The inhabited pueblos, like the ruins, usually consist of a number of rooms built adjoining or on top of each other, like a lot of square boxes placed in rows or in a pyramidal pile, or like a series of steps, with the total height at the back often straight down. One or two single-room houses are first built, and then additions are made from time to time till the pile grows to a considerable height; three or four stories.[197] Groups of these groups built near each other form courts between. The lower tier of rooms, in olden times, was not entered from the ground, but from the roof through a hatchway, a ladder leading up on the outside and down on the inside. The upper rooms, or houses, were entered from the roofs of the lower ones; that is, the roofs of the lower rooms formed the floors of the upper ones, and also balconies in front of the rooms. I occupied for a time one of these upper rooms in Tewa, on the “East Mesa” at the Moki towns, and I found the roof in front of my door a delightful place, commanding a view of the whole mesa and a hundred miles beyond. I could also reach the top of my house easily, by a sort of stairway formed on the edge of the prolonged wall that separated me from my neighbour, and as this was the summit of the village my view was superb. Such stairways are common in all the villages. The ladders by which the various roofs are reached are now much like our own, but rudely made, and the upper ends are often very long, extending in many cases far above the house-top. The walls, about a foot thick, are of stone slabs laid in adobe mortar, and are generally built up by the women, who take their own time to the work, adding a few stones whenever they feel like it. Beams of small tree-trunks, six to eight inches in diameter, form the basis of the flat roof.[198] They are laid across the top of the walls and the ends, if too long, usually allowed to project beyond. These are covered with smaller poles laid about a foot apart, and on these are spread slender willows or reeds, with a layer of grass or twigs next, on which a layer of adobe mortar is laid and earth trodden down on top till it is firm, when a finish is made with another coat of adobe mortar. A slight pitch is given to the roof. No plumb-line, level, or square was used by the Amerinds anywhere on the continent so far as is now known. Sometimes the floors are paved with irregular flat sandstone slabs, but in most houses the floor is formed by a coat of adobe mortar which is patched and renewed as needed. Moccasined feet are not hard on such a surface, but my heavy soled shoes were the despair of the owner of my habitation. The hand is used as a trowel. The chimney is usually at one corner, and did not exist in America previous to the sixteenth century. A hood is built down from the roof to within about three or four feet of the floor, to catch the smoke, and outside the chimney is built up about three feet, sometimes with stones, but more frequently with large earthen pots with the bottoms knocked out. The hood is formed of sticks plastered with adobe mortar. Doorways were formerly of the notched variety[199] closed by a curtain, and the hatchways were closed by a mat of reeds. In later times the doorways have become like our own, and doors, too, have been made out of sawed boards. My door at Tewa was hung on hinges and had a latch and string. Glazed windows have also been adopted in many houses. The Rio Grande pueblos are built of adobe bricks, and so, largely, is Zuñi, but there is little adobe in the Moki towns, except in the form of plaster and mortar. The Rio Grande pueblos were largely constructed of adobe when first visited in 1540. The Pueblo Amerind frequently abandoned his village for one cause or another and built a new one elsewhere, so while his village may be called a permanent one it was not much more so than villages of the Algonquins and Iroquois.

