[1] In his Annual Report for 1908. See also Bulletin 41 of the Bureau of American Ethnology.

[2] A few holes that have been dug here and there in these mounds have brought to light sections of walls with good masonry, but no excavations that could be called extensive or scientific have yet been attempted on this site. The excavation of these mounds might reveal a pueblo like Walpi, and a comparison of objects from them with those from Cliff Palace would be important in tracing the relationship of cliff-dwellings and pueblos.

[3] Bulletin 41 of the Bureau of American Ethnology.

[4] All potable water for camp had to be brought from Spruce-tree House, about 2 miles away.

[5] It is generally stated by stockmen and others who claim to have seen Cliff Palace "years ago," that the walls of the buildings were much higher in the early eighties than they are at present.

[6] Appalachia, vi, 28-30, May, 1890, Boston, 1892.

[7] Jour. Amer. Geog. Soc., xxiii, no. 4, 598, New York, 1891.

[8] In The Cliff Dwellers of the Mesa Verde (a translation in English from the Swedish edition, Stockholm, 1893), (pp. 59-66), unfortunately not accessible to most readers on account of the limited edition and the cost. For this reason the description is here reproduced in extenso. (The references to illustrations and the footnotes in this excerpt follow NordenskiÖld.)

[9] The room marked 48 in the plan is visible in Pl. XIII. Almost in the center of the plate, but a little to the right, two small loopholes may be seen, and to their right a doorway, all of which belong to room 48; the walls of 49 and 50 are much lower than those of 48. Behind 48 the high walls of 43 may be distinguished.

[10] They are shown in the plate just to the left of the fold at its middle, rather low down.

[11] A part of this wall may be seen to the extreme right of Pl. XIII, and also in Fig. 34 behind and to the right of the tower.

[12] The illustrations referred to in this paragraph are in NordenskiÖld's work.

[13] As stated in a note (Peet, p. 133) Chapter VII is a reprint of Doctor Birdsall's article in the Journal of the American Geographical Society, op. cit.

[14] In Les CommunautÉs Anciennes dans le DÉsert AmÉricain. In this work may be found a ground plan of Cliff Palace by Morley and Kidder, the interior of kiva Q (pl. viii, e), and a large view of the ruin taken from the north (pl. i, b). (Plate and figure designations from Hewett.)

[15] Clearings in the forest indicate the positions of the former farms of the inhabitants of Cliff Palace.

[16] Access to Cliff Palace from the bottom of the canyon, although difficult, is possible, and a pathway might be constructed down its sides or along the top of the talus to several other cliff-dwellings. In the vicinity of Cliff Palace there are at least 20 ruins, large and small.

[17] One of these rooms had been chosen by eagles for their nests, but both nests and eggs were abandoned by the birds after the repair work was begun.

[18] In some of these waterways are found good examples of "potholes," some of considerable size, which often retain water for a long time. Their capacity was increased in prehistoric times by the construction of dams.

[19] While there has probably been considerable erosion in the bed of the canyon since Cliff Palace was constructed, this does not mean that "the action of the water carved out the valley, leaving at an inaccessible height buildings originally constructed on almost level land." See History N. Y. State Chapter, Colorado Cliff Dwellings Assoc., p. 11.

[20] Some, possibly considerable, of this mutilation may be ascribed to the former occupants. The Ute Indians will not now enter cliff-dwellings and probably are not responsible for their destruction.

[21] The author's hope is to excavate and repair in different sections of the Southwest a number of "type ruins," each of which will illustrate the major antiquities of the area in which it occurs. From an examination of these types the tourist and the student may obtain, at first hand, an accurate knowledge of the prehistoric architecture.

[22] In "Report, House of Representatives, No. 3703, 58th Congress," Mr. Coert Dubois ascribes to Cliff House (Cliff Palace) 146 rooms and 5 estufas (kivas). Unfortunately the error in the count of kivas has been given wide circulation. As stated in the present article, there are at least 23 rooms in Cliff Palace that may be called kivas.

[23] Fragments of mortar from the walls and floors, ground to powder, were used in the repair work.

