larger rectangular opening. Fig. 84 shows the usual type of terraced doorway in Tusayan, in which one jamb is stepped at a consideral)ly greater height than the other. In Tusayan large openings occur in which only one jamb is stepped, jtroducing an effect somewhat of tliat t)f the large balcony openings with tiiglits of stone steits at om^ side, previously illustrated. An opening of this form is shown in Fig. 8.5. Both of the stepjied door- ^^^'^ *^' V '' t ^* v^A^* -a ^Jjfijiu. .^^e' ^. Wii Af, >•^ '^Tv'^ ^^^^^''*" ^^^ i ^^-^'^Sfl \ vi / 'iff' '; .„ if','"' '^Ka" MINDELEFF.] DOORWAYS. 191 ways, illustrated above, are provided with transom openings extending from one roof beam to another, lu the absence of a movable door the openings were made of the smallest size consistent with convenient use. The stepped form was very likely suggested by the temporary partial blocking up of an opening with loose, flat stones iu such a manner as Fig. 84. A Tusayan notohod doorwav- to least impair its use. This is still quite commonly done, large open- ings being often seen in which the lower portion on one or both sides is narrowed by means of adobe bricks or stones loosely piled up. Iu this connection it may be noted that the secondary lintel pole, previously described as occurring in both ancient and modern doorways, serves the additional purpose of a hand-hold wheu supplies are brought into the house on the backs of the occuitauts. The stepping of the doorway, while diminishing its exposed area, does not interfere with its use in bringing in large bundles, etc. Series of steps, picked into the faces of the clitts, and aftordiug access to clift' dwellings, frequently have a supplementary series of narrow and deep cavities that fiu-nish a secure hold for the hands. The requirements of the precipitous environment of these people have led to the carrying of loads of produce, fuel, etc., on the back by means of a suspending band passed across the forehead; 192 PUEBLO ARCHITECTURE. this left the hands free to aid iu the dititiiult task of dimbiug. These couditioiis seem to liave Ijrought alxiut the use, in some eases, of liaiid- holds ill the luargiual frames of interior trapdoors as an aid in elimb- ing the ladder. Fig. 85. A large Tusayan doorway with one notched janih. One more charaeteristic type of the aneient pueblo doorway remains to be deseribed. During the autumn of 1883, when the ruined jmeblo of Kin-tiel was surveyed, a number of excavations were made in and about the pueblo. A small room on the east side, near the brink of the arroyo that traverses the ruin from east to west, was completely cleared out, exposing its fireplace, the stone pa\ing of its floor, and other details of construction. Built into an inner partition of this room was found a large slab of stone, pierced with a circular hole of sufficient size for a man to squeeze tliroiigli. This slab was set on edge and incorporated into the masonry of the partition, and evidently served as a means of communication with another room. The position of this doorway and its relation to the room in which it occurs may be seen from the illustration in Tl. 0, which shows the stone in situ. The doorway or "stone-dose" is shown in Fig. 8(5 on a sufficient scale to indicate the degree of technical skill in the architectural treatment of stone i)ossessed by the builders of this old pueblo. The \\Titer visited ZuFii in October of the same season, and on describing this ffiid to Mr. Frank II. Cashing, learned that the Zuni Indians still preserved tradi- tional knowledge of this device. Mr. Gushing kindly furnished at the 1 ,4 K ^ \ ,j3lC>i .}iM^. ' fi'^4i t^ J. a > o o i ^^^1 . ' fU ?r ^i^ ^ tflNDELEFF.] DOORWAYS. 193 time the following extract from the tale of "The Deer-Slayer and the Wizards," a Zufii folk-tale of the early occupancy of tlic valley of Zuui. "'How will they enter V said the young man to his wife. 'Through the stone-close at the side,' she an- swered, lu the days of the ancients, the doorways were often made of a great slab of stone with a round hole cut through the middle, and a round stone slab to close it, which was called the stone-close, that the en- emy might not enter in times of war." Mr. Gushing had found displaced fragments of such circular stone doorways at ruins some distance F"»-86- ^n-^i^ient 'i""i'i> 'i'""»'>y or "stone- ,1 ,/. rr ^' \ J. 1 J -L. close" in Kin-tiul.northwest from Zuni, but had been under the impression that they were used as roof openings. All exam- ples of this device known to the writer as having been found in place occurred iu side walls of rooms. Mr. B. W. Nelson, while making collec- tions of pottery from ruins near Springerville, Arizona, found and sent to the Smithsonian Institution, in the autumn of 1884, "a flat stone about 18 inches square with a round hole cut in the midille of it. This stone was taken from the wall of one of the old ruined stone houses near Springerville, iu an Indian ruin. The stone was set in the waU between two inner rooms of the ruin, and evidently served as a means of com- munication or jjerhaps a ventilator. I send it on mainly as an example of their stone-working craft." The position of this feature in the exca- vated room of Kin-tiel is indicated on the ground plan, Fig. 00, which also shows the position of other details seen iu the general view of the room, PI. c. A small fragment of a " stone-close" doorway was found incorporated into the masonry of a flight of outside stone steps at Pescado, indicat- ing its use in some neighboring ruin, thus bringing it well within the Cibola district. Another point at which similar remains have been brought to light is the pueblo of Halona, just across the river from the present Zuni. Mr. F. Webb Hodge, recently connected with the Hemen- way Southwestern Archeological Exposition, under the direction of Mr. F. H. Gushing, describes this form of opening as being of quite common occurrence in the rooms of this long-buried puelilo. Here the doorways are associated with the round slabs used for closing them. The latter were held in place by props within the room. No slabs of this form were seen at Kin-tiel, but quite possibly some of the large slabs of nearly rectangular form, found within this ruin, may have served the same purpose. It would seem more reasonable to use the rectangular 8 ETH 13 194 PUEBLO ARCHITECTUEE. slabs for this purpose when the openings were conveniently near the floors. No example of the stone-elose has as yet been found in Tusayan. The annular doorway described above afl:brds the only instance known to the writer where access openings were closed with a rigid device of aboriginal invention; and fvoni the character of its material this device was necessarily restricted to openings of small size. The larger rect- angular doorways, when not partly closed by masonry, probably were covered only with blankets or skin rugs suspended from the lintel. In the discussion of sealed windows modern examples resembling tlie stone- close d(»vice will be noted, but these are usually employed in a more permanent manner. The snnill size of the ordinary pueblo doorway was iierhaps due as much to the fact that there was no convenient means of closing it as it was to defensive reasons. Many primitive habitations, even quite rude ones built with no intention of defense, are characterized by small doors and windows. The planning of dwellings and the distribution of open- ings in such a manner as to protect and render comfortable the inhabited rooms implies a greater advance in architectural skill than these build- ers bad achieved. The inconveniently small size of the doorways of the modern x>ueblos is only a survival of ancient conditions. The use of full-sized doors, admitting a nmn without stooping, is entirely iiracticable at the present day, but the conservative builders persist in adhering to the early tj'pe. The ancient i^osition of the door, with its sill at a considerable height from the ground, is also retained. From the absenceof any convenient means of rigidly closing the doors and windows, in early times external openings were restricted to the smallest practicable dimensions. The convenience of these openings was increased without altering their di- mensions by elevating them to a certain height above the ground. In the ruin of Kin-tiel there is marked uniformity in the height of the openings aliove the ground, and such openings were likely to be quite uniform when used for similar purposes. The most common elevation of the sills of doorways was such that a man could readily step over at one stride. It vnW be seen that the same economy of space has effected the use of windows in this system of architecture. WINDOWS. In the pueblo system of building, doors and windows are not always clearly differentiated. Many of the openings, while used for access to the dwellings, also answer all the piu-poses of windows, and, both in their form and in their pt)sition in the walls, seem more fidly to meet the recjuirements of openings for the adnussion of light and air than for access. We have seen in the illustrations in Cluipters iii and iv, openings of considerable size so located in the face of the outer wall as to unfit them for use as doorways, and others whose size is wholly in- adequate, but which are still i)rovided \\ith the tji)ical though diminu- f>':-^^^^^- MINDELErF.] WINDOWS. 