A notice of the ancient remains of southwestern Colorado examined during the summer of 1875. By W. H. Holmes. A notice of the ancient ruins in Arizona and Utah lying about the Rio San Juan. By W. H. Jackson. The human remains found near the ancient ruins of southwestern Colorado and New Mexico. By D. Emil Bessels.

A NOTICE OF THE ANCIENT RUINS OF SOUTHWESTERN COLORADO, EXAMINED DURING THE SUMMER OF 1875.

By W. H. Holmes.

In addition to my duties as geologist to the southwest or San Juan division of the survey for 1875, I was assigned the very agreeable task of making examinations of such ancient remains as might be included in the district surveyed.

Much information had already been given to the public in relation to the ruins of Southwestern Colorado by Mr. Jackson, who paid them a short visit in 1874, and many similar remains had been described by early explorers in New Mexico and Arizona, but nothing like a complete survey of this particular region had been made.

The district examined by our party covers an area of nearly 6,000 square miles, chiefly in Colorado, but including narrow belts in the adjacent Territories of New Mexico, Utah, and Arizona. It lies wholly on the Pacific slope, and belongs almost entirely to the drainage-system of the Rio San Juan, a tributary of the Colorado of the West.

Lying along the west base of the mountains is a comparatively flat country, the eastern border of the great plateau-region that reaches westward toward the Sierras. The surface-geology is chiefly Cretaceous, and the various large streams formed on the west slope of the Rocky Mountains have cut long canoned valleys down through the nearly horizontal beds. In the greater part of this region, there is little moisture apart from these streams, and, as a consequence, vegetation is very sparse, and the general aspect of the country is that of a semi-desert. Yet there is bountiful evidence that at one time it supported a numerous population; there is scarcely a square mile in the 6,000 examined that would not furnish evidence of occupation by a race totally distinct from the nomadic savages who hold it now, and in every way superior to them.

At first, it seems strange that a country so dry and apparently barren could support even a moderate population, and it is consequently argued that the climate has grown less moist since the ancient occupation. Be this as it may, I observe the fact that the great bulk of remains are on, or in the immediate neighborhood of, running streams, or by springs that furnish a plentiful supply of water during the greater part of the year. The ever-present pottery may in many cases have been broken and left by hunting and wandering parties, and the remnants of dwellings far out from water may be but temporary abodes used only in the winter or during rainy seasons.

I also notice that the country is by no means an entire desert. All along the stream-courses there are grass-covered meadows and broad belts of alluvial bottom, affording, if properly utilized, a considerable area of rich, tillable land.

The ruins of this region, like most others of the extreme West and South, are the remnants in a great measure of stone structures. To

what extent wood and adobe were used can hardly be determined. It is evident, however, that the greater portion of the villages and dwellings of the lowlands have been of material other than stone, frequently doubtless of rubble and adobe combined.

As to situation, they may be classified very properly under three heads: (1) lowland or agricultural settlements; (2) cave-dwellings ; and (3) cliff-houses or fortresses.

Those of the first class are chiefly on the river-bottoms, in close proximity to water, in the very midst of the most fertile lands, and located without reference to security or means of defense.

Those of the second are in the vicinity of agricultural lands, but built in excavations in low-bluff faces of the Middle Cretaceous shales. The sites are chosen also, I imagine, with reference to security; while the situation of the cliff-houses is chosen totally with reference to security and defense, built high up in the steep and inaccessible cliffs, and having the least possible degree of convenience to field or water.

As to use, the position for the most part determines that. The lowland ruins are the remains of agricultural settlements, built and occupied much as similar villages and dwellings would be occupied by peaceable and unmolested peoples of today. The cave-dwellers, although they may have been of the same tribe and contemporaneous, probably built with reference to their peaceable occupations as well as to defense, and it is impossible to say whether or not they made these houses their constant dwelling-places. The cliff-houses could only have been used as places of refuge and defense. During seasons of invasion and war, families were probably sent to them for security, while the warriors defended their property or went forth to battle; and one can readily imagine that when the hour of total defeat came, they served as a last resort for a desperate and disheartened people.

In form, the parallelogram and circle predominate, and a considerable degree of architectural skill is displayed. Where the conformation of the ground permits, the squares are perfect squares and the circles perfect circles. A greater part of the ordinary structures are square or rectangular; while attached to each group, and sometimes without indications of contiguous buildings, are the circular ruins frequently resembling towers. These are the most pretentious structures, being often as much as forty feet in diameter, and in many cases having double or triple walls. They are solidly built of hewn stone, dressed on the outside to the curve, neatly jointed, and laid in mortar.

The space between the outer walls is invariably divided by heavy partition-walls into a number of apartments, while a circular depression, or estufa*, occupies the center of the inclosure.

It seems evident, from the extraordinary form of these structures and the unusual care shown in their construction, that they were not designed for the ordinary uses of dwelling or defense. It has been observed that, among nearly all the ancient tribes of North America, the grandest and most elaborate works of art were the offspring of their superstitions, and it does not seem at all improbable that these great towers had a religious origin.

It is stated that the eternal fire—an essential of their worship—has always been kept in circular inclosures, and that the circle symbolizes the sun, their deity. The occurrence, therefore, of one or more of these circular inclosures in each of their settlements can be rationally accounted for; but it is with less certainty we arrive at conclusions in regard to the triple walls and the cell-like apartments. In the inhabited

--------------------------------------------------------------------------------

* A Spanish word signifying "sweat-house" or council-house.

pueblos of to-day, there are underground rooms, mostly circular, used as council-chambers as well as for the performance of the mysterious rites of their religion. Similar chambers occur, according to Lieutenant Simpson, in all the ruined cities of New Mexico, but having single walls of no great height or thickness. It is stated by Squier and Davis* that in Mexico the sacred inclosures were also used for defensive purposes, and it certainly seems probable that these curious structures served the double purpose of temples and fortifications, and that the apartments between the walls were the cells of the priesthood or the receptacles of sacred or valuable property.

The smaller single-walled towers, which are scattered at intervals along the river-courses and canons, frequently in commanding situations, were probably watch or signal towers.

The cave-dwellings are made by digging irregular cavities in the faces of bluffs and cliffs formed of friable rock, and then walling up the front, leaving only a small doorway for entrance and an occasional small window at the side or top.

The cliff-houses conform in shape to the floor of the niche or shelf on which they are built. They are of firm, neat masonry, and the manner in which they are attached or cemented to the cliffs is simply marvelous. Their construction has cost a great deal of labor, the rock and mortar of which they are built having been brought for hundreds of feet up the most precipitous places. They have a much more modern look than the valley and cave remains, and are probably in general more recent, belonging rather to the close than to the earlier parts of a long period of occupation. Their position, however, has secured them in a great measure from the hand of the invader as well as from the ordinary effects of age.

Of works of art other than architectural that might assist in throwing light upon the grade of civilization reached by these people, but meager discoveries were made; although I imagine that careful search and well-conducted exhumation might develop many things of great interest. A small number of arrow-heads, stone-implements, ornaments, and articles of fictile manufacture, that may fairly be attributed to the age of the cliff-builders, were collected. The greater part of these are figured in plates XIII and XIV. There are no evidences whatever that metals were used.

Numerous hieroglyphics were observed, both engraved and painted upon the cliffs. Drawings of a large number were made, and some of the more notable examples are given in plates XI and XII.

A great number of burial-places were noted, but of the graves examined few yielded further evidences of occupation than small quantities of charcoal and bits of painted pottery. These burial-places, which are in a number of cases covered by a heavy growth of full-grown pinon pines and cedars, are usually found on the summits of high ridges and promontories, and are still marked by slabs of sandstone set on edge and arranged in circles, and parallelograms of greatly varying dimensions. But that they did not always bury their dead in high places is proven by the frequent discovery of human remains in the arroyos or deep washes in the valleys. Three skeletons were obtained, in the vicinity of ruined villages, from the sides of recent washes.

The accompanying plates are, with one exception, reproductions of pen-drawings, and are arranged for convenience, rather than from any method of classification. The plans are not drawn to a uniform scale, because of the inconvenience of such an arrangement ; but measure

-------------------------------------------------------------------------------

* Ancient Monuments of the Mississippi Valley, page 102.

merits are so frequently given on the plates themselves that no confusion need occur. Measurements were taken by tape-line in all the more important structures; but in many of the ordinary ruins, where exact dimensions were not considered essential, the distances were estimated. It is to be greatly regretted that extreme haste frequently prevented close and accurate work.

The accompanying map will give the location of all the more important-groups of ruins.

The fragments of information collected in this notice are given with the hope that they may throw a little additional light upon the very interesting problems of ancient American history.

RUINED VILLAGE ON THE RIO LA PLATA. Plate I.

The first group of ruins observed is situated on the Rio La Plata, about twenty-five miles above its junction with the San Juan, and three or four miles south of the New Mexican line. It is doubtless the remains of a large, irregular village, and stands on a low terrace, some 20 feet above the river-bed, and near the center of a large, fertile valley.

It will be seen, by reference to the plate, which includes only the more important part of the town, that the buildings have been isolated, and, in a measure, independent of each other, differing in this respect from most of the groups of ruins farther south.

The forms are chiefly rectangles and circles; one or two seem to have been elliptical, while a number have consisted of irregular groups or clusters of apartments. All that now remains to mark the site of these ancient structures is the low, rounded heaps and lines of debris, composed of earth, water-worn pebbles, and small fragments of sandstone. The walls of four of the main structures are quite distinctly marked. That of the circle c is still 4 feet high on the outside, and incloses a depression, probably an estufa, which, in the center, is 2 or 3 feet below the terrace-level.

North of this, about 300 feet, is a truncated rectangular mound, 9 or 10 feet in height and 50 feet in width by 80 in length. On the east end, near one of the angles, is a low, projecting pile of debris that may have been a tower. There is nothing whatever to indicate the use of this structure. Its flat top and height give it more the appearance of one of the sacrificial mounds of the Ohio Yalley than any other observed in this part of the West. " It may have been, however, only a raised foundation, designed to support a superstructure of wood or adobe.

North of this again, and 100 feet distant, is a rectangular inclosure about 60 by 100 feet. It is slightly excavated in the center, and the rounded and irregular wall is from 4 to 6 feet in height. The space between this and the last-mentioned structure is filled in to the depth of 2 or 3 feet, and the amount of debris about their bases indicates original walls of considerable height. North of this are scattered a number of inferior ruins, the walls of which are not always distinctly marked. These extend back toward a row of low hills, the remnants of a superior terrace, on the summits of which a number of artificial depressions were found. Such "dug holes ^v are generally quite numerous in the vicinity of these ruins, and have doubtless in many cases been made by throwing up earthworks for defensive purposes. South of the large circle is a mass of ruins covering some 15,000 square feet, but so much reduced that nothing further could be determined than the fact that it had contained a large number of irregular apartments. Next to this is a rect-

angular ruin, containing three well-marked apartments. Its walls are 6 or 7 feet high, and, unlike those of the preceding examples, do not coincide with the cardinal points. South of this, and occupying the extreme southern end of the terrace, are a number of small circles and mounds, while an undetermined number of diminutive mounds are distributed among the other ruins.

To the east of the Indian trail, as shown in the plate, are a number of inclosures of lesser importance, which, from want of time, were not closely examined.