[24] See Bulletin 50, Bureau of American Ethnology.

[25] In at least one of the Oraibi kivas the plastering of the wall is laid on sticks that form a kind of lathing. Whether this is a survival of an older method of construction or is traceable to European influence has not been determined, but it is believed to be a survival of prehistoric wall construction.

[26] The red color is derived from the red soil common everywhere on the mesa. Yellow was obtained from disintegrated rock, and white is a marl which is found at various places. The mortar used by the ancient masons became harder, almost cement, when made of marl mixed with adobe.

[27] This figure resembles closely that on the outside walls of the third story of room 11 of Spruce-tree House. (See pls. 4, 5, 6, Bulletin 41, Bureau of American Ethnology.)

[28] Isolated cliff-dwellings are scattered throughout the Southwest, but there are several areas, as the Mesa Verde, in which they are concentrated. Among these clusters may be mentioned the Canyon de Chelly, the Navaho National Monument, the Red Rocks area, and that of the upper Gila. One characteristic feature in which the cliff-dwellings of the Mesa Verde differ from some others is the independence of all of the upright walls from support of the sides of the cliffs. In the cliff-houses of the Navaho Monument a large majority of the houses have the rear wall of the cave as a wall of the building; a few of the houses in Cliff Palace have the same, but the largest number are entirely free from the cliff. This separation on all sides is due largely to the geological structure of the rear of the cavern in which the cliff-house stands.

[29] Workmen could operate in these parts only by tying sponges over their nostrils, so difficult was it to breathe on account of the fine dust.

[30] Possibly some of these may have been used sometimes for ceremonial purposes, or rather for the less important rites.

[31] Probably both men and women of one clan worked together in the construction of houses, the men being the masons, the women the plasterers. Each clan built its own rooms, and there were no differentiated groups of mechanics in the community.

[32] Some of the doorways were filled with rude masonry; evidently the rooms were thus closed in some instances before the buildings were deserted.

[33] The placing of the sill at a level with the floor is a modern innovation at Walpi. The oldest houses still have it elevated, as in Cliff Palace. In some of the cliff-houses of the Navaho Monument sills and floor levels are continuous.

[34] Smoke on the walls of certain second and third stories shows that fireplaces were not restricted to the ground floor.

[35] Among the Hopi the oldest woman, as a clan representative, owns the living rooms, but kivas are the property of the men, the kiva chief of certain fraternities being the direct descendant of the clan chief of the ceremony when limited to his clan.

[36] Genetically the room for storage of property was of earliest construction. This custom, which was necessary among agriculturists whose food supply was bulky, may have led to the choice of caves, natural or artificial, for habitation.

[37] See Property Rights in Eagles, American Anthropologist, vol. ii, pp. 690-707, 1907.

[38] While only one place where bodies were burned was found in Cliff Palace, several such places were found on top of the mesa. Evidences of similar inclosures occur at Spruce-tree House and at Step House.

[39] This type of building is believed to be the oldest in those sections of the Southwest where cliff habitations occur.

[40] The Mongkiva at Walpi has such a chamber which is closed by a door and is opened only when paraphernalia for certain ceremonies are desired. In the Warrior House at Walpi there is a similar chamber, ordinarily closely sealed, in which the fetishes of the Warrior Society are kept. Masked dancers among the Pueblos are called Katcinas, and the masks they wear would naturally be kept in a house (kihu) called "Katcinakihu."

[41] A passage or inclosure surrounded by high walls is called kisombi by the Hopi.

[42] On the top of the rock that forms the foundation of the walls of these rooms, and south of them, are hollows or grooves where the metates were ground, and shallow pits used in some prehistoric game. There are similar pits in some of the kiva floors.