195 tive single-paneled door. Many of tbese small openings, occurring most frequently in the back walls of house rows, liave the jambs, lin- tels, etc., characteristic of the typical modern door. However, as the drawings above referred to indicate, there are many openings concern- ing the use of which there can be no doubt, as they can only i>rovide outlook, light, and air. In the most common form of window in jiresent use in Tusayan and Cibola the width usually exceeds the height. Although found often in what appear to be the older x'ortious of the present jjueblos, this shape probably does not date very far back. The windows of the ancient pue- blos were sometimes square, or nearly so, when of small size, but when larger they were never distinguishable from doorways in either size or finish, and the height exceeded the width. This restriction of the width of openings was due to the exceptionally small size of the building stone made use of. Although larger stones were available, the builders had not sufficient constructive skill to successfully utilize them. The failure to utilize this material indicates a degree of ignorance of mechanical aids that at first thought seems scarcely in keeiiing with the niassiveness of form and the high degree of finish characterizing many of the remains ; but as already seen in the discussion of masonry, the latter results were attained by the patient industry of many hands, although laboring with but little of the spirit of cooperation. The narrowness of the largest doors and windows in the ancient pueblos suggests timidity on the part of the ancient builders. The apparently bolder construction of the jiresent day, shown in the prevailing use of horizontal oi)enings, is not due to greater constructive skill, but rather to the markedly greater carelessness of modern construction. The same contrast between modern and ancient practice is seen in the disposition of openings in walls. In the modern pueblos there does not seem to be any regularity or system in their introduction, while in some of the older pueblos, such as Pueblo Bonito on the Chaco, and others of the same group, the arrangement of the outer openings ex- hibits a certain degree of symmetry. The accompanying diagram, Fig. 87, illustrates a portion ofthe northern outer wall of Pueblo Bonito, Fig. 87. DiagTam illustrating 8_\-iumetrical arrangement of small openings in Pueblo Bonito. in wliicli the small windows of successive rooms, besides being niiiform in size, are grouped in pairs. The degree of technical skill shown in the execution of the masonry about these openings is in keeping with the precision with which the openings themselves are placed. PL cv, gives a view of a portion of the wall containing these openings. 196 PUEBLO ARCHITECTURE. In marked fontraist to the above examiiles is the slovenly practice of the modern i)neblos. Thert^ are rarely two openings of the same size, even in a single room, nor are these usnally placed at a uniform height fi'om the floor. The placing appears to be purely a matter of individual taste, and no trace of system or uniformity is to be found. Windows occur sometimes at considerable height, near or even at the ceiling in some cases, whUe others are placed almost at the base of the wall ; ex- amples may be found occupying all inteririediate heights between these extremes. Many of the illustrations show this characteristic irregu- larity, but Pis. Lxxix aud Lxxxii of Zuiii perhaps represent it most clearly. The ft'aming of these openings differs but little from thal> of the an- cient examples. The modern opening is distinguished ^jrincipally by the more careless method of combining the materials, and by the intro- duction in nuiny instances of a rude sasli. A numljcr of small poles or sticks, usually of cedar, with the bark peeled off, are laid side by side in contact, across the opc^ning, to form a support for the stones and earth of the superposed masonry. Frequently a particularly large tablet of stone is placed imm(>dia-tely upon the sticks, but this stone is never long enough or thick enough to answer the ijurpose of a lintel for larger openings. The nximber of small sticks used is sufficient to reach from the face to the back of the wall, and in the simplest openings the sur- rounding masonry forms jambs and sill. American or Spanish in- fluence occasionally shows itself in the employment of sawed boards for lintels, sills, and jambs. The wooden features of the windows exhibit a curiously light and flimsy construction. A large ijercentage of the windows, in both Tusayan and Cibola, are furnished with glass at the i)resent time. Occasionally a jn'imitive sash of several lights is found, but fre(pieiitly the glass is used singly; in some instances it is set directly into the adobe without any intervening sasli or frame. In several cases in Zufa the primitive sash or frame has been rudely decorated with incised lines and notches. An example of this is shown in Fig. 88. The frame or sash is usually built solidly into the wall. Hinged sashes do not seem to have been adopted as yet. Often the introduction of lights shows a curious and awkward com- pronuse between aboriginal methods and foreign ideas. Characteristic of Zuiii win- dows, a7id also of those of the neighboring imeblo ofAcoma, Fig. 88. Inciseil .Ifcoiatioi. on a ruilo window sasli in ZuBi. jy ^Jjg use of SemitransluCCnt slabs of selenite, about 1 inch in thickness and of irregular form. Pieces are occasionally met with about 18 inches long and 8 or 10 inches MINEELEFF.] WINDOW GLAZING. 197 wide, but usually they are much smaller and very irregnilar in outline. For wiudows pieces are selected that approximately fit against each other, and thin, flat strips of wood are fixed in a vertical position in the openings to serve as supports for the irregular fragments of selenite, which could not be retained in place without some such provision. The use of window openings at the bases of walls probably suggested this use of vertical sticks as a support to slabs of selenite, as in this position they would be particularly useful, the windows being generally arranged on a slope, as shown in Fig. SO. Similar glazing is also employed in the related, obliquely pierced openings of Zuiii, to be described later. -II Fig. 89. Sloping selenite window at base of Zuiii wall on upper terrace. Selenite, in all probability, was not used iu lire-Spauish times. No examples have as yet been met with among ruins in the region where this material is found and now used. Tliroughout the south and east portion of the ancient pueblo region, explored by Mr. A. F. Bandelier, where many of the remains were iu a very good state of i^reservation, no cases of the use of this substance were seen. Fig. 90 illustrates a typical selenite window. Fig, 90. A Ziifii window glazeil with selenite. In Zuni some of the kivas are provided with small external windows framed with slabs of stone. It is likely that the kivas would for a long time perpetuate methods and practices that had been superseded in the construction of dwellings. The use of stone jambs, however, would necessarily be limited to openings of small size, as such use for large openings was beyond the mechanical skill of the pueblo builders. 198 PUEBLO ARCHITECTUEE. Fig. 91 illustrates the manner of making small openings in external exposed walls in Ziiiii. Stone frames occur only occasionally in what seem to be the older and least modified ijortions of the village. At Tusayan, however, this method of framing windows is much more notice- able, as the exceptional crowding that has exercised such an influence on Zuiii construction has not occurred there. The Tusayan houses are arranged more in rows, often with a suggestion of large inclosures resembling the courts of the ancient pueblos. The inclosures have not been encroached ujion, the streets are wider, and altogether the earlier methods seem to have been retained in greater purity than in Zuiii. The unbroken outer wall, of two or three stories in height, like the same feature of the old villages, is pierced at various heights with small open- ings that do not seriously impair its efficiency for defense. Tusayan examples of these loop-hole-like openings may be seen in Pis. xxu, XXIII, and xxxix. Fig. 91. Small openings in the back wall of a Zufii bouse-clnster. In some of the ancient pueblos such openings were arranged on a dis- tinctly defensive plan, and were constructed with great care. Openings of this type, not more than i inches square, pierced the second story outer wall of the pueblo of Wejegi in the Chaco Canyon. In the pueblo of Kiu-tiel (PI. LXiii) similar loop-hole-like openings were very skill- fully constructed in the outer wall at the rounded northeastern corner of the pueblo. The openings pierced the wall at an oblique angle, as shown on the plan. Two of these channel-like loopholes may be seen in PI. Lxv. This figure also shows the carefully executed jamb corners and fiices of three large openings of the second story, which, though greatly undermined by the falling away of the lower masonry, are still held in position by the bond of thin flat stones of which the wall is built. It is often the practice in the modern pueblos to seal up the windows of a house with masonry, and sometimes the doors also during the tem- porary absence of the occupant, which absence often takes place at the seasons of planting and harvesting. At such times many Zuni families occupy outlying fanning pueblos, such as Nutria and Pescado, and the MTNDELEFF.] SEALED OPENINGS. 