Nowhere about these ruins are there any considerable indications of defensive works, and the village, which is scattered over an area fully two miles in circuit, has no natural advantages whatever. Neither are there traces of ditches, nor of anything that might throw important light upon the habits or occupations of the people. A few arrow-heads and minute cutting-implements were picked up. Countless chips of jasper, obsidian, and flint were scattered around, and the soil was literally full of fragments of painted and ornamental pottery.

On the opposite side of the river, and at intervals above and below, are isolated groups of ruins and heaps of debris—certainly the remains of dwellings. These seem to have been distributed very much as dwelling-houses would be in the rural districts of civilized and peaceable communities.

It is possible that there are undiscovered ruins on this stream equally important with that described; for, in pursuing my geologic investigations, I was compelled to take a long detour to the westward from this point, returning to the La Plata again a few miles above its junction with the San Juan. On this occasion, while riding through a desert-like locality, quite naked and barren, much resembling the well-known Mauvaises terres, I was surprised to observe fragments of pottery strewn around, and presently a number of ruins, in a very reduced state and almost covered by the drifting sand, and this six or eight miles from water. On the high, dry table-lands, on all sides, fragments of pottery were picked up. What could have induced people to build and dwell in such a locality it is useless to surmise.

GROUP OF CAVE DWELLINGS AND TOWERS ON THE RIO SAN JUAN.

Plate II.

On the San Juan River, about thirty-five miles below the mouth of the La Plata and ten miles above the Mancos, occurs the group of ruins figured in Plate II.

The river is bordered here by low lines of bluffs formed from the more compact portions of the Middle Cretaceous shales. At this particular place, the vertical-bluff face is from 35 to 40 feet in height.

I observed, in approaching from above, that a ruined tower stood near the brink of the cliff, at a point where it curves outward toward the river, and in studying it with my glass detected a number of cave-like openings in the cliff-face about half-way up. On examination, I found them to have been shaped by the hand of man, but so weathered out and changed by the slow process of atmospheric erosion that the evidences of art were almost obliterated.

The openings are arched irregularly above, and generally quite shallow, being governed very much in contour and depth by the quality of the rock. The work of excavation has not been an extremely great

one, even with the imperfect implements that must have been used, as the shale is for the most part soft and friable.

A hard stratum served as a floor, and projecting in many places made a narrow platform by which the inhabitants were enabled to pass along from one house to another.

Small fragments of mortar still adhered to the firmer parts of the walls, from which it is inferred that they were at one time plastered. It is also extremely probable that they were walled up in front and furnished with doors and windows, yet no fragment of wall has been preserved. Indeed, so great has been the erosion that many of the caves have been almost obliterated, and are now not deep enough to give shelter to a bird or bat.

This circumstance should be considered in reference to its bearing upon the question of antiquity. If we suppose the recess to be destroyed is six feet deep, the entire cliff must recede that number of feet in order to accomplish it. If the rock were all of the friable quality of the middle part, this would indeed be the matter of a very few decades; but it should be remembered that the upper third of the cliff-face is composed of beds of comparatively hard rocks, sandstones, and indurated shales. It should also be noted still further that at the base of the cliff there is an almost total absence of debris, or fallen rock, or even of an ordinary talus of earth, so that the period that has elapsed since these houses were deserted must equal the time taken to undermine and break down the 6 feet of solid rock, plus the time required to reduce this mass of rock to dust; considering also that the erosive agents are here unusually weak, the resulting period would certainly not be inconsiderable.

Figure 2 gives a fair representation of the present appearance of these dwellings, while their relations to the group of ruins above will be understood by reference to figure 1. These ruins are three in number— one rectangular and two circular. The rectangular one, as indicated in the plan 0, is placed on the edge of the mesa, over the more northern group of cave-dwellings; it is not of great importance, being only 34 by 40 feet, and scarcely 2 feet high; the walls are 1 1/2 feet thick, and built of stone.

The small tower B is situated on the brink of the cliff, directly above one of the principal groups of cave-houses. It is neatly built of stone, which, although not hewn, is so carefully chosen and adjusted to the curve that the wall is quite regular. That the stone was procured from the neighboring cliffs is indicated by the presence of great numbers of characteristic fossils. The wall is 18 inches thick and from 2 to 6 feet in height.

Long lines of debris, radiating from all sides, indicate that it has been much higher, and has but recently fallen. This tower is inclosed by a wall, also circular in form, but open toward the cliff, as seen in the drawing; the ends projecting forward and irregular and broken as if portions had fallen. Its construction is like that of the inner wall, but the height is not more than 3 feet at any point. The diameter of the inner circle is 12 feet, that of the outer 22 feet; the distance, therefore, between the walls is a little less than 4 feet. In this space, there are indications of partition-walls that have originally divided it into a number of apartments.

About one hundred and fifty yards to the southwest of this ruin are the remains of another similar structure. It has been, however, on a much grander scale. The walls are 26 inches thick, and indicate a diameter in the outer wall of about 140 feet. They are not above 4 feet high at any point, and in the parts toward the cliff can only be traced

by a low ridge of earth. The remaining fragments of wall are at the remoter parts of the circles, and are in every respect like the walls already described. The inner wall, which can be traced but a short distance, is 8 1/2 feet from the outer, and has been connected by partition-walls, as in the other case.

The first impression given by this curious inclosure is that it was designed for a "corral," and used for the protection of herds of domestic animals; but since these people are not known to have possessed domestic animals, and when we further consider that inclosures of pickets would have served this purpose as well as such a massive and extraordinary structure, we can hardly avoid assigning it to some other use, which use, doubtless similar to that of the smaller tower, is very naturally suggested by its location and construction.

That they both belonged to the community of cave-dwellers, and served as their fortresses, council-chambers, and places of worship, would seem to be natural and reasonable inferences. Being on the border of a low mesa country that rises toward the north, the strong outside walls were doubtless found necessary to prevent incursions from that direction, while the little community by means of ladders would be free to pass from dwelling to temple and fortress without danger of molestation.

The original height of these structures must necessarily be a matter of conjecture, and it is true that although there is every evidence of age, both in the cave-dwellings and in the walled inclosures above, the lack of great quantities of crumbling walls and debris, and the general bareness of the ruins, give rise to the notion that they were but meager affairs. If we conclude, however, that the outer walls were constructed for defense, and their thickness and form favor such a hypothesis, their height would probably have been as great as fifteen or twenty feet, while the inner walls, being equally heavy and well built, would be sufficiently high to accommodate two or three stories. With these conclusions in view, I have ventured to present a sketch showing a restoration of the smaller tower (plate III). This sketch illustrates the probable appearance of the dwellings and tower, and the supposed means of communication between them.

The manner of walling up the fronts of the cave-dwellings, as here given, was observed frequently on the Rio Mancos, where, in corresponding cliffs of shaly sandstones, there are many well-preserved specimens. A large group situated on this stream, about ten miles above its mouth, was subsequently examined. The walls were in many places quite well preserved and new-looking, while all about, high and low, were others in all stages of decay. In one place in particular, a picturesque outstanding promontory has been full of dwellings, literally honeycombed by this earth-burrowing race, and as one from below views the ragged, window-pierced crags, he is unconsciously led to wonder if they are not the ruins of some ancient castle, behind whose rnoldering walls are hidden the dread secrets of a long-forgotten people; but a nearer approach quickly dispels such fancies, for the windows prove to be only the doorways to shallow and irregular apartments, hardly sufficiently commodious for a race of pigmies. Neither the outer openings nor the apertures that communicate between the caves are large enough to allow a person of large stature to pass, and one is led to suspect that these nests were not the dwellings proper of these people, but occasional resorts for women and children, and that the somewhat extensive ruins in the valley below were their ordinary dwelling-places. On the brink of the promontory above stands the ruin of a tower, still twelve feet high, and similar in most respects to those already described. These

round towers are very numerous in the valley of the Mancos. From this point alone at least three others are in view, some on the higher promontories, others quite low, within twenty or thirty feet of the riverbed. I visited and measured seven along the lower fifteen miles of the course of this stream. In dimensions they range from ten to sixteen feet in diameter and from five to fifteen feet in height, while the walls are from one to two feet in thickness. They are in nearly every case connected with other structures, mostly rectangular in form. At the mouth of the Mancos, however, a double circle occurs, the smaller one having been the tower proper. It is fifteen feet in diameter, and from eight to ten in height. The larger circular wall is forty feet in diameter and from two to four feet high, and is built tangent to the smaller. This ruin is at the point where the Mancos reaches the alluvial bottom bordering the Rio San Juan, and about one mile above its junction with that river. On the opposite or south side of the river are traces of somewhat extensive ruins, but so indistinct that the character of the original structures cannot be made out, and indeed no single mile of the lower fifty of the Man cos is without such remains.

CANYON OF THE RIO MANGOS.

Fifteen miles from its junction with the San Juan, this stream emerges from the southwest border of the Mesa Verde, through which it has cut its way, producing a most remarkable canyon. This canon has already been once or twice described; but in order to make my subsequent descriptions better understood, I shall give here an outline. The Mesa Verde is a somewhat irregular table-land, comprising an area of about seven hundred square miles, and is formed of a great series of nearly horizontal sedimentary rocks, of which the surrounding country has been denuded. This series of strata consists, in the upper part, of massive sandstones; in the middle part, of alternating sandstones and shales; and in the lower one thousand feet, chiefly of shales and clay. These softer beds are, when once exposed to the erosive agents, carried away with great rapidity, and, as a consequence, the firmer rocks above are undermined and break down in vertical cliffs, and, where soft and hard beds alternate, a series of steps, with intervening slopes, is formed. It will readily be seen that a canyon thus formed would consist in general of a narrow, irregular river-bottom, long steep slopes of debris rising like the arms of a letter V from this, then a succession of steeps and slopes, culminating above in a series of lofty, embattled cliffs. The canyon is nearly thirty miles in length, and ranges from one to two thousand feet in depth. It seems to have been a favorite resort of the cliff-building people, and traces of their industry may be found everywhere, along the bottoms, in the cliffs, and on the high, dry tablelands above.

The four following plates will be taken up in the delineation of the more interesting portions of these remains.

PLATE IV.

Figure 1 illustrates the method of walling up the cave-fronts as described on a preceding page. This sketch was made at the last-mentioned locality on the Rio Man cos. The group occurred in the cliff about thirty feet from the base. The three doorways opened into as many small apartments, and these were connected with each other by very small passage-ways. The farther door could not be reached from the outside, as the platform of rock had broken away.

Figure 2 gives a plan of the double tower near the mouth of the Man-cos; it has already been described.

The ruin, of which a plan is given in figure 3, occurs on the left bank of the Mancos about eight miles above the foot of the canyon. It is one of the best preserved specimens of the ruined towers, and seems to have been built with much skill. It is 9 feet in diameter on the inside and about 16 feet high. There are three rectangular apartments attached, the walls of which are almost leveled with the ground. In the side of the tower facing the river is a window, about 8 feet from the ground and 2 feet high by 1 1/2 wide. I had been previously led to the conclusion that these towers were in all cases built without windows or openings of any kind within reach of the ground from without, and it is not improbable that this opening did not communicate with the outside, but served as a doorway between the tower and one of the adjoining apartments. The advantage of such an arrangement in a defensive work, such as we may suppose this to have been, is clearly apparent, and evinces not a little intelligence and forethought on the part of the builders. Being built in connection with dwellings and places of resort, they could, in case of alarm, be reached with ease from within, but be altogether secure from without.