Besides houses, some of the Amerinds of the South-west dwelt in shelters excavated wholly or in part in the face of a cliff or mountain, or hill. There are four localities where these cavate lodges occur in numbers, the northern Rio Grande valley, the San Juan River valley, the San Francisco mountain region, and the Rio Verde valley in Arizona. There are in these places thousands of cavate lodges. They average in size two or three rooms, sometimes communicating by a ledge, sometimes, often, in fact, with excavated connections. Some of the Verde group[200] are cut back a long distance into the rock—forty or fifty feet. The rooms are both oblong and circular, about seven feet high and ten by seventeen feet in size, or eight or ten feet diameter, according to the shape. There were no chimneys, the fire-pits being near the entrances. Nor were there any windows, the doorway being the only opening to the outside. Floors were levelled by filling depressions with adobe clay and low ridges were built up of the same material, probably to keep the inmates off the bottom, which must have sometimes been damp. Poles or willows laid across the ridges with skins on them would have made a flooring. Depressions at the back walls appear to have been made to hold water, and doubtless at times there was a “seepage” of considerable amount, as I have suggested regarding the open conchoidal caves occupied by the Cliff-dwellers. What appear to be stepping-stones are found in some entrances, as if water at times flowed out. The Verde group are in a soft grey sandstone, the Rio Grande in tufa, the San Francisco in cinder hills. These cavate dwellings are simply another form of Amerind residence due to necessity or expedience.[201] In other places there are some that were undoubtedly merely farming outlooks, occupied only during the crop season, just as there are cliff houses for this purpose, and also houses erected singly in open valleys. But many cavate lodges were actual residences for a period of years, owing to circumstances of one kind or another. The Cliff-dwellers may still be found among the Tarahumaris of northern Mexico. Schwatka describes some who “had walled up the outward face of a cave nearly to the top, leaving the latter for ventilation.” Many small cliff-dwellings in other places were so made to allow the smoke to escape. That is, the wall along the outer edge of the cavity was not carried quite up to the rock above, so that the smoke could drift out. There was, therefore, no roof over the dwelling, but it was sheltered by the overhanging rock. Many more examples of this adaptation of the dwelling to circumstances might be added.

There are ruins scattered all over the South-west, many of which were built by the same set of Amerinds, and do not represent a vanished population. Still, I believe that the population was at one time much greater than it was when our acquaintance with it began. Internecine wars resulting from a diminution of water-supply; diseases introduced by the whites; and also the attacks and absorption of tribes by the wilder Amerinds, being some of the causes of the diminution. It would not be possible to describe even all the prominent ruins here, but I will mention several. Beginning easterly of the Rio Grande, we find the Pecos Ruins first of importance. There are also remains of a large adobe Catholic church and a convent here, not finally and fully abandoned till about 1840. The ruins consist of two chief buildings on a low table, surrounded by an artificial wall. The buildings were in the form of rectangles, with courts within, one 55 by 440 feet, and the other 170 by 350 feet. In places, they were three or four stories high, terraced, Pueblo fashion. The construction was slightly different from the ordinary, as the upper floor and roof beams rested mainly on heavy upright posts set into the walls, and not directly on the walls themselves. The whole framework was thus independent of the enclosing walls, very much as our modern steel frame buildings are. The walls were of sandstone slabs, and were from one to two feet thick. Another group of important ruins, and about the finest specimens of the stone buildings of the ancient Pueblos, is that of the Chaco, in north-western New Mexico.[202] There are eleven chief ruins, and many smaller ones. The principal ruins were once houses three, four, or perhaps five stories high, all built of sandstone slabs and blocks obtained from the débris of the cliffs. Some of the walls are still standing to the height of thirty or forty feet. All are not uniform in the way the stones are laid, the variation being due to building at different times, and to a variation of the available supply of slabs. The stones were usually laid so closely, and so carefully chinked with spalls, that the outside of the walls resembles a smooth mosaic; though adobe mortar and rubble were freely used in the interior. Lintels, as was generally the case throughout America, were of wood. The date of the abandonment of these buildings is not known. They were first mentioned by Gregg, in 1844.[203]

There were many round towers of stone in the South-west, also the work of the Pueblos. Some stand alone but most of them are near other ruins. Often they were built with two or three concentric walls from two to five feet apart. A double-walled tower on the Mancos had an outer diameter of forty-three feet. Some of them may have been watch-towers, but those connected with other buildings were perhaps religious structures, or were used somewhat as the kiva[204] is to-day. The kiva is a room now usually square, in part, or wholly, below the general surface of the locality, used as a kind of club-house, council-house, lounging place, and meeting place for members of the society to which it belongs; and also a lodging place for the men; women are generally excluded. In Zuñi, the kivas are rooms on the ground floor. Many ancient kivas were round.