199 Tiisayans, in a like manner, live in rnde snmmer shelters close to their fields. Such absence from the home pueblo often lasts for a month or more at a time. The work of closing the opening is done sometimes in the roughest manner, but examjiles are seen in which carefully laid masonryhas been used. The latter is sometimes plastered. Occasionally the sealing is done with a thin slab of sandstone, somewhat larger than the opening, held in place with mud plastering, or propped from the inside after the manner of the " stone close" j)reviously described. Fig. 92 illustrates specimens of sealed openings in the village of Hauo of Fig. 92. Sealed openings in Tusayan. the Tusayau group. The upper window is closed with a single large slab and a few small chinking stones at one side. The masonry used in closing the lower opening is scarcely distinguishable from that of the adjoining walls. PI. CVI illustrates a similar treatment of an opening in a detached house of Kutria, whose occupants had returned to the home pueblo of Zuni at the close of the harvesting season. The door- way in this case is only partly closed, leaving a window-like aperture at 200 PUEBLO ARCHITECTURE. its top, and the stones used for the |mi))os(> arc simjily lu'lcd up without tlio use of adobe mortar. Windows and doors elosed with masonry are often met with in the remains of aneient imebh)s, snajiestinu', perhaps, that some of the oeeu- pants wcvc absent at tlie time of the destruetion of the viHas'e. AVhen large door-like oi)eninijs in upi)er external walls were built up and plastered ov(>r in this way. as in some ruins, the imrjxise was to econo- mize heat durinji' the wint(>r, as bhiiikets or rugs made of skins would be inadequate. Besides tlie closing and reopeniuii' of doors and windows just de- scribed, the mo(h'in [lueblo liuihlcrs tVcipiently make permanent changes in such openings. I )oors are often converted into windows, and windows are reduced in size or enlarged, or new ones are broken through the walls, apparently, with the greatest freedom, so that they do not, from their finish or method of construction, furnish any clue to the antiquity of the mud-covered wall in which they are found. Occasionally surface weathering of the walls. i)articularly in Zufii, exposes a bit of horizontal pole embedded in the masonry, the lintel of a window long since sealed up and obliterated by successive coats of mud finish. It is probable that many openings are so covered up as to leave no trace of their ex- istence on the external wall. In Zuni particularly, where the original arrangement for entering and lighting many of the rooms must have been wholly lost in the dense clustering of later times, such changes are very numerous. It often happens that the addition of a new room will shut oti" one or more old windows, and in such cases the latter are often converted into interior niches which serve as open cupboards. Such niches were sometimes of considerable size in the older pueblos. Changes iu the character of openings are quite common in all of the pueblos. Usually the evidences of such changes are much clearer in the rougher and more exposed work of Tusayan than in the adobe-finished houses of Zuni. PI. cvii illustrates a large balcony like opening iu Oraibi that has been reduced to the size of an ordinary door by filling in with rough masonry. A small window has been left immediately over the lintel of the newer door. PI. cviii illustrates two large openings in this village that have been treated in a somewhat similar manner, but the filling has been carried farther. Both of these openings have been used as doorways at one stage of their reduction, the one on the right hav- ing been i)rovi(led with a small transom; the combined oiieiiing was arranged wholly within the large one and under its transom. In the fiuther conversion of this doorway into a small window, the secondary transom was blocked u]> with stone slabs, set on edge, and a small loop- hole window in the upinnlcfthand corner of the large opening was al.so closed. The masonry filling of the large opening on the left in this illustration shows no trace of a transom over the smaller doorway. A small loophole in the corner of this large opening is still left open. It will be noted that the original transoms of the large openings have in all these cases been entirely filled up with masonry. BURFAU OF ETHNOLOG'^ EIGHTH ANNUAL REPORT PI. XCVIII CROSS PIECES ON ZUNI LADDERS. UINDELEFF.] ROOF OPENINGS. 201 The clearness with which all the steps of the gradual reduction of these openings can be traced in the exposed stone work is in marked contrast witli the oljscurity of such features in Zuni. In tlie latter group, however, examples are occasionally seen where a doorway has been partly closed with masonry, leaving enough space at the top for a window. Often in such cases the fllled-in masonry is thinner than that of the adjoining wall, and cdnsecjuently the form of the original doorway is easily traced. Fig. 93, from an adobe wall in Zuiii, gives an illustration of this. The entrance doorway of the detached Zuiii house illustrated in PI. LXXXiii, has been similarly reduced in size, leaving traces of the orignal form in a slight oii'set. In modern times, both in Tusayan and Cibola, changes in the form and disposition of openings seem to have been made with the greatest freedom, but in the ancient pueblos altered doors or windows have rarely been found. The original placing of these Fig. 93. A Zuiii doorway converted into a window. features was more carefully considered, and the buildings were rarely subjected to unforeseen and irregular crowding. In both ancient and modern pueblo work, windows, used only as such, seem to have been universally quadrilateral, ofl'sets and steps being con- fined exclusively to doorways. ROOF OPENINGS. The line of separation between roof openings and doors and windows is, with few exceptions, sharply drawn. The origin of these roof-holes, whose use at the present time is widespread, was undoubtedly in the simple trap door which gave access to the I'ooms of the first terrace. PI. xxxviii, illustrating a court of Oraibi, shows in the foreground a kiva hatchway of the usual form seen iu Tusayan. Here there is but little difference between the entrance traps of the ceremonial chambers and those that give access to the rooms of the first terrace ; the former are in most cases somewhat larger to admit of ingress of costumed dan- 202 PUEBLO ARCHITECTI'RE. cers, and tlie kiva traps are usually on a somewhat sharper slope, eon- forming to the pitch of the small dome-roof of the kiva.s, while those of the house terraces have the scarcely perceptible fall of the house roofs in which they are placed. In Zufii, however, where the development and use of openings has been carried further, the kiva hatchways are distinguished by a specialized form that will be described later. An examination of the plans of the modern villages in Chapter ii and iii will show the general distribution of roofopenings. Those used as hatch- ways are distinguishable by their greater dimensions, and in many cases by the i^resence of the ladders that give access to the rooms below. The smaller roof oi^enings in their simplest form are constructed in essen- tially the same manner as the traj) doors, and the width is usually regu- lated by the distance between two adjacent roof beams. The second FiQ. 94. Zuui roof-openings. series of small roof poles is interru^jted at the sides of the opening, which sides are finished by means of carefully laid small stones in the same manner as are projecting copings. This finish is often carried several inches above the roof and crowned with narrow stone slabs, one on each of the four sides, forming a sort of frame which protects the mud plas- tered sides of the opening from the action of the rains. Examples of this simple type may be seen in many of the figures illustrating Chap- ters II and III, and in PI. xcvii. Fig. 9-i also illustrates common types of roof openings seen in Zuui. Two of the examjiles in this figure are o> o fe*^ I If ^\V MIKDELEFF.] ROOF OPENINGS. 