This ruin was visited by Mr. Jackson last year, and a sketch of it has been published.

Figure 4. The large circular ruin, of which a ground-plan is given in this plate, was also visited by Mr. Jackson, photographs were made, and a brief description given; but I deem it best to give a more detailed description, the result of such observations and measurements as could be made in a period of time entirely too short for a work of such importance.

This ruin is situated on a narrow strip of alluvial bottom about midway in the canyon of the Mancos. On first approaching it, one does not observe that it differs greatly from the ordinary fragmentary structures below, as it is much decayed and almost hidden by artemisia and vines. Closer inspection, however, develops the greater part of the outline, and I imagine that a little excavation would bring all the foundations to light. The inner wall can be traced throughout the entire circle, and is in places 6 or 8 feet high. A portion of the outer wall, at the point farthest from the river, is still 12 feet in height and in a fair state of preservation. The space between the walls has been divided into cells_ras in the two examples given in plate II. Four of the cross walls are still a number of feet high, while others can be traced by lines of debris. The diameter of the outer wall is 43 feet; that of the inner, 25 feet. They are faced up with larger stones than usual (the heaviest of which, however, could be lifted with ease by a single workman), and have been filled in with rubble, adobe, and wood. The outside courses have been dressed to the curve, and the implements used, judging from the appearance of the picked surfaces, have been of stone. The main wails are 21 inches in thickness, while the partition walls are somewhat lighter, and seem to have been but slightly built into the circular walls.

In order to determine the probable number of these cells, I measured the two having complete walls, and found the inner side of each to be 8 feet. As these were both on one side of the circle, I had but to measure the remaining space to complete the semicircle, and on so doing found that there was just room for three additional cells and the necessary partition-walls; two of these were still traceable. To complete the circle, therefore, ten apartments would be necessary. Being desirous of confirming this conclusion, I took the diameter of the inner circle

as given in my notes, and, by adding twice the thickness of the wall, obtained a circumference of 89 1/2 feet; just sufficient space to accommodate ten apartments, with an equal number of partition-walls a fraction less than 12 inches in thickness.

By adding to the diameter of the inner circle the total thickness of the walls, plus twice the distance between them, I obtained a diameter of 43 feet for the outer circle. The circuit of the structure is, therefore, 135 feet. Although these figures are greater than those previously given (estimated), I am confident that they cannot vary greatly from the truth.

There were no indications of windows or doors in the fragment of outer wall, but two nearly rectangular openings in the inner wall seem to have served as doorways between the central inclosure and the cells. We may suppose that each cell had similar means of communication with the interior. The one doorway that remains entire is 6 feet from the ground, and measures 2 feet in width by 3 in height. The stone-work of the facing is very neat and exact, and the lintel is of a single slab of sandstone. It may be fairly presumed that the outer wall had no doorways or windows within reach of the ground, and that entrance was obtained by means of ladders, through high windows or by way of the roof. The central inclosure has doubtless served as an estufa, and there are still evidences of a considerable depression.

That this ruin is quite ancient is attested by the advanced stage of decay, and that it has been of considerable height may be inferred from the large quantities of debris. A similar and somewhat more perfect example of double-walled tower is illustrated in plate VIII.

There seem to have been no buildings of importance in connection with this ruin, but many in the vicinity. On the point of a low rocky promontory that extends down from the mesa on the west to within a few yards of the circular ruin are some masses of decaying wall, and a large circular depression, not differing in appearance from the usual estufa.

It is probable that there are other remains higher up on the rocky slope; indeed, others could be seen fro, the trail, but I found no time to visit them.

A few hundred yards below the great tower, and very near the trail, a smaller tower occurs, having other ruins connected with it, and in a weather-worn cavity in a massive crag near by is the cosy little dwelling shown in figure 5.

The rude little fire-place illustrated in figure 6 was observed by Mr. Brandegee in connection with a cliff house on the opposite side of the canon, a little farther up. It is remarkable as being the only example discovered by our party. There seem to be no traces whatever of fireplaces, ovens, furnaces, or chimneys in or about any of the ruins described, which is rather remarkable, since fires must have been used in baking pottery and for domestic purposes, and we cannot suppose that a people so well advanced in architectural skill were unable to build fire-places and furnaces.

PLATE V.—CLIFF-HOUSES OF THE MANCOS.

In this plate are given sketches of two of the more interesting groups of cliff-houses, or fortresses, as they may quite as appropriately be called. Figure 1 represents a portion of a group found about ten miles from the foot of the canon in a subordinate cliff on the west side. This low cliff is of massive sandstone, and is washed by the river, the trail being crowded back against the steep wall. At the height of about

forty feet above the river, a bed of shale occurs in the sandstone, which,being easily disintegrated, has been weathered out and carried away, leaving a sort of horizontal groove some four feet high and from four to six feet deep. In this a row of diminutive houses has been built. Three of these are almost perfect, having a fresh, new look that certainly belies their age. Four others are much more decayed, and fragments of wall only cling to the cliffs. They have been made to occupy the full height and depth of the crevice, so that when one reaches it at the only accessible point, he is between two houses and must pass through these to get at the others. The doorways are quite small and bear no evidence of the fitting or hanging of doors; and the windows, of which a number open to the front, are but a few inches square.

The walls are strongly built and are from eight to ten inches thick. The stones are small, dressed roughly on the outside, and laid in mortar.

In many places the heavier seams of mortar have been chinked with bits of pottery and small flakes of sandstone. The marks of the mason's pick are as fresh as if made within a few years, and the fine, hard mud-mortar, which has been applied with the bare hands, still retains impressions of the minute markings of the cuticle of the fingers.

The house at the left hand in the drawing has two apartments, the farthest of which has a curved wall conforming with the rounded end of the crevice floor, which, beyond this for some distance, is broken down.

Specimens of the mortar and of the dressed stone were procured from this house and brought east. Below the middle part of this line of houses, on an irregular projection, are the remains of a number of walls, in such a state of ruin, however, that the character of the original structure could not be made out. In digging among the debris of this ruin, I came upon a bin of charred corn, in which the forms of the ears were quite perfect. It seems to be of a variety similar to that cultivated by the tribes of the neighborhood at the present time.

That this corn had been placed there by the ancient occupants seems probable from the fact that it occupied a sort of basement apartment or cellar, and had been buried beneath the fallen walls of the superstructures. Imbedded in this mass of charcoal, I found the very perfect specimen of stone implement figured in plate XIV (figure 3). Many large fragments of the ordinary painted pottery were also picked up here. A certain new look about portions of this group leads one to suspect that it cannot boast of great antiquity ; but it is very difficult to calculate the effects of age upon walls so perfectly protected and in such a climate.

The group given in figure 2 is of a much more interesting and remarkable character. It was first observed from the trail, far below and nearly three fourths of a mile away. From this point, by the aid of a field glass, the sketch given in the plate was made. So cleverly are the houses hidden away in the dark recesses, and so very like the surrounding cliffs in color, that I had almost completed the sketch of the upper house before the lower or " sixteen-windowed ^v one was detected. They are at least eight hundred feet above the river. The lower five hundred feet is of rough cliff-broken slope, the remainder of massive bedded sandstone full of wind-worn niches, crevices, and caves. Within one hundred feet of the cliff-top, set deep in a great niche, with arched, overhanging roof, is the upper house, its front wall built along the very brink of a sheer precipice. Thirty feet below, in a similar but less remarkable niche, is the larger house, with its long line of apertures, which I afterward found to be openings intended rather for the insertion of beams than for windows.

PLATE VI.

I subsequently climbed the canon-walls to make a closer examination of these ruins, and the plans given in plate VI were obtained.

The lower house was easily accessible, and proved to be of a very interesting character. It occupies the entire floor of a niche which is about sixty feet long and fifteen in depth at the deepest part. The front walls are built flush with the precipice, and the partition-walls extend back to the irregular wall of rock behind. Portions of the wall at the left, viewing the house from the front, are greatly reduced; but the main wall, that part which contains the window-like openings, is still thirteen or fourteen feet high.

The arrangement of the apartments is quite complicated and curious, and will be more readily understood by a reference to the ground-plan (figure 1). The precipice-line, or front edge of the niche-floor, extends from a to b. From this the broken cliffs and slopes reach down to the trail and river, as shown in the accompanying profile (figure 3). The line bcd represents the deepest part of the recess, against which the walls are built. To the right of b, the shelf ceases, and the vertical face of rock is unbroken. At the left, beyond a, the edge is not so abrupt, and the cliffs below are so broken that one can ascend with ease. Above, the roof comes forward and curves upward, as seen in the profile.

The most striking feature of this structure is the round room, which occurs about the middle of the ruin and inside of a large rectangular apartment.

The occurrence of this circular chamber in this place is highly significant, and tends greatly to confirm my previously-stated opinion that the circle had a high significance with these people. Their superstitions seem to have been so exacting in this matter that, even when driven to the extremity of building and dwelling in the midst of these desolate cliffs, an inclosure of this form could not be dispensed with; a circular estufa had to be constructed at whatever cost of labor and convenience.

Its walls are not high and not entirely regular, and the inside is curiously fashioned with offsets and box-like projections. It is plastered smoothly, and bears considerable evidence of having been used, although I observed no traces of fire. The entrance to this chamber is rather extraordinary, and further attests the peculiar importance attached to it by the builders, and their evident desire to secure it from all possibility of intrusion. A walled and covered passage-way,/,/, of solid masonry, ten feet of which is still intact, leads from an outer chamber through the small intervening apartments into the circular one. It is possible that this originally extended to the outer wall, and was entered from the outside. If so, the person desiring to visit the estufa would have to enter an aperture about twenty-two inches high by thirty wide, and crawl, in the most abject manner possible, through a tube-like passage-way nearly twenty feet in length. My first impression was that this peculiarly-constructed doorway was a precaution against enemies, and that it was probably the only means of entrance to the interior of the house; but I am now inclined to think this hardly probable, and conclude that it was rather designed to render a sacred chamber as free as possible from profane intrusion. The apartments l, k, m, n, do not require any especial description, as they are quite plain and almost empty. The partition-walls have never been built up to the ceiling of the niche, and the inmates, in passing from one apartment to another, have climbed over. The row of apertures indicated in the main front wall are about five feet from the floor, and were doubtless entered for the

insertion of beams, although there is no evidence that a second floor has at any time existed. In that part of the ruin about the covered passage-way, the walls are complicated, and the plan can hardly be made out, while the curved wall inclosing the apartment e is totally overthrown.

In digging among the debris with our hammers, we came upon a large earthen vessel at ft, and shortly afterward discovered another near i. They were so situated in a small recess under the sheltering walls that the falling rubbish had not reached them. Roughly-hewn stone lids were fitted carefully over the tops, but both were empty. One had been slightly broken about the rim, while the other had been pierced on the underside by some sharp instrument, and had been mended by laying a small fragment of pottery over the aperture on the inside and cementing it down with clay. They are of the ordinary corrugated pottery, and have a capacity of about three gallons.

Beneath the vessels, spread out on the floor, was a large piece of rush-matting, and beneath this a quantity of fine vegetable tissue from the interior bark of some kind of tree. The vessels are illustrated in plate XIII, and the matting in plate XIV.