Adobe brick and adobe clay in various forms were largely employed by the South-western and Mexican Amerinds, and there are evidences that some tribes in the Mississippi valley also used it. In the Rio Grande valley the adobe is made into large bricks, sun-dried and laid up with a mortar of the same material. Otherwise the villages are much the same as those described. One of the best modern examples of the adobe construction is the village of Taos in north-eastern New Mexico. (See illustration page 3.) Another method of employing adobe is seen in the famous ruin called Casa Grande, near Florence, Arizona, which our government recently repaired so that it will endure for a considerable time. This was made by the cajon method; that is, the adobe mud was rammed into large chests or boxes of wicker, without top or bottom, and when the material was sufficiently dried to hold its shape the frame was removed and the operation repeated till the wall was finished. The ruin referred to is only one of a number that were still standing in an area of about sixty-five acres in 1744, when Father Sedelmair saw them. He described the present ruin as having four stories, but only three are now distinguishable at the highest part. Its age is unknown. Its builders are supposed to have been the ancestors of the present Pimas, though probably there was considerable difference in the matter of culture. Father Kino, in 1694, was the first European to see the place and it was a ruin then. It was doubtless a communal dwelling like all the other large structures of the Amerinds of this region. Its size on the ground is forty-three by fifty-nine feet. Partitions three or four feet thick divide the interior into five rooms, the middle one having higher walls than the rest. The adobe blocks are two feet high, three to five long, and three or four across, and are almost as hard as sandstone while dry. There may have been upper stories of plastered wattled posts. Another famous ruin similar to this is the Casas Grandes in Chihuahua, Mexico. It is built in much the same way as Casa Grande, and there are more buildings there standing. Probably there were at one time a great many structures of this kind in that region, and there may be others still standing in less explored parts. In the Salt River valley many of the buildings were of a somewhat different type again, as concerns their wall construction.[205] The Hemenway Expedition excavated a great many sites and found that the walls were often adobe rammed in between two series of posts wattled with reeds and cross-braced with sticks, the outer part of the wattled frames being plastered with adobe mortar. The thinner walls were constructed with only one line of wattled posts plastered on both sides, after the manner of the Mexican construction known as jacal, which is a set of poles fixed in the ground and then plastered on one or both sides with mud. The upper stories of some of the Rio Grande structures in the early times were made of wood probably plastered this way, which explains why in the southern part of New Mexico there are not now found higher standing walls of ruins or evidences of several stories.[206] Examples also have been seen in South-western Colorado, where a kind of wicker-work was built on the top of a wall and plastered on both sides. In the Salt River ruins the existence of the wood-work was indicated by the cavities left by its decay. There were also other structures built without the wattled frames. The cajon and pisé construction are very much alike, one being a Spanish and the other a French term, except that any pounded or rammed earth construction might be pisé, while the cajon is distinctly made by ramming earth into a box.[207] Therefore the Casa Grande would be a clear example of cajon, while the Salt River construction of adobe rammed between the wattled frames would be pisé; and the plastered wicker-work would be jacal. The pisé and cajon method is very old all over the world. It is still to be found in France and England. In France the pisé box is about three yards long, one yard high, and about half a yard wide. The readiness with which the Amerind took advantage of his resources in the architectural line is again apparent in these great adobe structures of the Amerinds of northern Mexico and the South-western United States. It is not sensible, therefore, when some style of construction is discovered differing from that which we have been accustomed to see, to ascribe it to some mysterious race.

In southern Mexico they erected extensive cities or pueblos because there they were more crowded together than anywhere to the northward, but these cities were essentially the same as the more simple towns in the northern country. At Tlascala “the houses were built, for the most part, of mud or earth; the better sort of stone and lime, or bricks dried in the sun. They were unprovided with doors or windows, but in the apertures for the former hung mats fringed with pieces of copper, or something which by its tinkling sound would give notice of anyone’s entrance. The streets were narrow and dark.”[208] This extract from Prescott might picture a New Mexican pueblo instead of one of the towns encountered by Cortez which have been often so romantically described. The copper on the mats was probably more for Amerind ornament than for the purpose stated by Prescott. While in some respects the Aztec towns may have been more elaborate than the New Mexican towns, there was probably not much difference in their method of construction. “The principal buildings and temples of the city were covered with a hard white stucco which glistened like enamel in the ... morning sun.”[209] This was perhaps a wash of gypsiferous clay similar to that used by the Mokis, or it may have been similar to the zahcab of the Mayas, which was a singular and abundant white earth used by them as a stucco. It was found in pockets.