[43] The word kiva, now universally employed in place of the Spanish designation "estufa" to designate a ceremonial room of the Pueblos, is derived from the Hopi language. The designation is archaic, the element ki being both Pima and Hopi for "house." It has been sought to connect this word with a part of the human body, and esoterically the kiva represents one of the underworlds or womb of the earth from which the races of man were born. It is highly appropriate that ancient ceremonies should take place in a kiva, the symbolic representation of an underworld, for many of the ceremonies are said to have been practiced while man still lived within the Earth Mother. The word kiva is restricted to subterranean chambers, rectangular or circular, in which secret ceremonies are or were held, and the term kihu is suggested for ceremonial rooms above ground. The five kivas at Walpi are examples of the true kiva, while the Flute chamber may be called a kihu.

[44] The so-called "warrior room" in Spruce-tree House belongs to the second type.

[45] In certain ceremonies of Hopi women's societies the kiva has also come to be a meeting place for these sororities and where they erect their altars.

[46] These small holes, generally square, are usually found in the wall below the banquette.

[47] The fire in these rooms was more for light than for heat, for when roofed a large fire would have produced so much smoke and heat that the occupants would be driven out. The character of the ashes indicates that logs were not used as firewood, but that the prescribed kiva fuel was, as at Walpi, small twigs or brush. No evidence of lamps has been found in cliff-dwellings, the lamp-shaped pottery objects having been used for purposes other than illumination.

[48] Cosmos Mindeleff quotes from NordenskiÖld a description of a Mesa Verde kiva, the deflector of which was made in the same way.

[49] On this supposition the large banquette may have been the forerunner of the spectator's section in the modern rectangular Hopi kivas, of which it is a modification.

[50] The measurements of the kivas here given were determined by Mr. R. G. Fuller, who served as voluntary assistant during the summer.

[51] With the exception of that in kiva Q there has not been found in any deflector a large stone ("fire stone") forming the cap or top. In deflectors formed of a slab of stone such a "fire stone" on top would be impossible.

[52] This kiva, which is one of the best in Cliff Palace, is illustrated by NordenskiÖld.

[53] A similar deflector is recorded by Mr. Morley as existing in the Cannonball ruin, and is figured by NordenskiÖld from the Mesa Verde.

[54] This entrance in the wall appears in all photographs of this portion of Cliff Palace.

[55] For a note on a similar vase and its use, see remarks on kiva S. It is probable that these dried lizards were regarded by the Cliff Palace priests as very potent "medicine."

[56] From all appearances the kivas were plastered from time to time after the walls had become blackened.

[57] This kiva, one of the finest and in some features the most exceptional in Cliff Palace, is not indicated in NordenskiÖld's plan.

[58] NordenskiÖld describes a ventilator constructed in the same way.

[59] In ceremonial rooms of ruins in the Navaho National Monument this curve is represented by a raised step.

[60] Among the Hopi at the present day certain fetishes, as the effigies of the Great Plumed Serpent, are regarded as so sacred that when not in use they are kept in jars set in a banquette, the surface of which is level with the neck of the jar. These receptacles are closely sealed with a stone slab when the images are deposited in them. Possibly the jars set in the kiva banquettes of Cliff Palace may have been used for a similar purpose: i. e., were receptacles for fetishes held in such veneration that, as is the case with the Great Serpent effigies of the Hopi, one even touching them may, in the belief of the people, be afflicted with direful disorders.

[61] See Bulletin 50, Bureau of American Ethnology.

[62] The two circular kivas of Kukuchomo, near Sikyatki, have this large banquette and in other respects resemble the ruins of Canyon de Chelly. Kukuchomo marks the site of a settlement, of the Coyote clan of the Hopi in prehistoric times.

[63] As a huge rock had fallen from the roof of the cave in which this kiva lies, since it was first occupied, it would appear that the place was abandoned on that account.

[64] NordenskiÖld's description of this kiva has been quoted earlier in this paper. In the description of a ceremonial room of a somewhat similar or of the same type in Spruce-tree House the term "warrior room" is used; there is nothing to warrant this designation, however, and it would be better to consider it simply as a kiva of the second type.

[65] For instance, the complicated reredos of many of the modern Hopi altars is made of flat wooden slabs, the manufacture of which would be very difficult for a people ignorant of iron. These probably replaced painted stone slabs of simpler character, examples of which have been found in ruins and indeed still survive in some of the oldest rites.