The rock-face between this ruin and the one above is smooth and vertical, but by passing along the ledge a few yards to the left a sloping face was found, up which a stairway of small niches had been cut; by means of these, an active person, unincumbered, could ascend with safety. On reaching the top, one finds himself in the very doorway of the upper house (a, figure 2) without standing-room outside of the wall, and one can imagine that an enemy would stand but little chance of reaching and entering such a fortress if defended, even by women and children alone. The position of this ruin is one of unparalleled security, both from enemies and from the elements. The almost vertical cliff descends abruptly from the front wall, and the immense arched roof of solid stone projects forward 15 or 20 feet beyond the house (see section, figure 3). At the right the ledge ceases, and at the left stops short against a massive vertical wall. The niche-stairway affords the only possible means of approach.

The house occupies the entire floor of the niche, which is about 120 feet long by 10 in depth at the deepest part. The front wall to the right and left of the doorway is quite low, portions having doubtless fallen off. The higher wall f g is about 30 feet long, and from 10 to 12 feet high, while a very low rude wall extends along the more inaccessible part of the ledge, and terminates at the extreme right in a small inclosure, as seen in the plan at c.

In the first apartment entered, there were evidences of fire, the walls and ceiling being blackened with smoke. In the second, a member of the party,* by digging in the rubbish, obtained a quantity of beans, and in the third a number of grains of corn, hence the names given. There are two small windows in the front wall, and doorways communicate between rooms separated by high partitions.

The walls of these houses are built in the usual manner, and average about a foot in thickness.

The upper house seems to be in a rather unfinished state, looking as if stone and mortar had run short. And when one considers that these materials must have been brought from far below by means of ropes, or carried in small quantities up the dangerous stairway, the only wonder is that it was ever brought to its present degree of finish.

* Mr. Brandegee.

Figure 3 is given for the purpose of making clear the geologic conditions that give shape to the cliffs as well as to show the relations of these houses to the cliffs. The hard and massive beds of rock resist the erosive agents; the soft and friable beds yield, hence the irregularity— the overhanging cliffs, the niches and benches. a is a section of the lower house, b of the upper.

It has heretofore been supposed that the occupants of these houses obtained water, either from the river below or from springs on the mesa above; but the immense labor of carrying water up these cliffs, as well as the impossibility of securing a supply in case of siege, made me suspect the existence of springs in the cliffs themselves. In three or four cases these springs have been found, and it is evident that with a climate a very little more moist than the present, a plentiful supply could be expected. Running water was found within a few yards of the group of houses just described, and Mr. Brandegee observed water dripping down the cliffs near a group of small houses on the opposite side of the canon.

About one mile farther up the canon, I came upon the ruin photographed by Mr. Jackson in 1871, and minutely described by him as the two-story cliff-house of the Rio Mancos. It is also in the cliffs of the north side, about seven hundred feet above the river, and although not so large or complicated in design as the houses just described, it shows higher skill in construction and is in a better state of preservation. It is also exceedingly difficult of access. It seems hardly necessary for me to enter into a detailed description, as little can be added to what has already been published ;* but for the purpose of having as much of the matter together as possible I present plate VII, illustrating some of the interesting features of this house.

Figure 1 gives the ground-plan, and shows the position of the house in relation to the floor of the niche. There are four small apartments only; the front one, a, being 10 feet long by 6 wide. Of the back rooms, one is 9 by 10 and the other 6 by 6 feet, while the apartment with the curved wall is much smaller. The walls are about twelve feet high and reach within from two to three feet of the overhanging roof. They are built in the ordinary manner of stone and adobe mortar, and what is rather remarkable are plastered both inside and out. This plaster does not differ greatly from the common mortar, is lightly spread over the walls, probably with the hands, and in color imitates very closely the hues of the surrounding cliffs, a pleasing variety of red and yellow grays. Whether this was intended to add to the beauty of the dwelling or to add to its security by increasing its resemblance to the surrounding cliffs, I shall not attempt to determine.

Another remarkable feature of this house is the consummate skill with which the foundations are laid upon and cemented to the sloping and overhanging faces of the ledge. The buttresses b, b, which have probably at one time supported a superstructure of wood or stone, now totally obliterated, are most striking illustrations of this; and just here is a fact that has an interesting bearing upon the question of the antiquity of this structure. These wall-supports, or buttresses, have originally been four in number, one evidently having falling off, and are built in continuation of the front wall, on a smooth, sloping surface of rock. Now the sandstone of which this rounded slope is composed is rather coarse and soft, and hence easily disintegrated. It is here also not greatly protected from the weather, since the cliffs above do not

* Bulletin No. 1, second series, p. 20.

overhang to any extent, and must, year by year, yield a little to the elements ; but I observe that since the construction of these foundations no perceptible change has taken place; the thickness of a sheet of paper has hardly been washed from the surface of the rock, and the mortar, which is of almost equal firmness with the rock, lies upon it as if placed there within a dozen years, and the plaster on the outer wall, although somewhat cracked and broken off, does not add greatly to our impressions of antiquity.

There is also a fact worthy of notice in regard to the question of occupancy. I have already stated my impressions that these houses were not used as constant dwelling places, but rather as places of occasional resort. I notice that, although the building seems complete, and has had its floors laid and its door-ways and windows conveniently and carefully arranged, the plastering of the interior is almost untouched, that with the exception of three names scratched in the soft, thick coat of adobe by Mr. Jackson's party, there is almost no trace of the presence of man ; yet this plaster may have been applied only shortly before the final desertion, and hence no definite conclusion can be drawn.

A sketch of one of the doorways is given in figure 2. The outline is accurately drawn, but there is a little too much regularity in the stonework. It will be seen that the aperture is of very nearly the same width above and below, which is rather unusual, since, in these ruins, as well as in those farther south, the door ways and windows are, as a rule, narrower at the top. This drawing also shows the manner of employing a number of small straight beams of wood as lintels, for the purpose, evidently, of strengthening the masonry above.

There are two of these exterior door-ways only, one op3ning into each story of the front room from the unoccupied part of the niche; these are shown in figure 3, a sketch of the interior of the front room taken from the side f. There is only a low wall between this room and the room c, while small door-ways communicate with the other apartments. There is a small rectangular window, 22 inches high by 30 wide, in the front wall, from which a fine view can be had of the deep narrow valley below.

Figure 4 is designed to show the extraordinary situation of these houses. Whether viewed from below or from the heights above, the effect is almost startling, and one cannot but feel that no ordinary circumstances could have driven a people to such places of resort.

There are no ruins of importance in the canon of the Mancos above the two-story house. Indistinct remains occur on the bottoms in a number of places, and a few small houses were observed in the cliffs. The most interesting of these is built upon a ledge about 40 feet above the trail, and is nearly midway between the two-story house and the head of the canon. It does not differ in any essential point from the ruins already described. I shall therefore pass it by, in order to take up two very interesting groups of ruins that occur about twenty miles to the northwest.

Between the Mesa Verde and the Late Mountains, of which Ute peak is the culminating summit, there is a long, deep valley or strip of low land that connects the great lowland of the Lower Mancos with the canon-cut plain that rises toward the Dolores. The southern end of this depressed strip drains into the Mancos, the northern into the McElmo. The latter stream heads along the north base of the Mesa Verde within five miles of the Mancos at the point where it enters the canon, and flows westward, passing along the north base of Ute Mountain, curving around to the southwest and reaching the San Juan nearly ten miles

beyond the Utah line. The large depressed area drained by this stream contains a great number of ruins, many of which have not yet been examined.

PLATE VIII.—THE TRIPLE-WALLED TOWER.

The group partially illustrated in this plate is situated on a low bench within a mile of the main McElmo, and near a dry wash that enters that stream from the south. It seems to have been a compact village or community-dwelling, consisting of two circular buildings and a great number of rectangular apartments. The circular structures or towers have been built, in the usual manner, of roughly-hewn stone, and rank among the very best specimens of this ancient architecture. The great tower is especially noticeable on account of the occurrence of a third wall, as seen in the drawing and in the plan at a. In dimensions it is almost identical with the great tower of the Rio Mancos. The walls are traceable nearly all the way around, and the space between the two outer ones, which is about five feet in width, contains fourteen apartments or cells. The walls about one of these cells are still standing to the height of 12 feet; but the interior cannot be examined on account of the rubbish which fills it to the top. No openings are noticeable in the circular walls, but door-ways seem to have been made to communicate between the apartments; one is preserved at d.

The inner wall has not been as high or strong as the others, and has served simply to inclose the estufa. This tower stands back about one hundred feet from the edge of the mesa and near the border of the village. The smaller tower, b, stands forward on a point that overlooks the shallow gulch; it is 15 feet in diameter; the walls are 3 1/2 feet thick and 5 feet high on the outside. Beneath this ruin, in a little side gulch, are the remains of a wall twelve feet high and twenty inches thick. The remainder of the village is in such a state of decay as to be hardly traceable among the artemisia and rubbish. The apartments number nearly a hundred, and seem, generally, to have been rectangular. They are not, however, of uniform size and certainly not arranged in regular order. The walls are marked by low lines of loose rubble which show no stone in place, and I am inclined to believe that they have never been raised to any great height. It is not impossible that they have been, originally, of a species of rubble-masonry such as is seen in some of the great casas farther south, and that these meager remains are all that is left of an imposing structure, but the total want of regularity both in the form and size of the apartments seems incon-sistent with such a conclusion. In reality they are more like a cluster of pens such as are used by the Moqui tribes for the keeping of sheep and goats. The site of this village can hardly have been chosen on account of its defensive advantages, nor on account of the fertility of the surrounding country. The neighboring plains and mesas are as naked and barren as possible. The nearest water is a mile away, and during the drier part of the season the nearest running water is in the Rio Dolores, nearly fifteen miles away. To suppose an agricultural people existing in such a locality, with the present climate, is manifestly absurd. Yet every isolated rock and bit of mesa within a circle of miles is strewn with remnants of human dwellings.

PLATE IX.—RUINS AT "AZTEC SPRINGS."

Another very important group of ruins is located in the depression between the Mesa Verde and the Late Mountains, and near the divide

between the McElmo and Lower Mancos drainage. It is stated by Captain Moss and others who have been in this locality that up to within two or three years there has been a living-spring at this place, and the spot has been christened by them Aztec Springs.

The site of the spring I found, but without the least appearance of water. The depression formerly occupied by it is near the center of a large mass of ruins, similar to the group last described, but having a rectangular instead of a circular building as the chief and central structure. This I have called the upper house in the plate, and a large walled inclosure a little lower on the slope, I have, for the sake of distinction, called the lower house.

These ruins form the most imposing pile of masonry yet found in Colorado. The whole group covers an area of about four hundred and eighty thousand square feet, and has an average depth of from three to four feet. This would give in the vicinity of one million five hundred thousand solid feet of stone-work. The stone used is chiefly of the fos-siliferous limestone that outcrops along the base of the Mesa Verde a mile or more away, and its transportation to this place has doubtless been a great work for a people so totally without facilities.

The upper house is rectangular, measures 80 by 100 feet, and is built with the cardinal points to within five degrees. The pile is from 12 to 15 feet in height, and its massiveness suggests an original height at least twice as great. The plan is somewhat difficult to make out on account of the very great quantity of debris.