“The dwellings of the common people were also placed on foundations of stone which rose to the height of a few feet, and were then succeeded by courses of unbaked bricks, crossed occasionally by wooden rafters.”[210] These rafters were the projecting ends of the poles, as in the Pueblo country. The adobe houses in Mexico are now often built on stone foundations, for it is the foundation that is sapped and undermined by the rains. The upper walls of adobe stand well in a climate of that sort. Prescott says of the houses of the “dignitaries” and of the “principal nobles” that “They were low, indeed; seldom of more than one floor, never exceeding two. But they were spread over a wide extent of ground; were arranged in a quadrangular form, with a court in the centre,”[211] all of which sounds suspiciously like a communal dwelling, as Morgan maintains the Aztec houses were. The Aztecs were crowded around the lake of Mexico, and also built out over the water on piles. Houses raised above the water or ground were nothing unusual in America. Some of the North-west coast Amerinds built dwellings which were “raised and supported near thirty feet from the ground by perpendicular spars of very large size” with “access formed by a long tree in an inclined position from the platform to the ground, with notches cut in it by way of steps about a foot and a half asunder.”[212]

So far as the Aztec houses are concerned, “None of the Spanish descriptions,” asserts Morgan, “enable us to realise the exact form and structure.... But for the pueblos, occupied or in ruins, in New Mexico, and the more remarkable pueblos in ruins in Yucatan and Central America, we would know very little concerning the house architecture of the Sedentary Village Indians.”[213]

Morgan believes all were joint tenements, but in this he may be mistaken, for the life of the Aztecs seems to have passed to a point somewhat higher than that of the New Mexican Amerinds, and a further development of Aztec life certainly included a further development of their house-life also.

Within a day’s journey of the City of Mexico, Saville investigated some interesting ruins of an old “temple” erected, according to a tablet found there, in 1502, the signs on the tablet representing a rabbit and ten dots, or ten tochtli, corresponding to this date. It was built of rubble stone covered in many places inside with stone carving that had been painted.[214] There were also ornaments in stucco. The outer walls are nearly six feet thick. It is on the top of a high, cliff-like mountain difficult of access, near the Mexican town of Tepoztlan. Another splendid ruin near this is the Temple of Xochicalco. See illustrations, pages 23 and 31.

The greatest group of architectural remains on this continent is that of the Maya region, mainly in Yucatan. For a full description of many of these buildings the reader is referred again to the admirable work of Maudsley. The Mayas were the greatest architects as well as the greatest artists and greatest in almost everything of all the Amerinds, and if Goodman is correct in his rendering of some of their chronology they occupied the region more than ten thousand years.[215] Mound-like foundations supported the buildings, which generally rose as from a terrace, though sometimes the mound was very high and very steep, with small space around the building crowning it. At Copan,[216] which was in ruins before the Spaniards arrived, there is a great main terrace from which mounds rise, the latter bearing the buildings. The casing of the mound and the walls of buildings are of nicely dressed oblong stones usually without mortar. The joints were not broken here, nor in other Maya work. The mound slopes were terraced at five-foot intervals and the steps were about five feet high. The so-called “triangular arch” probably existed here as it did at the other Maya ruins. It was made by advancing the courses, several feet above the base of an opening, gradually toward each other till they met above, where a large slab was usually laid across to bind the whole together. The ceilings or roofs of many rooms in Maya ruins were wholly made this way. It has also been called a corbel arch, though it is, in fact, not an arch at all. See illustrations, pages 210, 235, and 237. There was no arch in Amerindian architecture besides the one the Eskimo constructed in his snow hut. The rooms are generally long and narrow in all the Maya structures and no windows existed. The Maya inability to span wide spaces was the cause of the narrow rooms and buildings. At Uxmal the two main rooms of the so-called Governor’s Palace are sixty feet long and only eleven to thirteen feet wide. The walls of all the structures are very thick, though certain walls, as the rear ones, are usually thicker than the others and have no openings, the latter, as a rule, being along one, two, or three sides. This was a probable survival of earlier defensive constructions similar to the communal fortresses of the Puebloan type as particularly exemplified in the ruins of the Chaco in New Mexico, where there were no rear openings. See ground plans, page 232. At Palenque are some fine examples of the Maya construction. The largest is called the palace and is 180 feet wide, 228 feet long, and 25 feet high, with fourteen doorways on the side and eleven at the ends. It was one story in height, as were all Maya buildings. There is a vast amount of carving and stucco modelling around them. One of the most unique constructions is that called the “Temple of the Cross,” number one, or two, or three, by different explorers, there being two structures much alike. See note 2, page 184. This is on top of a high mound, and is fifty feet front, thirty-one feet deep, and about forty feet high. The roof was something like our gambrel type, being the same all around without gables, with a level platform about three feet wide along the ridge, from which arose a peculiar stone and stucco, latticed, superstructure in two stories, the first about seven and the second about eight feet high. See illustrations, pages 210 and 235. There was abundant stucco ornamentation over the exterior, and on each side of the entrance was one of the figures referred to in the last chapter.