203 of openings that give access to lower rooms. Occasional instances are seen in this pneblo in which an exaggerated height is given to the cop- ing, the result slightly approaching a square chimney in effect. Fig. 95 illustrates an example of this form. 1 Flo. 95. A Zufii ruof oiu-iiiug, with raised coping. In Zufii, where many minor variations in the forms of roof openings occiir, certain of these variations appear to be related to roof drainage. These have three sides crowned in the usual manner with coping stones I^^^ZIL. PlQ. 96. Zuiii roof-openings, with one elevated end. laid flat, but the fourth side is formed by setting a thin slab on edge, as illustrated in Fig. 90. Fig. 94^ also embodies two specimens of this form. 204 PUEBLO ARCHITECTURE. The special object of tlii.s arraiijieinent is in some cases difficult to determine; tlie raised end in all the examples on any one roof always takes the same direction, and in many cases its position relative to drainage suggests that it is a provision against flooding l)y rain on the slightly slojiing roof; but this relation to drainage is by no means con- stant. Eoof holes on the west side of the village in such positions as to be directly exposed to the violent sand storms that prevail here during certain months of the year seem in some cases to have in view protec- tion against the flying sand. We do not meet with evidence of any fixed system to guide the disposition of this feature. In many cases these trap holes are provided with a thin slab of sandstone large enough to cover the whole opening, and used in times of rain. During fair weather these are laid on the roof, near the hole they are designed to cover, or lie tilted against the higher edge of the trap, as shown in Fig. 97. Pig. 97. A Zimi roof hole with cover. When the cover is jilaced on one of these holes, with a high slab at one end, it has a steex) pitch, to shed water, and at the same time light and air are to some extent admitted, but it is very doubtful if this is the result of direct intention on the part of the builder. The possi- ble development of this roof tra]) of unusual elevation into a ruilinien- tary chimney has already been mentit)ne(l in the discussion of chimneys. A development in this direction would possibly be suggested by the desirability of separating the access by ladder from the inconvenient smoke hole. This must have been brought very forcibly to the atten- tion of the Indian when, at the time a fire was burning in the fireplace, they were compelled to descend the ladder amidst the smoke and heat. y- •\ I A is, 1 / f ^ ^ MINDELEFF.J ROOF OPENINGS. 205 The survival to the present time of siidi an inconvenient arrangement in the kivas can be explained only on the groniid of the intense con- servatism of these people in all that pertains to religion. In the small roof holes methods of construction are seen which would not be so i)rac- ticable on the larger scale of the ladder holes after which they have been modeled. In these latter the sides are built up of masonry or adobe, but the framing around them is more like the usual coping over walls. The stone that, set on edge in the snuill ()i)euings built for the admission of light, forms a raised end never occurs in these. The ladder for access rests against the coping. Wlien occurring in connection with kivas, ladder holes have certain I^eculiarities in which they differ from the ordinary form used in dwell- ings. The opening in such cases is made of large size to admit dancers in costume with full paraphernaha. These, the largest roof openings to be found in Zuiii, are framed with pieces of wood. The methods of Fia. 98. Kiva trapdoor iu Zufli. holding the pieces in place vary somewhat in minor detail. It is quite likely that recent examples, while still x)reserving the form and general appearance of the earher ones, would bear evidence that the builders had used their knowledgeof improved methods of joining and finishing. As may readily be seen from the illustration, Fig. 9S, this framing, by tlie addition of a cross piece, divides the opening unequally. The smaller aperture is situated immediately above the fireplace (which conforms to the ancient type without chimney and located in the open floor of the room) and is very evidently designed to furnish an outlet to the smoke. In a chamber having no side doors or windows, or at most ' 206 PUEBLO ARCHITECTURE. very small square windows, and consequently no drafts, the column of smoke and flame can often on still nij^iits be seen rising vertically from the roof. The other portion of the opening containing the ladder is used for ingress and egress. This singular combination strongly suggests that at no very remote period one opening was used toanswerboth purposes, as it still does in the Tusayan kivas. It also suggests the direction in which differentiation of functions began to take place, which in the kiva was delayed and held back by the conservative religious feeling, when in the civil architecture it may have been the initial point of a develop- ment that culminated in the chimney, a development that was assisted in its later steps by suggestions from foreign sources. In the more primitively constructed examples the cross pieces seem to be simply laid on without any cutting ui. The central piece is held in place by a peg set into each side piece, the weight and thrust of the ladder helping to hold it. The primitive arrangement here seen has been sonu'what im- proved upon in some other cases, but it was not ascertained whether these were of later date or not. In the best made frames for kiva entrances the timbers are " halved" in the manner of our carpenters, the joint being additionally secured by a pin as shown in Fig. 99. The use of a frame of wood in these trap- doors dates back to a comparatively high antiquity, and is not at all a modern innova- tion, as one would at first be incliiu'd to be- lieve. Theii' use in so highly developed a form in the ceremonial chamber is an argu- ment in favor of antiquity. Only two exam- ples were discovered by Mr. L. H. Morgan in a ruined i)uoblo on the Animas. "One of these measured IG by 17 inches and the other was 16 inches square. Each was formed in the floor by pieces of wood put together. The work was neatly done." Unfortunately, Mr. Morgan does not de- scribe in detail the manner in which the join- ing was effected, or whether th«^ pieces were halved or cut to fit. It seems hardly likely, considering the rude facihties possessed by the ancients, that the enormous labor of re- no. 99. Halved and pinned trapdoor duciug large picCCS of WOod to SUCll iutcrflt- frame of a Zuni kiva. j-jug shapes would liavc been Undertaken. A certain neatness of finish would undoubtedly be attained by arranging the principal roof beams and the small poles that cross them at right angles, in the usual careful manner of the ancient builders. The kiva roof opening, with the hole serving for access and smoke exit, is paral- ' Contributions to N. A. Ethnology, vol. i. Honse Life, etc., p. 182. MINDELEFF.J ROOF OPENINGS. 207 leled in the excavated lodges of the Sau Francisco Mountains, where a single opening served this doulile jmrpose. A slight recess or excava- tion in the side of the entrance sluift evidently served for the exit of smoke. At the village of Acoiiia the kiva trapdoors differ somewhat from the Zufii form. The survey of this village was somewhat hasty, and no opportunity was afforded of ascertaining from the Indians the .si^ecial purpose of the mode ot construction adopted. The roof hole is divided, as in Zuui, but the portion against which tlie ladder leans, instead of being made into a smoke vent, is provided with a small roof. These roof holes to the ceremonial chamber are entered directly from the open air, while in the dweUing rooms it seems customary (much more cus- tomary than at Zufii) to enter the lower stories through trapdoors within upper rooms. In many instances second-story rooms have no exterior rooms but are entered from rooms above, contrary to the usual arrangement in both Tusayan and Cibola. All six of the kivas in this village are provided with this peculiarly constructed opening. In Zuni close crowding of the cells has led to an exceptionally fre- quent use of roof-lights and trapdoors. The ingenuity of the builders was greatly taxed to admit sufflcieut light to the inner rooms. The roof hole, which was originally used only to famish the means of access and light for the first terrace, as is still the case in Tusayan, is here used in all stories indiscriminately, and principally for light and air. In large clusters there are necessarily many dark rooms, which has led to the employment of great numbers of roof holes, more or less directly modeled after the ordinary trapdoor. Their occurrence is particularly frequent in the larger clusters of the village, as in house No. 1. The exceptional size of this pile, and of the adjoining house No. 4, with the consequent large proportion of dark rooms, have taxed the ingenuity of the Zuni to the utmost, and as a result we see roof openings here assuming a degree of importance not found elsewhere. In addition to roof openings of the type described, the dense clus- tering of the Zuni houses has led to the invention of a curious device for lighting inner rooms not reached by ordinary external openings. This consists of an opening, usually of oval or subrectangular form in elevation, placed at the junction of the roof with a vertical wall. This opening is carried down oblicpiely between the roofing beams, as shown in the sections. Fig. 100, so that the light is admitted within the room just at the junction of tlie ceiling and the inner face of the wall. With the meager facilities and rude methods of the Zuni, this peculiar arrangement often involved weak construction, and the open- ings, placed so low in the wall, were in danger of admitting water ft-om the roof. The tlifftculty of obtaining the desired light by this device was much lessened where the outer roof was somewhat lower than the ceil- ing within. 