The walls seem to have been double, with a space of 7 feet between ; a number of cross-walls at regular intervals indicate that this space has been divided into apartments, as seen in the plan.

The walls are 26 inches thick, and are built of roughly-dressed stones, which were probably laid in mortar, as in other cases.

The inclosed space, which is somewhat depressed, has two lines of debris, probably the remains of partition-walls, separating it into the three apartments, a, b, c. Inclosing this great house is a net-work of fallen walls, so completely reduced that none of the stones seem to remain in place; and I am at a loss to determine whether they mark the site of a cluster of irregular apartments, having low, loosely-built walls, or whether they are the remains of some imposing adobe structure built after the manner of the ruined pueblos of the Rio Chaco.

Two well-defined circular inclosures or estufas are situated in the midst of the southern wing of the ruin. The upper one, A, is on the opposite side of the spring from the great house, is 60 feet in diameter, and is surrounded by a low stone wall. West of the house is a small open court, which seems to have had a gate-way opening out to the west, through the surrounding walls.

The lower house is 200 feet in length by 180 in width, and its walls vary fifteen degrees from the cardinal points. The northern wall, a, is double, and contains a row of eight apartments about 7 feet in width by 24 in length? The walls of the other sides are low, and seem to have served simply to inclose the great court, near the center of which is a large walled depression, (estufa B.) No other ruins were observed in the neighborhood of these, although small groups are said to exist along the base of the Late Mountains, a few miles to the southwest.

PLATE X. —RUIN AT OJO CALIENTE, NEW MEXICO.

For the sake of comparison, I present in Plate X the ground-plan of a ruined pueblo found at Ojo Caliente, New Mexico. It occurs on a high,

almost isolated fragment of terrace near Caliente Creek. It has been constructed chiefly of adobe, and has consisted of rows of apartments surrounding a number of large open courts. Individual walls cannot be traced, and the rows of houses are reduced to smooth rounded ridges of earth. These are indicated on the plan, and are often as much as 8 feet high, and 30 feet wide at the base. The courts contain a number of small circles and mounds, a, a, and the single estufa is identical in appearance with those among the ruins of Colorado. A number of openings, b, b, through the walls indicate the location of gate-ways. Metates, arrow-heads, and many fragments of pottery were found. Many other groups of ruins similar to this occur in this as well as in the neighboring valleys. Near Abiquiu, a large pueblo occurs, at which I found a stone axe and a number of arrow-heads and metates. A couple of skeletons were also obtained here. This ruin is described at length by Dr. Yarrow in his report for 1874.*

PLATES XI AND XII.

Although it is quite impossible to read the curious rock-inscriptions of unknown tribes, or even to conjecture to any extent their meaning, yet it is conceded that in most cases they have a meaning and represent an idea or record an event. Aside from this, however, they are valuable to the historian as records of the grade of civilization reached by the tribes who executed them.

That the examples given in the two following plates belong to the age of the cliff-builders, cannot be satisfactorily proved, but, at the same time, evidence that they do, is not wanting. Some are found on the cliffs and in the niches with the cliff-dwellings, while all are in localities that must have been frequently visited by these people. Some are found in the canon of the Mancos, others on the bluffs of the San Juan, and many in the canons farther west.

Figures 1, 2, and 3, Plate XI, occur on the Mancos near the group of cliff-houses figured in Plate V. They are chipped into the rock, evidently by some very hard implement, and rudely represent the human figure. They are certainly not attempts to represent nature, but have the appearance rather of arbitrary forms designed to symbolize some imaginary being.

Figures 4,5, and 6 were found in the same locality, not engraved, but painted in red and white clay upon the smooth rocks. These were certainly done by the cliff-builders, and probably while the houses were in process of construction, since the material used is identical with the plaster of the houses. The sketches and notes were made by Mr. Brandegee. The reproduction is approximately one-twelfth the size of the original.

The examples given in figures 7, 8, 9, 10, and 11, as well as those in Plate XII, occur on the Rio San Juan, about ten miles below the mouth of the Rio La Plata. A low line of bluffs, composed of light colored massive sandstones that break down in great smooth-faced blocks, rises from the river-level and sweeps around toward the north. Each of these great blocks has offered a very tempting tablet to the graver of the primitive artist, and very many of them contain curious and interesting inscriptions. Drawings were made of such of these as the limited time at my disposal would permit. They are all engraved or cut into the face of the rock, and the whole body of each figure has generally been chipped out, frequently to the depth of one-fourth or one-half an inch.

The work on some of the larger groups has been one of immense

* Report of the Chief of Engineers for 1875, page 1064.

labor, and must owe its completion to strong and enduring motives. With a very few exceptions the engraving bears undoubted evidence of age. Such new figures as occur are quite easily distinguished, both by the freshness of the chipped surfaces and by the designs themselves. Figure 11 gives a specimen of the modern work; it is evidently intended to represent a horse, and is done in the manner of the Navajoes. It will readily be seen that among all the figures given of the ancient work there is no animal that resembles a horse, and we can hardly suppose that artists who could so cleverly delineate birds and deer and men, would fail in an attempt to represent an animal of so marked a character. The curious designs given in figure 10 have a very perceptible resemblance to many of the figures used in the embellishment of pottery.

The most striking group observed is given in figure 1, Plate XII. It consists of a great procession of men, birds, beasts, and fanciful figures. The whole picture as placed upon the rock is highly spirited, and the idea of a general movement toward the right, skillfully portrayed. A pair of winged figures hover above the train as if to watch or direct its movements; behind these are a number of odd figures, followed by an antlered animal resembling a deer, which seems to be drawing a notched sledge containing two figures of men. The figures forming the main body of the procession appear to be tied together in a continuous line, and in form resemble one living creature about as little as another. Many of the smaller figures above and below are certainly intended to represent dogs, while a number of men are stationed about, here and there, as if to keep the procession in order.

As to the importance of the event recorded in this picture, no conclusions can be drawn; it may represent the migration of a tribe or family or the trophies of a victory. A number of figures are wanting in the drawing at the left, while some of those at the right may not belong properly to the main group. The reduction is, approximately, to one-twelfth.

Figures 2 and 3 of the same plate represent only the more distinct portions of two other groups. The complication of figures is so great that a number of hours would have been necessary for their delineation, and an attempt to analyze them here would be fruitless.

PLATE XIII.—POTTERY.

It is generally conceded that the ancient tribes of the San Juan produced fictile fabrics superior to those of the town-building tribes of to-day. There is, however, great similarity between the ancient and modern work, both in material and execution, and the differences are not greater than could be expected in the manufactures of the same tribe at periods separated by two or three centuries of degeneration.

The study of the fragmentary ware found about the ruins is very interesting, and its immense quantity is a constant matter of wonder. On one occasion, while encamped near the foot of the Mancos Canon, I undertook to collect all fragments of vessels of manifestly different designs within a certain space, and by selecting pieces having peculiarly marked rims, I was able to say with certainty that within ten feet square there were fragments of fifty-five different vessels. In shape these vessels have been so varied that few forms known to civilized art could not be found. Fragments of bowls, cups, jugs, pitchers, urns, and vases, in infinite variety, maybe obtained in nearly every heap of debris.

The art of ornamentation seems to have been especially cultivated, as

very few specimens are found that are not painted, indented, or covered with raised figures. Indeed, these ornamental designs are often so admirable, and apparently so far in advance of the art-ideas of these people in other respects, that one is led to suspect that they may be of foreign origin. This suspicion is in a measure confirmed when we discover the scroll and the fret struggling for existence among the rude scrawlings of an artisan who seems to have made them recognizable rather by accident than otherwise. It is not improbable, however, that the specimens referred to are but rude copies of models designed by more accomplished artists or procured from some distant tribe. There is certainly no conclusive evidence that these people ever came in contact with Europeans.

The material used in the manufacture of pottery is generally a fine clay, (in which the country abounds,) tempered with sand or pulverized shells. The modeling is done almost exclusively with the hand; no wheel has been used, and no implement whatever, except for surface creasings or indentings.

The thickness of the ware varies from 1/8 to 1/2 an inch. Lightness has evidently been greatly desired, and vessels having a capacity of many gallons are not more than 1/4 of an inch thick in any part.

Nearly all of the vessels and fragments collected have been baked or burned, but not to such a degree as to greatly change the color of the clay. Most, if not all, of the painted pottery is glazed with a very thin vitreous coating that gives a beautiful enamel-like surface of great hardness; upon this the coloring-matter is laid, apparently with a brush. With one or two exceptions the corrugated pottery is without the glazing, and in no instance contains painted figures. The peculiarities of this variety can be described more readily by reference to the examples in the plate.

Figure 1 represents a large vessel obtained in one of the Mancos cliff-houses (Plate VI). It is of the corrugated variety; has a capacity of about three gallons, and was probably used for carrying or keeping on hand a supply of water. It is quite light, not weighing more than a common wooden pail, and is made of a light-gray clay tempered with coarse sand, and but slightly burned. The corrugated appearance is given by laying on strips of clay, in somewhat regular succession, and pressing them into place and indenting them with the thumb or a stick. Whether a thin shell of clay is first constructed and the strips laid on and pressed down so as to unite with it, or whether the vessel is built up by the strips alone, cannot be determined, since the inside is perfectly smooth, excepting finger-marks, and the strips are so welded into the general texture of the vessel that individual strips cannot be detected beneath the surface when examined on broken edges.

In the specimen figured the workman has begun near the center of the rounded bottom and laid a strip in a continuous but irregular spiral (see Fig. 3) until the rim was reached, indenting the whole surface irregularly with the finger. Two small conical bits of clay have been set in near the rim, as if for ornament. Other specimens have small spirals, while others have scrolls, and still others very graceful festoons of clay (Figs. 2 and 2a). A number of the more distinct styles of indentation are given in connection with this figure (Figs. 3, 3a, 3b, 3c, and 3d).

Figure 4 is a bowl restored from a large fragment. It is painted both inside and out, and the designs are applied with rather more than usual care.

Figures 5, 5a, and 5b are prominent among the ornamental designs. I have corrected the drawing, but have introduced no new element.

Figure 6 represents a very usual pattern of mug or cup. It is of the ordinary painted ware, and is made to contain about a pint. The specimen is not entire.

Figure 7 is apparently a pipe. It was found by Mr. Aldrich, near a ruin on the San Juan, and is made of the ordinary potter's clay; it is 2 inches in length.

Figure 8 represents part of an ornamental handle, formed by twisting together three small rolls of clay.

Figure 9 represents a small spoon or ladle. Fragments of similar implements are quite numerous.

Figure 10 is a portion of the handle of some small vessel.

As to whether the manufacture of pottery was carried on in certain favorable localities only, or whether each village had its own skilled workmen or workwomen, I cannot determine, since, as previously stated, no remains of kilns or manufactories were discovered. The forms and styles of ornament are pretty uniform, which is to be expected in either case, since the inhabitants of the various villages must have had constant communication with each other.

PLATE XIV.

This plate contains drawings of a number of stone implements; arrowheads, ornaments, and other articles manufactured or used by the ancient inhabitants of this region. Nearly all were found so associated with the architectural remains that I do not hesitate to assign them to the same period.