The mortar used is said to have been a cement made of one part slaked lime to two parts of zahcab. This was used by all the ancient Mayas and is used still in that country. It is, however, doubtful if slaked lime was known to the ancients. There is no evidence of it. At Mitla is yet another type of house ascribed to the Zapotecs.[217] It is in the Mexican State of Oaxaca. The human figures and animal carvings and forms seen in the Yucatan ruins are absent. The rooms are the same, long and narrow, with no openings except the doors. One of the most unusual features is a great hall 12 by 121 feet, with six round stone columns standing at intervals of about fifteen feet down the middle. See illustrations, pages 9 and 209. These average about twelve feet high and nine feet in circumference. The walls are forty-eight inches thick, of roughly broken stones laid in courses in plenty of adobe mortar, the outer parts of all the buildings being faced by slabs of stone containing the ornamentation, which is wholly geometrical. Some adobe brick walls are forty-six inches thick. The columns are out of the common because they are single stones, but built up piers are often used in Pueblo architecture, and the North-west coast Amerinds use the column in wood very frequently to support their large longitudinal rafters. One of these which I sketched in an Alaskan house at Cape Fox is given in the illustration, page 162. The roofs at Mitla were wooden beams covered with earth and stone slabs. See illustration, page 230. There are other ruins all through Honduras and Nicaragua and the rest of Central America. Squier says: “In Honduras, as also in San Salvador, I heard of remains and monuments equal to those of Copan in extent and interest.”

At the time the Spaniards came into Yucatan the Amerinds, according to Herrera, were dwelling in timber huts thatched with grass or something similar. The dense unexplored forests of the Yucatan region are filled with ruins which have never been seen by white men, at least that is the supposition of archæologists like Saville and Charnay. The Maya house was divided, according to Landa, from side to side by a wall with doors, the back part being sleeping quarters. The front portion was whitewashed or painted in designs and was open the whole length, with low sheltering eaves. In the rear there was a doorway leading from that part. A lengthwise division into two main parts was a characteristic of almost all the Maya buildings now found in ruins. The structures were generally wide and shallow, and subdivided into a great many rooms. It is more in the ornamentation of the buildings and the stone roofs than in anything else that they differ from structures farther north. The interior masonry is frequently a rubble, with the dressed and carved stones on the outside as a facing. Bandelier thinks that some of the stone walls in New Mexico are quite as well constructed as some in Mexico proper. But however this may be, there is nothing north of the City of Mexico that compares in architectural excellence with the Yucatan structures, albeit in some respects there is a strong resemblance between the latter in plan and conception, and the Pecos and other northern ruins.

The communal principle of living had much to do almost everywhere with the size and character of the Amerind houses. Situation was determined by expedience and necessity; material of construction by environment. Throughout the continent the Amerind was a village dweller, and except in the Far North and on the northern Californian and North-west coasts he was generally a tiller of the soil, growing, often in large quantities, maize, beans, squashes, cotton, and some other products according to locality. His large communal buildings were in part fortresses to protect the families against marauding Amerinds of a less prosperous and cultivated type, and against the occupants of other towns, for in general it may be said that there was little political cohesion in the various tribes, though the Aztecs and Iroquois are examples of exceptions that arose from time to time.

There is nothing in any of the remains, so far developed, that indicates foreign influence, prior to the Discovery. Every architectural work on the continent is purely Amerindian or modified by contact with other races subsequent to 1492.