208 PUEBLO ARCHITECTURE. These oblique opening's oecur not only in the larjicer clusters of houses Nos. 1 and 4, but also in the more ojtenly i)lanneil portions of the vil- lage, though they do not occur (iither at Acoma or in the Tusayan vil- lages. They afford an interesting example of the transfer and continu- ance in use of a constructional device developed in one place by unusual conditions to a new field iu which it was uncalled for, being less efficieut and more difiBcult of introduction than the devices in ordinary use. . v--^^,^ ^^^^^^>^^^^A>-> FiQ. 100. Tj-pical sectiona of Zufii oblique openings. FURNITURE. The pueblo Indian lias little household furniture, in the sense in which the term is commonly employed ; but his home contains certain featiires which are more or less closely embodied in the house construction and which answers the purpose. The suspended pole that serves as a clothes rack for ordinary wearing apparel, extra blankets, robes, etc., has already been described in treating of interiors. Religious costumes and cere- monial para])hernalia are more carefully jtrovided for, and are stored away in some hidden corner of the dark storerooms. The small wall niches, which are formed by closing a window with a thin filUng-iu wall, and which answer the purpose of cupboards or recep- :>?1^^. r'^'i^i^i ;i* '^if '^'a f% M^ y> ?M * - (* tf-x 'til I ^.V^1 f! ' I 4 ^^^t't-'-yt MINDELEFF.) MEALING STONES AND GRAIN BINS. 209 tacles for many of the smaller household articles, have also been described and ilhistrated in connection witli the Znni interior (PI. Lxxxvi). In many houses, both in Tusayan and in (Jibola, shelves are constructed for the more convenient storage of food, etc. These are often constructed in a very primitive manner, particularly in the former province. An un- usually frail example may be seen in Fij^. (il, in connection with a fire- place. Fig. 101, showing a series of meaUng stones in a Tusayan house, also illustrates a rude shelf in the corner of the room, supported at one Fio. 101. Arrangement of mealing atones in a Tusayan house. end by an upright stone slab and at the other by a projecting wooden peg. Shelves made of sawed boards are occasionally seen, but as a rule such boards are considered too valuable to be used in this manner. A more common arrangement, particularly in Tusayan, is a combination of three or four slender poles placed side by side, 2 or 3 inches apart, forming a rude shelf, upon which trays of food are kept. Another device for the storage of food, occasionally seen in the pueblo house, is a pocket or bin built into the corner of a room. Fig. 101, illus- trating the plan of a Tusayan house, indicates the })osition of one of these cupboard-like inclosiu'es. A sketch of this specimen is shown in 8 BTH U 210 PUEBLO AECHITECTUliE. Fig. 102. This bin, used for the storage of beans, grain, and the like, is formed by cutting oil' a corner of the room l)y setting two stone slabs into the floor, and it is covered with the mud plastering which extends over the neighboring walls. A curious modilication ofthis device was seen in one of the inner rooms in Zuni, in the house of Jose I'ie. A large earthen jar, apparently an ordi- nary water vessel, was built into a |)roJecting masonry bench near the «/ corner of the room in such a manner that its rim projected less than half an inch above its surface. Tlus jar was used for the same i)urpose as the Tu- Fig. 102. A Tnsayan grain 'Jin. sayan corner bin. Some of the Indians of the present time have chests or boxes in which their ceremonial blankets and paraphernalia are kept. These of course have been introduced since the days of American boards and boxes. In Fm. 103. A Zimi plume box. Zuni, however, tlie Indians stdl use a small wooden rece])tacle for tlie jire- cious ceremonial articles, such as feathers and beads. Tiiis is an oblong box, provided with a countersunk lid, and usually carved from a single single piece of wood. Typical specimens are Qlustrated in Figs. 103 and . 104. A Znfli plwme box. 101:. The workmanshij) displayed in these objects is not beyond the aboriginal skill of the native workman, and their use is undoubtedly ancient. MINDEI.EFF.] MEALING TROUGHS. 211 Perhaps the most important article of furniture in the home of the jnieblo Indian is the mealing tnmgh, containing the household milling apparatus. This trough usually contains a series of three nictates of varying degrees of coarseness firmly fixed in a slanting position most convenient for the workers. It consists of thin slabs of sandstone set into the floor on edge, similar slabs forming the separating i>artitious between the compartments. This arrangement is shown in Fig. 105, illustrating a Tusayau mealing trough. Those of Zuni are of tlie same form, as may be seen in the illustration of a ZuDi interior, Fig. 105, Flo. 105. A Tusayan mealing trough. Occasionally in recently constructed specimens the thin inclosing walls of the trough are made of planks. In the example illustrated one end of the series is bounded by a board, all the other walls and divisions being nuide of the usual stone slabs. The metates themselves are not usually more than 3 inches in thickness. They are so adjusted in their setting of stones and mortar as to slope away from the operator at the proper angle. This arrangement of the mealing stones is characteristic of the more densely clustered communal houses of late date. In the more primi- tive house the mealing stonewas usually a single large piece of cellular basalt, or similar rock, in which a broad, slop- ing dei)ression was carved, aiul which Fig. 106. An ancient pueblo form of metat*. couldbetransported from ])lace to place. Fig. 10(5 illustrates an example of this type from the vicinity of Globe, in southern Arizona. The stationary mealing trough of the present day is undoubtedly the successor of the earlier moveable form, yet it was in use among the pueblos at the time of the first Spanish expedition, as the following extract from Gastaneda's account' of Gibola will show. He says a special room is designed to grind the grain : " This last is apart, and contains a furnace and three stones made fast iu masonry. > Given by W. W. H. Davis in El Gringo, p. 119. 212 PUEBLO ARCHITECTURE. Three women sit down before these stones; the first crushes the grain, the second brays it, and the third reduces it entirely to powder." It ^^^ll Ije seen liow exactly this description fits both the arrangenuuit and the use of this mill at the present time. The jierfection of mechanical devices and the refinement of methods here exhibited would seem to be in advance of the a('hievement of this people in other directions. The grinding stones of the mealing apparatus ai'e of correspondingly varying degrees of roughness; those of basalt or lava are used for the first crushing of the corn, and sandstone is used for the final grinding ou the last metate of the series. By means of these pi'imitive appli- ances the corn meal is as finely ground as our wheaten flour. The grind- ing stones now used are always flat, as shown in Fig. 105, and differ from those that were used with the early massive type of metate in being of cylindrical form. One end of the series of milling troughs is usually biult against the wall near the corner of the room. In some cases, where the room is quite narrow, the series extends across ft'om wall to wall. Series com- prising four mealing stones, sometimes seen in Zuni, are very generally arranged in this manner. In all cases sufficient floor space is left be- hind the mills to accommodate the women who kneel at their work. PI. Lxxxvi illustrates an unusual arrangement, in which the fourth mealing stone is set at right angles to the other stones of the series. Mortars are in general use in Zuiii and Tusayan households. As a rule they are of considerable size, and made of the same material as the rougher mealing stones. They are employed for crushing and grind- ing the chile or red pepper that enters so largely into the food of the Zuiii, and whose use has extended to the Mexicans of the same region. These mortars liave the ordinary circular depressions and are used with a round pestle or crusher, often of somewhat long, cylindrical form for convenience in handling. Parts of the apparatus for indoor blanket weaving seen in some of the pueblo houses may be included under the heading of furniture. These consist of devices for the attachment of the movable parts of the loom, which need not be described in this connection. In some of the Tusayan houses may be seen examples of posts sunk in the floor pro- vided with holes for the insertion of cords for attaching and tightening the warp, similar to those built into tlie kiva floors, ilhistrated in Fig. 31. No device of this kind was seen in Zuni. A more primitive appli- ance for such work is seen in both groups of pueblos in an occasional stump of a beam or short pole projecting from the wall at varying heights. Ceiling beams are also used for stretching the warp both in blanket and belt weaving. Tlie f'urnisliing.s of a pueblo house do not inchide tables and chairs. T]ie meals are eaten dii-ectly from the stone-paved floor, the participants rarely having any other seat than the blaidvct that they wear, rolled up or folded into convenient torm. Small stools are sometimes seen, but MINDELEFF.] STOOLS AND CHAIRS. 