Figure 1 represents a small fragment of rush matting, a large piece of which was found on the floor of one of the cliff-houses of the Rio Mancos. (See Plate VI.) It is probably manufactured from a species of rush, Scirpus validus, that grows somewhat plentifully along the Mancos bottoms.

Figure 2 represents a bundle of small sticks, probably used in playing some game. They are nearly a foot in length, and have been sharpened at one end by scraping or grinding. They were found in one of the cliff-houses of the Mancos, buried beneath a pile of rubbish. The bit of cord with which they are tied is made of a flax-like fiber, carefully twisted and wrapped with coarse strips of yucca bark; beside this a number of short pieces of rope of different sizes were found, that in beauty and strength would do credit to any people. The fiber is a little coarser and lighter than flax, and was probably obtained from a species of yucca, which grows everywhere in the southwest.

Figure 3 is a very perfect specimen of stone implement found buried in a bin of charred corn in one of the Mancos cliff-houses (figure 1, Plate V).

It is 8 inches in length and 2 1/2 inches broad at the broadest part; its greatest thickness is only 1/2 an inch. One face is slightly convex, while the other is nearly flat. The sides are neatly and uniformly rounded, and the edge is quite sharp. It is made of a very hard, fine-grained, siliceous slate, is gray in color, and has been ground into shape and polished in a most masterly manner.

Although its use is not positively determined, it belongs, in all prob-ability, to a class of implements called " scrapers," which are employed by most savage tribes in the dressing of skins. This specimen may have been used for other purposes, but certainly not for cutting or striking, as the material is very brittle. The most conclusive proof of its use is the appearance of the edge, which shows just such markings as would be produced by rubbing or scraping a tough, sinewy surface.

Figure 4 represents a part of a metate or mill-stone. The complete im-

plement consists of two parts—a large block of stone with a concave surface, upon which the maize is placed, and a carefully-dressed but ooarse-grained slab of stone for grinding. This slab is generally from 8 to 12 inches long by 3 to 6 wide, and from 1 to 2 inches thick. The specimen illustrated is made of black cellular basalt, and was found with many others at the ruined pueblo near Ojo Caliente, N. Mex. (Plate X.)

Figure 5 is a very much worn specimen of stone axe, which was found at an ancient ruin near Abiquiu, N. Mex. It is made of light-colored chloritic schist, and measures 2 inches in width by 3 in length.

Figures 6 and 6a are specimens of ear-ornaments, such as are found in connection with very many of the ruins of Southern Colorado. These are made of fine-grained gray slate, only moderately well polished, and measure an inch and a quarter in length.

Figure 7 represents a marine shell of the genus Olivella, obtained probably from the Pacific coast. Large numbers of this and allied shells are found about these ruins. They are generally pierced, and were doubtless used as beads.

Figure 8 represents a small carved figure found on the Rio Mancos. It is made of hard gray slate. Its use or meaning cannot be determined.

Figure 9 represents a stone ring 5/8 of an inch in diameter, and probably intended for the finger. It is made of hard gray slate; is shaped like the usual plain gold ring, and is quite symmetrical.

Figure 10. Arrow-heads were found associated with nearly every ruin examined. They present a great variety of forms ; some of the more striking of these are given in the figure. The materials used in their manufacture are principally the more beautiful varieties of obsidian, jasper, and agate.

A NOTICE OF THE ANCIENT RUINS IN ARIZONA AND UTAH LYING ABOUT THE RIO SAN JUAN.

By W. H. Jackson.

In continuing the investigation, commenced last season, of the very interesting ruins scattered throughout the San Juan basin, I proceeded to Parrott City, a frontier mining-camp on La Plata River, where I procured the services of Harry Lee as guide and interpreter. Mr. E. A. Barber, naturalist and special correspondent of the New York Herald, was also of the party. Providing ourselves with the supplies which had been forwarded to this point via Tierra Amarilla, we started out late in July, journeying westwardly to that point on the Hovenweep from which we had turned back last year, and where we shall also resume our explorations.

The Hovenweep (a deserted valley) is a tributary of the McElmo, which, together with the wide-spreading arms of the Montezuma, drains into the San Juan all that portion of the country lying between the Mesa Verde and the Sierra Abajo, covering in the aggregate some two thousand five hundred square miles. Their labyrinthine canons head close upon the Dolores on the north, and ramify the plateaus in every direction with an interminable series of deep, desolate gorges, and wide, barren valleys. There is not a living stream throughout this whole region. During the summer months water occurs in but very few places, generally in pockets, sometimes in springs, where the excess, if any, is soon swallowed by the hot and thirsty sands. The rainy season is in winter and the early spring months, when the water is more generously distributed, being then found in the many basins scattered over the bare tops of the mesas, as well as in the beds of the canons, the lower temperature of the colder season preventing the rapid evaporation of summer and autumn weather. As a great proportion of the surface of this region is a bare bed of rock, with a soil in the lowlands nearly impervious to moisture, the winter showers soon gather their waters together in great floods in the main channels, and then, rushing down in a solid body, form those deep " washes" so characteristic of the country. But these torrents are short-lived, and it is only by noting the height of the drift material lodged upon the trunks of the venerable cotton woods bordering the banks that we can fully realize such great bodies of water ever having existed in so dusty a bed. Every canon and valley has its corresponding wash, worn perpendicularly down through the dry, easily-eroded soil, forming circuitous but excellent pathways. In some valleys, where the drainage is considerable, these washes frequently attain a depth of from 30 to 40 feet, and are impassable for miles.

The intervening table lands obtain a very nearly uniform height of 500 feet, running up to over 1,000 feet as we approach the Dolores divide. In the wider valleys the maximum is reached by successive steps, or benches, rising one back of the other, while in the narrow

canons the ascent is more abrupt; the upper third of the escarpment being generally perpendicular, with the lower two-thirds composed of talus. Their summits and sides are usually clothed with a scrubby growth of pinon and juniper trees, increasing in density and size as we approach the divide on the north, while the valleys below sustain dense masses of sage-brush and greasewood, that, in some places, attain a height of from 10 to 12 feet. Vigorous, fresh-looking cottonwoods line the main channels, and are as deceptive to the thirsty traveler as a mirage. One may travel for miles in the parched bed of the wash at their feet, while overhead their wide-spreading branches cast a grateful shade, and yet not be able to find a drop of water anywhere in their vicinity.

West of the Montezuma two or three small tributaries of the San Juan head from the southern face of the Sierra Abajo, and then comes Epsom Greek, rising among the plateaus farther to the west—so called from the water in one portion of its bed having the effect and tasting like that salt. For a distance of some twenty-five miles above its mouth the valley of this creek presents upon its eastern side a remarkable wall, some 400 feet in height, inaccessible throughout its whole length with the exception of one place where the Indians have made a way for themselves. It is caused by an immense fold in the sandstones, running north and south in a semicircular line, for some forty miles, and this valley has been eroded from that portion of it where the strata stood nearly perpendicular. On the west the beds sweep up in graceful curves to a nearly horizontal position, upon which isolated mesas rise up above the general level in bold relief against the sky.

The Rio San Juan drains all of this great interior basin, covering over twenty thousand square miles, as well as several great mountain masses bordering it nearly all around. It has at the mouth of the McElmo an average width of fifty yards, and a depth of from 4 to 6 feet; its current moving somewhat sluggishly in great sweeping curves that almost touch upon themselves again. The water is warm, and well freighted with the soil which it is continually undermining—a great contrast to the clear, ice-cold tributaries which give it existence. The bottoms are from three to five miles in width, and, bordering the stream, covered with dense growths of cottonwoods and willows. The broad and fertile alluvial lands, well covered with grass, and the low sagebrush benches bordering them, will undoubtedly prove a rich agricultural possession at no distant day. Back of all, upon either hand, rise up the precipitous sandstone bluffs, picturesque in outline and color, that gradually close down upon the river until it is ingulfed in the great canon commencing just below the mouth of the Rio De Chelly, and is then lost to all knowledge until it re-appears mingling its waters with those of the still more turbid and turbulent Colorado. South of the San Juan, the Rio De Chelly, coming in opposite the mouth of Epsom Creek, does not differ in its canyon character from those of the north. The bordering plateau, however, is more massive and less cut up by side canons. The same aridity prevails throughout nearly its whole length.

Having thus superficially surveyed the region on which are to be found a vast number of prehistoric ruins, we will now return to the Hovenweep and examine, in such detail as our rapid reconnaissance will allow, the more prominent of the abundant remains.

Starting from the pueblo of the Hovenweep, described on page 30 of Bulletin No. 1, second series, we do not find in the immediate neighborhood any other ruins of importance; but a short distance down the

canon they begin to occur quite frequently. We observe first, on the left, the remains of a tower perched upon a rock, jutting out into the valley, beneath and about which were other ruins, evidently belonging to the tower. In the vicinity are other "rock-shelters," occurring upon either side of the canon, some merely walled-up caves, while others are semicircular walls built out from the rock and protected overhead by an overhanging ledge.

Some seven miles from the pueblo, and about three above the McElmo, on the western side of the valley, is a jagged, butte-like promontory, of a brownish-yellow sand-rock, standing out from the mesa, upon the face of which are a number of benches and cave-like recesses. These have been built up and inclosed with neatly-laid walls, making six different houses or sets of rooms upon three benches, one above the other. Access was had from below, first by ascending a steep slope of debris for about 100 feet to the foot of the rock, where we find the first and largest of the houses. It is some 12 feet in length by 5 feet deep, divided midway into two rooms, but rendered somewhat indistinct by the falling-down of a portion of the rock back of it. The second bench was reached in the manner shown in Fig. 2, Plate 15, the little house there seen being the first of three strung along in a row. Above these were two other similar ones, very difficult to reach, the ledge upon which they stand projecting over the one beneath. The perfectly flat floor of the valley at the foot of the rock contained faint indications of having been occupied by buildings; and one of the curves of the wash, here some 10 feet in depth, in cutting away the soil disclosed a thin stratum of charcoal about 6 feet below the surface ; one piece that we picked out being about 3 inches thick, and the earth about the mass in which it occurred was much burnt, as though the fire had been long continued. About a mile farther down we came to an expansion of the valley with a canon opening in from the west. An examination up this for six miles failed in discovering any remains of stone buildings, but very numerous indications of probably adobe structures, or earthen foundations for wooden ones ; in every instance circular, with a diameter of from 15 to 25 feet. A dozen such were found within three miles of each other. Fragments of pottery of excellent quality and neatly ornamented were very abundant.

Opposite the mouth of this canon the mesa juts out prominently into the valley. Half-way up its face is a bench-like spur, upon which rests an almost perfectly rectangular block of sandstone fallen from the cliff above. It is 38 by 32 feet square and 20 feet high. The upper surface is entirely covered with the remains of a wall from 3 to 5 feet high, running around its outer edge; a diagonal line divides the interior into two nearly equal spaces, one of which is again subdivided into three smaller rooms, the passage between them formed by the dividing walls overlapping, their opposite ends being set off from each other about 20 inches, thus necessitating a zigzag course in passing from one to the other. At the foot of the south side of the rock, and directly beneath the subdivided half of it, there is a line of stone wall inclosing a space 40 feet square, the rocks forming one side, with the center depressed a couple of feet below the surrounding level. In the right-hand corner of this inclosure, against the rock, are the ruins of another building 20 feet square; 10 feet above the base of the rock, and over this ruin, four holes have been drilled into it, six inches deep and four inches in diameter, serving evidently to support the roof of the building below and to afford a means of access to the rock above, a door-way in the surrounding wall being plainly indicated at that point.