213 the need of such appliances does not seem to be keenly felt by these Indians, who can, for hours, sit in a pecidiar squatting- position on their haunches, without any apparent discomfort. Thougli moveable chairs or stools are rare, nearly all of the dwellings are provided with the low ledge or bench around the rooms, which in earlier times seems to have been confined to the kivas. A slight advance on this fixed form of seat was the stone block used in the Tusayan kivas, described on \k 132, which at the same time served a useful purpose in the adjustment of the warp threads for blanket weaving. The few wooden stools observed show very primitive workmanship, and are usually made of a single ])iece of wood. Fig. 107 illustrates two forms of wooden stool from Zuni. The small three-legged stool on Fig. 107. Zuni stools. the left has been cut fiom the trunk of a pinon tree in such, a manner as to utilize as legs the three branches into which the main stem sepa- rated. The other stool illustrated is also cut from a single piece of tree trunk, which has been reduced in weight by cutting out one side, leav- ing the two ends for support. A curiously worked chair of modern form seen in Zuiii is illustrated in Fig. 108. It was difficult to determine the antiqaity of this sijecimen, as its rickety condition may have been due to the clumsy work- manship quite as much as to the eflects of age. Rude as is the workmanship, how- ever, it was far beyond the unaided skill of the native craftsman to join and mor- tise the various pieces that go to make up this chair. Some decorative efiect has been sought here, the ornamentation, made up of notches and sunken grooves, Fig. 108. A Zufii chair. closely resembUng that on the window sash iUustrated in Fig. 88, and somewhat similiar in effect to the carving on the Spanish beams seen in the Tusayan kivas. The whole construction strongly suggests Span- ish influence. 214 PUEBLO ARCHITECTURE. Even the influence of Americans lias as yet failed to bring about the use of tables or bedsteads among the ]meblo Indians. The floor answers all the purposes of both these useful articles of furniture. The food dishes are placed directly upon it at meal times, and at night the blankets, rugs, and sheep skins that form the bed are spread directly upon it. These latter, during the day, are suspended upon the clothes pole previously described and illustrated. CORRALS AND GARDENS. The introduction of domestic sheep among the pueblos has added a new and important element to their mode of living, but they seem never to have reached a clear understanding as to how these aniruals should be cared for. No forethought is exercised to separate the rams so that the lambs will be born at a favorable season. The flocks consist of sheep and goats which are allowed to run together at all times. Black sheep and some with a grayish color of wool are often seen among them. No attempt is made to eliminate these dark-fleeced members of the flock, since the black and gray wool is utilized in its natural color in produc- ing many of the designs and patterns of the blankets woven by these people. The flocks are usually driven ui) into the corrals or inclosures every evening, and are taken out again in the morning, fi-equently at quite a late hour. This, together with the time consumed in driving them to and from pasture, gives them much less chance to thrive than those of the nomadic Navajo. In Tusayan the corrals are usually of small size and inclosed by thin walls of rude stone work. This may be seen in the foreground of PI. xxi. PI. cix illustrates several corrals just outside the village of Mashongnavi similarly constructed, but of somewhat larger size. Some of the corrals of Oraibi are of still larger size, approaching in this respect the corrals of Cibola. The Oraibi pens are rudely rectangular in form, with more or less rounded angles, and are also built of rude masonry. In the less important villages of Cibola stone is occasionally used for inclosing the corrals, as in Tusayan, as may be seen in PI. lxx, illns- trating an iuclosure of this character in the court of the farming pueblo of Pescado. PI. ex illustrates in detail the manner in M-hich stone work is combined with the use of rude stakes in the construction of this iuclosure. On the rugged sites of the Tusayan villages corrals are placed wherever favorable nooks happen to be found in the rocks, but at Zuni, built in tlie comparatively open plain, they form a nearly crush held in place by horizontal poles tied on with strips of rawhide. The rudely contrived gateways are supported in natural forks at the top and sides of posts. Often one or two small inclosures used for burros or horses occur near these sheep corrals. The construction is identical with those above described and is very riulc. It is illustrated in Fig. 109, which shows the manner in which the stakes are arranged, and also M1^^)ELEFF.] CORRALS. 215 the nietliod of attachinfc tlie horizontal tie-pieces. The construction of these inclosures i.*; frail, and the danger of pushing the stakes over by pressiu'e from witliin is guarded against by employing forked braces that abut agamst horizontal pieces tied on 4 or .5 feet from the ground. Reference to PI. lxxiv will illustrate this construction. r'--t- -^ m FiG. 109. Coustructiou of a Zuiii corral. Within the village of Zuni inclosures resembling miniature corrals are sometimes seen built against the houses; these are used as cages for eagles. A number of these birds are kept in Zuhi for the sake of their plumage, which is highly valued for ceremomal i)urposes. PI. cxr illustrates one of these coops, constructed partly with a thin a^lobe wall and partly with stakes arranged lilce tliose of the corrals. In both of the pueblo gi'oups under discussion small gardens contigu- ous to the villages are frequent. Those of Tusayan are walled in with stone. Within the pueblo of ZiiQi a small group of garden patches is inclosed by stake fences, but the majority of the gardens in the vicinity of the 21fi PUEBLO ARCIIITECTURK. principal villages aiv jn'ovidcd with low walls of iriud masonry. The small terraced gardens here are near the river bank on the southwest and southeast sides of the village. The inclosed spaces, averaging in size about 10 feet scjuare, are used for the cultivation of red peppers, beans, etc., which, during the dry season, are watered by hand. These inclosiires, situated close to the dwellings, suggest a probable explana- tion for similar inclosurcs found in many of the ruins in the southern and eastern portions of tlie ancient jiueblo legion. Mr. Bandelier was informeil by the Piinas ' tliat these inclosures were ancient gardens. He »; -5^. "^o-- Vui. no. (iardens of Zufii. concluded that since acequias were frequent in the iuunediate vicinity these gardens must have been used as reserves in case of war, when the larger fields were not available, but the manner of their occurrence in Zuni suggests rather that they were intended for cultivation of special crops, such as pepper, beans, cotton, and perhaps also of a variety of ' Fifth Ann. Kept. Arch. Inst. Am., p. 92. >^Mv^ MINDELEFF. 1 KISI CONSTRUCTION. 217 tobacco—corn, melons, squashes, etc., being cultivated elsewhere in larger tracts. There is a large groui) of gardens on the bank of the stream at the southeastern corner of Zuiii, and here there are slight in- dications of terracing. A second group on the steeper slope at the southwestern corner is distinctly terraced. Small walled gardens of the same type as these Zuiii examples occiu- in the vicinity of some of the Tusayan villages on the middle mesa. They are located near the springs or water pockets, apparently to facilitate watering by hand. Some of them contain a few small peach trees in addition to the vegetable crops ordinarily met with. The clusters here are, as a rule, smaller than those of Zuiii, as there is much less si)ace available in the vicinity of the springs. At one point on the west side of the first mesa, a few miles above Walpi, a coi^ious si)ring serves to irrigate quite an extensive series of small garden patches distributed over lower slojies. At several points around Zuni, usually at a greater distance than the terrace gardens, are fields of umch larger area inclosed in a similar manner. Their iuclosure was simply to secure them against the depre- dations of stray burros, so numerous about the village. When the crops are gathered in the autumn, several breaches are made in the low wall and the burros are allowed to luxuriate on the remains. PI. lix indicates the position of the large cluster of garden jjatches on the southeastern side of Zuni. Fig. 110, taken from photographs made in 1873, shows several of these small gardens with their growing crops and a large field of corn beyond. The workmanshij) of the garden walls as contrasted with that of the house masonry has been already de- scribed and is illustated in PI. xc. "KISl" CONSTEUCTION. Lightly constructed shelters for the use of those in charge of fields were probably a constant accompaniment of pueblo horticulture. Such shelters were built of stone or of brush, according to which material was most available. In very i^recipitous localities, as the Canyon de Chelly, these outlooks naturally became the so-called cliff-dwellings or isolated shelters. In Cibola single stone houses are in common use, not to the exclusion, how- ever, of the lighter structures of brush, while in Tusayan these lighter forms, of which there are a number of well defined varieties, are almost exclusively used. A detailed study of the methods of construction em- ployed in these rude shelters would be of great interest as affording a comparison both with the building methods of the ruder neighboring- tribes and with those adopted in constructing some of the details of the terraced house; the writer, however, did not have an opportunity of making an examination of all the field shelters used in these pueblos. Two of the simpler types are the " tuwahlki," or watch house, and the "kishoui," or uncovered si ,de. The former is constructed by first 218 PUEBLO ARCHITECTURE. planting a short forked stick in tlie ground, wliicli supports one end of a i)ole, the other end resting on the ground. The interval between this ridge pole and the ground is roughly lilled in with slanting sticks and brush, the inclosed space being not more than 3 feet in height, with a maximum width of four or live feet. These shelters are for the accom- modation of the children who watch the melou patches until the ftuit is harvested. The kishoni, or uncovered shade, illustrated in Fig. Ill, is perhaps ?f ~^ -^ --\ ^^.^ " " '^ "^^'~~"-11. -J'-^,^^: ^^>. Fig. 111. Kishoni, or unco%^pred sliado, of Tuaayan. the simplest form of shelter emjjloyed. Ten or a dozen Cottonwood saplings are set lirmly into the ground, so as to form a slightly curved iuclosure with convex side toward the south. Cottonwood and wiUow boughs in foliage, grease-wood, sage brush, and rabbit brush are laid with stems ui)wai(l in eveu rows against these .saplings to a height of G or 7 feet. This light material is held in place by bauds of small Cot- tonwood branches laid in continuous horizontal liiu's around the out- side of tlu>. shelter and these are attached to the upright saplings with cottouwood aud willow twigs. OH o o o :o > MINDELEFF.] FORMS OF FIELD SHELTER. 219 Figs. 112 and 1 !•'? illustrate a much more elaborate field shelter in Tusayau. A s may readily be seen from the figures this shelter covers a 'A'^'!^- JTiG. 112. A Tusayau Held shelter, from southwest. considerable area; it will be seen too that the upright branches that inclose two of its sides are of sufficient height to considerably shade the level roof of poles and brush, converting it into a comfortable retreat. is \ i| I 1^1 * ^y;j»«".i^^--"''<-f;« Fig. 113. A Tusayau lielU abeltur, trom northeast. : 220 PUEBLO AKCHITECTURE. ARCHITECTURAL NOMENCLATURE. The fdllowiiiK noineuclatuie, collected by Mr. Stephen, comprises the terms commonly used in designating the constructional details of Tusayan houses and kivas Kiko'li The ij;roiiiul floor rooms foniiins; thp first terrace. Tupii'bi The roofed recess at the eud of tlie first terrace. Ah'pabi i „, ,. : A terrace roof. Ih pol)i \ Tvipat'caih'polji The third terrace, used in common as a loitering place. Tumtco'kobi "The place of the flat stone;" small rooms in which "piki,'" or paper-bread, is baked. "Tuma,"the piki stone, and " tcok" describing its flat position. Tupa'tca "Where you sit overhead;" the third story. O'ml Ah'pabi The secoud story; a doorway always opens from it upon theroof of the "kiko'li." Kitcobi "The highest place ; " the fourth story. Tuhkwa A wall. Puce An outer corner. Apaphucua An inside corner. Lestabi The main roof timbers. Wina'kwapi Smaller cross poles. "Winahoya," a small pole, and "Kwapi," in place. Kaha'b kwapi The willow covering. Siiibi kwapi The brush covering. Si'hii kwapi The grass covering. Kiam' balawi Themudplasterof roof covering. "Balatle'lewini," to spread. Tcukat'ovewata Dry earth covering the roof. "Tcuka," earth, "katuto," to ait, and "at'cvewata," one laid above another. Kiami An entire roof. Kwo'pku The fireplace. Kwi'tcki "Smoke-house," an inside chimney-hood. Sibvu'tiituk'mula A series of bottomless jars piled above each other, and luted together as a chimney-top. Sibvu' A bottomless earthen vessel serving as a chimney pot. Bok'ci Any small hole in a wall, or roof, smaller than a doorway. Hi'tci An opening, such as a doorway. This term is also applied to a gap in a cliff. Hi'tci Kalau'wata A door frame. Tunau'iata A lintel; literally, "that holds the sides in place." WuwAk'pi "The place step;" the door sill. Ninuh'pi A hand hold; the small pole in a doorway below the lintel. Pana'ptca iitc'pi bok'ci A window; literally, "glass covered opening." Ut'cpi A cover. ,W,i,na,.., , . > A door. "Apiab," inside; wiua, a pole.ute pi i > > i( O'wa utc'ppi "Stone cover," a stone slab. o o MiNDELEFP.] ARCHITECTURAL NOMENCLATURE. 221 Tui'ka A projection in the wall of a room suggesting a par- tition, siicli as shown in PI. Lxxxv. The same term is applied to a projecting cliff in a mesa. Kiam'i An entire roof. The main beams, cross poles, aud roof layers have the same names as in the kiva, given later. Wina'kii'i Projecting poles; rafters extending beyond the walls. Bal'kakini " Spread out ; " the floor. O'tcokpii'h "Leveled with stones;" a raised level for the foundation. Ba'lkakini tii'wi "Floor ledge;" the floor of one room raised above that of an adjoining one. Hako'la " Lower place ; " the floor of a lower room. Sand dunes in a valley are called "Hakolpi." Ko'ltci A shelf. Owako'ltci A stone shelf. Ta'pii kii'ita A support for a shelf. Wina' koltci A hewn plank shelf. Kokiiini A wooden peg in a wall. Tiileta A shelf hanging from the ceiling. Tiilet'haipi The cords for suspending a shelf. TUkOlci A niche in the wall. Tiikftli A stone mortar. Ma'ta The complete mealing apparatus for grinding corn. Owa'mata The trough or o of remnants of various bands of builders. In Cibola, however, some of the Indians state that their ancestors, before reaching Zuni, built a number of pueblos, whose I'liins are distinguished from those illustrated in the l)reseiit paper by the presence of circular kivas, this form of ceremonial room l)eing, apparently, wholly absent from the Cibolan pueblos here discussed. The people of Cibola and of Tusayan belong to distinct linguistic stocks, but theii" arts are very closely related, the differences being no greater than would result from the slightly different conditions that have operated within the last few generations. Zuni, perhaps, came more directly under early Spanish influence than Tusayan. Churches were established, as has been seen, in both jirovinces, but it is doubtful whether their presence produced any lasting impression on the people. In Tusayan the sway of the Spaniards was very brief. At some of the pueblos the chui'clies seem to have been built outside of the village proper where ami)le space was available within the pueblo; but such an encroachment on the original inclosed courts seems never to have been attempted. Zuni is an ai)parent exception ; l)ut all the house clusters east of the church have ijrobably been built later than the church itself, the church court of the present village lieing a much hirger area than would be reserved for the usual ]iueljlo court. These early churches were, as a rule, built of adobe, even when occurring in stone pueblos. The only exception noticed is at Ketchipauan, where it was b)iilt of the characteristic Indian smoothly chinked masonry. The Spaniards usually intruded their own construction, even to the compo- sition of the bricks, which are nearly always made of straw adobe. At Tusayan there is no evidence that a church or mission house ever formed part of the villages on the mesa summits. Their plans are com- plete in themselves, and probably i-epresent closely the first pueblos built on these sites. These summits have been extensively occujjied only in comj)aratively recent times, although one or more small clusters may have been built here at an early date as outlooks over the fields in the valleys below. It is to be noted that some of the ruins connected traditionally and historically with Tusayan and Cibola differ Iti no particular from stone pueblos widely scattered over the soutliwestern plateaus which have been from time to time invested with a halo of romantic antiquity, and 0^ f>:£ll^ Ml i. ^a: MISDELEFK-1 CONCLUSION. 