Two miles farther down the McElmo comes in at nearly right angles from the east, and upon the point of the mesa included within the angle thus formed by the two canons or valleys—we cannot call them streams— are a group of ruins similar to ones above, but much less regularly built. An interesting inscription occurs upon the under face of a large rock that supports a ruin, covering some GO square feet of surface; animals resembling goats, lizards, human figures, and many hieroglyphical signs abound. While sketching these our attention was called to the top of the mesa by Mr. Holmes (who has accompanied us thus far with his division on his way to the San Juan, and who had ascended to the summit for the purpose of sketching), to some very interesting remains he had discovered there.

The perpendicular scarp of the mesa ran around very regularly, 50 to 100 feet in height, the talus sloping down at a steep angle. On cave-like benches at the foot of the scarp, is a row of rock shelters, much ruined, in one of which was found a very perfect polished stone implement. Gaining the top with some difficulty, we found a perfectly flat surface, 100 yards in width by about 200 in length, separated from the main plateau by a narrow neck, across which a wall had been thrown, but now nearly leveled. Very nearly the entire space fenced in by this wall was covered by an extended series of small squares, formed by thin slabs of sand-rock set up edgewise. (Fig. 3, Plate 20.) All were uniformly about 3 by 5 feet square, arranged in rows, two and three-deep, and adjusted to various points of the compass, but there were also a few circles disposed irregularly about the inclosed area, each about 20 feet in diameter, and formed of these same squares, leaving a circular space of 10 feet diameter in the center. These squares occur indiscriminately over the whole region that has come under our observation, upon the mesa tops and in the valleys, all of the same general shape and size, very seldom accompanied by even the faintest indication of a mound-like character, but nearly always in groups. We have always supposed them to be graves, but have not as yet found any evidence that would prove them such. Some that we excavated to a depth of 5 and 6 feet, into a solid earth that had never been disturbed, rewarded us with not the faintest vestige of any remains, excepting, in nearly every case, a thin scattered layer of bits of charcoal from 6 to 18 inches beneath the surface. In one instance, near the Mesa Verde, the upright slabs of rock which inclosed such a square were sunk 2 feet into the earth and projected 6 inches above it. In another, was found a mass of charred matter that promised to throw some light upon the subject, but a chemical analysis by Dr. Endlich proved it to be simply charred juniper-wood, without perceptible admixture of animal matter. In the present instance, as the soil was thin and sandy, in some places blown entirely off, leaving the bare bed-rock exposed, we excavated several of these with pick and shovel, there being only from 12 to 18 inches of earth to remove, but in no case finding anything more than the scattered charcoal spoken of above. In some, the earth was calcined, as though a fire had been made within them, while in others there was no vestige of a fire beyond the presence of the charcoal. The question very naturally arises as to whether they might not have been cremationists, a supposition that would have some appearance of likelihood, could we but find any trace of human remains among the bits of charred wood. Scattered over the whole surface of this mesa were a great many flint-chip-pings, from among which we picked up a number of very beautiful arrow-points. As the summit commands a wide sweep of country, it is not unlikely that sentries of old beguiled their tedious watch with arrow-making.

From the camp at the pueblo Mr. Chittenden, of Mr. Holmes's division, rode up the Hovenweep some eight miles, to where it divided equally into two branches, and upon the point between these forks he found the remains of a round tower, commanding an extended view down the main canon. No other ruins were noticed.

The parties, under the guidance of Mr. Gardner, camped one night near the head of the Hovenweep, and found there an important group of ruins, described as follows by Mr. Adams:

"The first of these we met are situated at the upper edge of the side of the canon, about one-third of the distance from the top, on a ledge about 300 feet long and 50 wide. On this small space were crowded some 40 houses, as well as we could judge from the ruins. The general plan of structure was circular, varying in size, but generally from 10 to 15 feet in diameter. The stone was dressed to three times the size of an ordinary brick and in the same shape. ^# * * The whole arrangement of the little town was for defense ; perched up high above, on the summits of bowlders, were little watch towers, which commanded the plateau above."

Between the Montezuma and the Hovenweep is a high plateau running north and south, from the San Juan to the Dolores ; the southern portion a level sage-covered plain, while the northern is more undulating and covered with junipers and pinon pine. Upon this we found the remains of many circular towns, generally occupying slight eminences, and in but one or two cases, as far as we observed, were so entirely demolished that not one stone remained upon another. In one of these exceptions, about half the circumference only of a tower remained, 15 feet in height and of average masonry. Broken pottery was but sparingly scattered about, showing them not to have been occupied as much as the very similar remains in the valleys below. This mesa, averaging 500 feet in height above the surrounding country, does not contain a spring or drop of water, except such as may remain in the holes in the rocks after a shower. The soil is thin and sandy, blown off clean to the rock-bed in places, yet what there is is well grassed, and sage-brush flourishes luxuriantly. As cultivation was out of the question, and permanent residence improbable, it is very likely these towns were lookouts or places of refuge for the shepherds, who brought their sheep or goats up here to graze, just as the Navajos used to, and the Utes do at the present time. Rude huts of a later day are now found scattered over its surface, by the side of the washes where water would be likely to collect.

In traveling down the San Juan, from the mouth of the McElmo, there are not within the first ten or twelve miles any ruins that would claim attention upon a rapid reconnaissance. Indistinguishable mounds of earth frequently occur along the bottom lands, surrounded by the ever-present fragments of pottery, showing them to be the sites or the remains of habitations; the quantity of pottery, domestic utensils, and arrow-points helping somewhat to determine the length of time they were occupied.

Crossing the mouth of the broad sandy wash of the Montezuma, that is here bordered with groves of brilliantly-green cottonwoods along its arid course, we pass about three miles below, and find camp under a grove of patriarchal trees within a well-grassed bend of the river. A wide gravelly bench, some 50 feet in height, and running back to the bluff line, rises up abruptly from the bottom lands; a few rods below camp, the river in its meandering sweeps close to the foot of this bench, producing an almost perpendicular face. Upon the top of the bench at

this point, overlooking the river, are the ruins of a quadrangular structure of peculiar design.

Referring to the ground-plan, as shown in Fig. 2, Plate 16, we see that it is arranged very nearly at right angles to the river, its greatest depth on the left, where it runs back 120 feet; the front sweeps back in a diagonal line, so that the right-hand side is only 32 feet in depth. The back wall is 158 feet long, and at right angles to the two sides. In the center of the building, looking out upon the river, is an open space 75 feet wide, and averaging 40 feet in depth, its depressed center divided nearly equally by a ridge running through it at right angles to the river. We judged it to have been an open court, because there was not the least vestige of a wall in front, or on the ridge through the center, while upon the other three sides they were perfectly distinct; although it is difficult to explain why it should have been hollowed out in the manner shown in the plan. Back of this court is a series of seven apartments of equal size, springing in a perfect arch from the heavy wall facing the court, leaving a semicircular space in the center, 45 feet across its greatest diameter. Each one is 15 feet in length, and the same in width across its center, the walls somewhat irregular in thickness, but averaging 20 inches, compact, and well laid. On the left are three rooms extending across the whole width of the building, each averaging 45 by 40 feet square; on the right only one was discernible. Back of the circle, our impression was that the walls diverged in the manner shown in the plan, although there is so much confusion resulting from the heaping up of the debris that much must be left to conjecture. There is also a slight shadow of doubt in regard to the wall facing the river on the right; it is barely possible that it extended somewhat farther out, although there is here a steep inclination to the brink of the bluff, and that it has become entirely obliterated by its foundations giving way. The remains of the wall above, however, led us to believe that it had been originally built in the way it is shown in the plan. Extreme massiveness is indicated throughout the whole structure by the amount of debris about the line of the walls, forming long rounded mounds, 4 to 5 feet high, with the stone-work cropping out, 20 to 24 inches in thickness. Portions of the outer wall have fallen outward almost in one solid piece, the stones remaining spread out in much the same order they occupied in the standing wall. The stones were of fair size, but yet not so large but that one man could handle the largest of them. They were obtained from the neighboring bluff, and probably undressed, but broken into very nearly rectangular blocks, so that when carefully laid and dressed up with adobe cement they would have all the effect of dressed stone. Their extreme age, which has crumbled a great many into dust and rounded the asperities of all into shapeless bowlders, renders any conjecture upon this point somewhat uncertain. Where portions of the undisturbed wall did appear above the rubbish it showed a solid, well-constructed masonry. No indications whatever could be found of any passage-ways, nor could we expect to find any so near their base, for all of the apartments were probably entered by ladders, the same as in other buildings of this order that we have found in other localities.

Upon either side and back of this building were low, indefinite lines of earth, not more than 12 to 18 inches above the surrounding surface, inclosing areas from 40 to 60 feet in diameter, which were probably corrals for domesticated animals, the walls being composed of adobe or turf brought from the valley below, and which would, of course, wash down to a barely perceptible ridge.

In the face of the bluff immediately under this ruin, and upon a recessed bench situated about half-way between top and bottom, is a row of little "rock shelters." A stratum of a rotten shaly sandstone has been weathered or dug out, probably both, for a distance of 300 feet along the bluff, to a depth of about six feet, leaving a firm floor and a projecting ledge overhead, with just room enough to walk along without stooping. A continuous row of buildings occupied this bench, although most of them have tumbled into the river, and none have their front walls remaining. Door-ways through each of the dividing walls afforded access along the whole line. A few rods up stream, and in the same line of the bluff as the preceding, was another little niched cave-house (figure 3, Plate 18), 14 feet in length, 5 feet high at the center and 6 deep, divided into two equal apartments; a small square window, just large enough for one to crawl through, was placed midway in the wall of each half. We well might ask whether these little " cubby-holes" had ever been used as residences, or whether, as seems at first most likely, they might not simply be " caches," or merely temporary places of refuge; and while, no doubt, many of them are such, yet in the greater majority the evidences of use and the presence of long continued fires, indicated by their smoke-blackened interiors, would prove them to have been quite constantly occupied. Among all dwellers in mud-plastered houses it is the practice to freshen up their habitations by repeated applications of clay, moistened to the proper consistency, and spread with the hands, the thickness of the coating depending upon its consistency. Every such application makes a building appear perfectly new, and many of the best-sheltered cave-houses have just this appearance, as though they were but just vacated.

A quarter of a mile back over the flat level bench, is a long narrow hill about 100 feet in height, commanding an extended view up and down the valley, upon the summit of which is one of the circular, mound-like inclosures which occur so frequently upon both the highlands and the lowlands. It evidently has some connection with the group below on the river's edge, for there are no other ruins within several miles.

Continuing down the river, under the great bluffs which border it closely, we find many ruins of the "rock-shelter " kind occurring frequently in all sorts of positions, from the level of the valley to a height of over 100 feet, and from the smallest kind of a " cache," not larger than a bushel-basket, to buildings that sheltered several families. We illustrate one group in figure 4, of Plate 18, that consists of a row of three small houses built upon a ledge running horizontally along the perpendicular face of the bluff, about 60 feet above the trail immediately below it. The ledge was so narrow that the buildings occupied every available inch of its surface. As near as we could judge from below, each was about 5 feet wide and 10 long, with apertures through their end walls, and in the two first ones windows in the outer wall. No possible means of access were discernible, and if ladders were ever used they were taller than any of the trees available for the purpose that grow in this vicinity.