225 regarded as iciiimkable acliievemeuts iu civilization by a vanished bnt once powerful race. These deserted stone houses, occurrini»- in the midstof desert .solitudes, appealed strongly to tlie imaginations of early explorers, and their stimulated faui^y connected the remains with "Az- tecs" and other my.sterious peoples. That this early implanted bias has caused the invention of many ingeniims theoiies concerning the origin and disappearance of the builders of the ancient pueblos, is amply attested m the {onelusions reached l)y many of the writers on this sub- ject. In connection with the architectural examination of some of these remains many traditions have been obtained from the present tribes, clearly indicating that some of the village rums, and even cliff dwell- ings, have been built and occupied by ancestors of the ])re.sent Pueblo Indians, sometimes at a date well within the historic period. The migrations of the Tusayan clans, as described in the legends collected by Mr. Steiihen, were slow and tedious. While they pursued their wanderings and awaited the favorable omens of the gods they halted many times and iilauted. They speak traditionally of stopping at certain places on their routes during a certain iiuml)er of " plant- ings," always Imilding the characteristic stone i)ueblos and then again taking up the march. When these Indians are qu(\stioned as to whence they came, their re- jjlies are various and contlictiiig; l)ut tliis is due to the fact that the members of one clan came, after a long series of wanderings, from the north, for instance, wliile those of other gcntes may have come last from the east. The tribe to-day seems to be made u]) of a collection or a. coii- federacyof many enfeebled remnants ofiiidependent phratries and groups once more nuiiieroiis and i)owerfiit. .Some clans traditionally referred to as having been important are now represented by few snr\ivors, and bid fair soon to become extinct. So the members of each phiatry have their own store of traditions, relating to the wanderings of their own an- cestors, which differ from those of other clans, and refer to villages suc- cessively built and occu])ied by them. In the case of others of tlie pueblos, the occupation of cliff dwellings and cav(^ lodges is known to have oc- curred within historic times. Both architectural iuid traditional eviileiicc ar(^ in accord in establish ing a continuity of descent from the ancient I'ueblos to those of the jucs- eiit day. Many of the coinmunities are now made n]> of the more or less scattered but interrelated remnants of gentes which in former times oc- cupied villages, the remains of which are to-day looked upon as the early homes of "Aztec colonies," etc. The adaptation of this architecture to the peculiar environment indi- cates that it has long been practiced under the same conditions that now prevail. Nearly all of the ancient i)ueblos were built of the sandstone found in 'natiiral quarries at the ba.ses of hundreds of cliffs throughout these table-lands. This stone readily breaks into small i)ieces of regular S KTH 1."> 226 PUEBLO ARCHITECTURE. form, suitable for use in the simple mas(jnry of the pueblos without re- ceiving any artificial treatment. The walls themselves give an exag- gerated idea of finish, owing to the care and neatness with which the component stones are placed. Some of the illustrations in the last chap- ter, from i^hotographs, show clearly that the material of the walls was much ruder than the appearance of the finished masonry would suggest, and that this finish depended on the careful selection and arrangement of the fragments. This is even more noticeable in the Chaco ruins, in which the walls were wrought to a liigh degrees of surface finish. The core of the wall was laid up with the laiger and more irregular stones, and was afterwards brought to a smooth face by carefully filling in and chinking the joints with smaller stones and fragments, sometimes uot more than a quarter of an inch thick ; this method is still roughly fol- lowed by both Tusayau and Cibolan builders. Although many details of construction and arrangement display remarkable ada^jtatiou to the physical character of the country, yet the Influence of such environment would not alone sufQce to produce this architectural tyi^e. In order to develop the results found, another ele- ment was necessary. This element was the necessity for defense. The pueblo population was probably subjected to the more or less continu- ous influence of this defensive motive throughout the period of their occupation of this territory. A strong independent race of jteoiile, who had to fear no invasion by stronger foes, would necessarily have been influenced more by the physical environment and would have progressed further in the art of building, but the motive for building rectangular rooms—the initial point of departure in the development of pueblo architecture—would not have been brought into action. The crowding of many habitations upon a small clift' ledge or other restricted site, re- sulting in the rectangular form of rooms, was most likely due to the conditions imposed by this necessity for defense. The general outlines of the development of this architecture wherein the ancient builders were stimulated to the best use of the exceptional materials about them, both by the difficult conditions of their semi-desert environment and by constant necessity for protection against their neighbors, can be traced in its various stages of growth from the primi- tive conical lodge to its culmination in the large communal village of many-storied terraced buildings which we find to have been in use at the time of the Spanish discovery, and which still survives in Zuiii, per- haps its most striking modern exauii)le. Yet the various steps have re- sulted from a simple and direct use of the material immediately at hand, while methods gradually improved as frequent experiments taught the builders more fully to utilize local facilities. In all cases the material was derived from the nearest available source, and often variations in the quality of the finished work are due to variations iu the quality of the stone near by. The results accomplished attest the jiatient and per- sistent industry of the ancient builders, but the work does not dis])lay great skill iu construction or in i)reparatiou of material. MiNDELEPP.] CONCLUSION. 227 The same desert environmeut that liiinished such au abundance of material for the ancient builders, also, from its difficult and inhospita- ble character and the constant variations in the water supply, com- pelled the frequent employment of this material. This was an impor- tant factor in bringing about the attained degi'ce of advancement in the building art. At the present day constant local changes occur in the water sources of these arid table-lands, while the general character of the climate remains unaltered. The distinguishing characteristics of Pueblo architecture may be re- garded as the product of a defensive motive and of an arid environment that tiirnished an abundance of suitable building material, and at the same time the climatic conditions that compelled its fiequent employ- ment. The decline of the defensive motive within the last few years has greatly aflected the more recent architecture. Even after the long practice of the system has rendered it somewhat fixed, comparative security from attack has caused many of the Pueblo Indians to i-ecog- nize the inconvenience of dwellings grouped in large clusters on sites difScult of access, while the sources of their subsistence are neces- sarily sparsely scattered over large areas. This is noticeable in the building of small, detached houses at a distance from the main villages, the greater convenience to crops, flocks and water outweighing the de- fensive motive. In Cibola particularly, a marked tendency in this direction lias sliown itself within a score of years; OJo Caliente, the newest of the farming ])ueblos, is perhaps the most striking example within the two provinces. The greater security of the pueblos as the country comes more fully into the hands of Americans, has also resulted in the more careless construction in modern examples as comiiared with the ancient. There is no doubt that, as time shall go on, the system of building- many-storied clusters of rectangular rooms wiU gradually be abandoned by these people. In the absence of the defensive motive a more con- venient system, einijloying scattered small houses, located near springs and fields, will gradually take its ]dace, thus returning to a mode of building that probably prevailed in the evolution of the pueblo prior to the clustering of many rooms into large defensive villages. PI. lxxxiii illustrates a building of the type described located on the outskirts of Zuni, across the river from the main pueblo. The cultural distinctions between the Pueblo Indians and neighbor- ing tribes gradually become less clearly defined as investigation pro- gresses. Mr. Cushing's study of the Zuni social, political, and religious systems has clearly established their essential identity in grade of cul- ture with those of other tribes. In many of the arts, too, such as weav- ing, ceramics, etc., these people in no degree surpass many tribes who build ruder dwellings. 228 PUEBLO ARCHITECTURE. Ill aichitecture, though, they have progicsscd far heyoiid their neigh- bors; many of the devices employed attest the essentially primitive character of the art, and demonstrate that the apparent distinction in grade of culture is mainly due to the exceptional condition of the en- vironment.