About twelve miles below the Montezuma we discovered, far away upon the opposite side of the river, a great circular cave, occupying very nearly the entire height of the bluff in which it occurred, and in which, by close inspection with the glass, we were enabled to make out a long line of masonry. Fording the river and approaching it, we found that the bluff-line at this place was a little over 200 feet in height, the upper half a light-colored, firm, massive sandstone, and the lower a dark red and shaly variety. The opening of the cave is almost perfectly cir-

cular, 200 feet in diameter, divided equally between the two kinds of locks, reaching, within a few feet, the top of the bluff above and the level of the valley below. It runs back in a semicircular sweep to a depth of 100 feet ; the top is a perfect half dome, and the lower half only less so from the accumulation of debris and the thick brushy foliage, the cool dampness of its shadowed interior, where the sun never touches, favoring a luxuriant growth. A stratum of harder rock across the central line of the cave has left a bench running around its entire half circle, upon which is built the row of buildings which caught our attention half a mile away. In figure 3, Plate 16, we have a plan of a horizontal section of the cave, showing the ledge and the manner of the disposition of the buildings upon it; in the drawing at the top of Plate 17, we have a view of them as seen from the opposite side of the cave.

It will be seen that the houses occupy the left hand or eastern half of the cave, for the reason, probably, that the ledge was wider on that side, and the wall back of it receded in such a manner as to give considerable additional room for the second floor, or for the upper part of the one-story rooms. It is about 50 feet from the outer edge in to the first building, a small structure 16 feet long, 3 feet wide at the outer end, and 4 at the opposite end; the walls, standing only 4 feet on the highest remaining corner, were nearly all tumbled in. Then came an open space 11 feet wide and 9 deep, that served probably as a sort of workshop. Four holes were drilled into the smooth rock floor, about 6 feet equidistantly apart, each from 6 to 10 inches deep and 5 in diameter, as perfectly round as though drilled by machinery. We can reasonably assume that these people were familiar with the art of weaving, and that it was here they worked at the loom, the drilled holes supporting its posts. At b, in this open space, are a number of grooves worn into the rock in various places, caused by the artificers of the little town in shaping and polishing their stone implements. The main building comes next, occupying the widest portion of the ledge, which gives an average width of 10 feet inside; it is 48 feet long outside, and 12 high, divided inside into 3 rooms, the first two 13 1/2 feet each in length, and the third 16 feet, divided into two stories, the lower and upper 5 feet in height. The joist holes did not penetrate through the walls, being inserted about six inches, half the thickness. The beams rested upon the sloping back-wall, which receded far enough to make the upper rooms about square. Window-like apertures afforded communication between each room, all through the second story, excepting that which opened out to the back of the cave. There was also one window in each lower room, about 12 inches square, looking out toward the open country, and in the upper rooms several small apertures not more than 3 inches wide were pierced through the wall, hardly more than peep-holes. The walls of the large building continued back in an unbroken line 130 feet farther, with an average height of 8 feet, and divided into 11 apartments, with communicating apertures through all. The first room was 9 1/2 feet wide, the others dwindling down gradually to only 4 feet in width at the other extremity. The rooms were of unequal length, the following being their inside measurements, commencing from the outer end, viz : 12 1/2, 9 1/2, 8, 7 1/2, 9, 10, 8, 7, 7, 8, 31 feet; the ledge then runs along, gradually narrowing, 50 feet farther, where another wall occurs across it, after which it soon merges into the smooth wall of the cave. The first of these rooms had an aperture leading outward large enough to crawl through; the wall around it had been broken away so that its exact size could not be determined; all the others, of which there were about two to each room, were mere peep-holes, about 3 inches in diameter, and

generally pierced through the wall at a downward angle. No sign of either roofing or flooring material could be found in any of the rooms. Everything of that kind has been thoroughly burnt out or removed, so that not a vestige of wood-work remains. We cannot be positively certain that they had ever been roofed, the mild temperature of this region hardly necessitating any other covering than such as the ample dome of the cave itself offered. In the central room of the main building we found a circular basin-like depression (a), 30 inches across and 10 deep, that had served as a fire-place, being still filled with the ashes and cinders of aboriginal fires, the surrounding walls being blackened with smoke and soot. This room was undoubtedly the kitchen of the house. Some of the smaller rooms appear to have been used for the same purpose, the fires having been made in the corner against the back wall, the smoke escaping overhead. The masonry displayed in the construction of the walls is very creditable; a symmetrical curve is preserved throughout the whole line, and every portion perfectly plumb; the subdivisions are at right angles to the front. The stones employed are of the size used in all similar structures, and are roughly broken to a uniform size; more attention seems to have been paid to securing a smooth appearance upon the exterior than the interior surfaces, the clay cement being spread to a perfectly plane surface, something like a modern stucco finish. In many places, of course, this has peeled away, leaving the rough, ragged edges of the stones exposed. Inside some of the subdivisions that appear to have been less used than others, the impressions of the hands and even the delicate lines on. the thumbs and fingers of the builders were plainly retained; in one or two cases a perfect mold of the whole inner surface of the hand was imprinted in the plastic cement. They were considerably smaller than our own hands, and were probably those of women or children. In the mortar between the stones several corn-cobs were found imbedded, and in other places the whole ear of corn had been pressed into the clay, leaving its impression; the ears were quite small, none more than 5 inches long. In the rubbish of the large house some small stone implements, rough indented pottery in fragments, and a few arrow-points were found. It is a wonder that anything is found, for it is more than likely that every house has been ransacked time after time by wandering bands of Utes and Navajos, who would search with keen eyes for any articles of use or ornament left after the first spoliation. The whole appearance of the place and its surroundings indicates that the family or little community who inhabited it were in good circumstances and the lords of the surrounding country. Looking out from one of their houses, with a great dome of solid rock overhead, that echoed and re-echoed every word uttered with marvelous distinctness, and all about a steep descent of 100 feet down to the broad fertile valleys, covered with waving fields of maize, the scattered groves of the majestic cotton wood, and the meanderings of the Rio San Juan, these old people, whom even the imagination can hardly clothe with reality, must have felt a sense of security that even the incursions of their barbarian foes could hardly have disturbed.

Soon after leaving the Casa del Eco, as we named the last ruins, our trail bore away to the right upon the plateaus, which now begin to encroach too closely upon the river to permit us to follow its course, and we come under a second line of bluff's, which were gradually surmounted also. The evidences of former occupation continue as numerous as ever, finding shape principally in cave-houses, all too near alike to bear much further repetition. A novel feature at one point is shown in Fig. 5, Plate 18, where we have a smooth bluff of cream-colored sandstone

about 150 feet high, with hardly a seam on its surface, over which has been cut a series of steps. Upon the pile of debris at the left are the ruins of some structure that had been built just beneath the line of footsteps, and was evidently placed there as an approach to them, as they only came down to within about 12 feet of the bottom. The large slab of rock lying against the bluff on the right was separated from it about 3 feet at the base, making a long, narrow passage-way, that could also be reached through the small opening between the rocks on the right; from within this place it was not difficult to reach the round bowlder lodged above, from which starts another line of steps. The surface of the rock has worn away to such an extent as to nearly entirely obliterate some of the holes, rendering ascent, at the present time, impossible; and as the bluff was inaccessible for two or three miles upon either side, we did nut reach the top or see from below any evidences of building.

Our trail over the bare plateau finally brought us down to the San Juan again, just at the junction of Epsom Creek with it, and but a short distance above the mouth of the Rio De Chelly, where we found a pleasant park like valley, about a mile in length, bordered by groves of cottonwood and willows. The beds of Epsom Creek and the Rio De Chelly were both perfectly dry, like all the tributaries of San Juan west of the Mancos, although in the latter were indications of occasional flooding, some of the deeper pockets retaining shallow pools of a very red muddy water. Upon every side—except where the broad valley of Epsom Creek opened northwardly, with the deep blue summits of the Sierra Abajo appearing in the vista—steep rugged bluffs of bare red rock rose up, weathering occasionally into sharp needle-like pinnacles, discernible for long distances in any direction. The San Juan emerged from a considerable canon at the head of this little valley but to sink into a still greater one below. The low stage of water encouraged us to explore this lower canon for a short distance, which we could readily do upon our riding-animals, the indefatigable little Mexico, our pack-mule, carrying the photographic instruments. We penetrated its exceedingly tortuous course about ten miles, meeting no serious obstruction, and it is likely could have gone much farther. The walls rose abruptly, generally perpendicularly, upon either side, from 1,000 to 2,000 feet in height, but always with a bench at the bottom bordering the stream, covered with a rough talus. Former floods and the drifting sands from the plateau above have filled up the interstices in the rocky mass, smoothing the way over them very considerably.

In the walls of the canon of the Chelly, where it opens into the park, are several great circular caves, averaging 100 feet in diameter, in which were the remains of walls and houses, but all very much dilapidated. Upon a ledge on the opposite side of the canon was a row of four houses, not easily reached, one of which still retained a roof ; and in another case, a shelter was formed by inclining a row of sticks across the opening of the cave, with the outside thickly plastered with clay. It had every appearance of being a more recent structure, yet it was in the midst of much older-looking ones, and in an almost, if not quite, inaccessible position.

Over the level surface of the valley the older form of ruins, indicated principally by broken pottery, occurred in several places, and also, on a bench bordering the San Juan, just above Epsom Creek, were a number of small squares arranged in circles, that we have heretofore assumed to be places of sepulture.

In going southward, up the Chelly, we find it necessary to avoid the canons and make a detour to the right, crossing a rugged depression in

the line of bluffs, to the valley of a small tributary, then over another divide across the upturned edges of the great fold spoken of in the first part of this article, to quite an expansion of the valley of the Chelly, about one mile square, covered with sage-brush and drifted sand, on the upper or right-hand side of which we are fortunate enough to find two springs of cool, fresh water, a most delicious luxury where the temperature of the water of the San Juan was 80 degrees, and the coldest to be had, and the temperature of the atmosphere away up in the hundreds during the day-time.

The surface of this valley, or small plain, contains indications of old ruins, about which we picked up many arrows, knives, and other stone implements, with the ever-present pottery. The wash of the Chelly skirts one side of the valley, with perpendicular bluffs 200 to 400 feet high, closely bordering its other bank. Above and below the opposite bluff's rise up again, throwing the wash into deep canyons. An examination of the exceedingly tortuous course of the wash and its accompanying bluff-line for a distance of 5 miles up and down revealed but one ruin (Plate 19), a very important and interesting one, however.

This cave-town occurs in a great bend of the encircling line of bluffs, where the wash makes a wide detour, perched upon a recessed bench about 70 feet above the valley, and overhung by a solid wall of massive sandstone extending up over 200 feet farther. The left-hand side of the bench supporting the buildings sweeps back in a sharp curve about 80 feet under the bluff, and then gradually comes to the front again until, on the extreme right hand, the buildings are built upon a mass of debris, but partially protected overhead. The total length over the solidly-built portion of the town is 545 feet, with a greater width in no place of more than 40 feet. There are somewhere in the neighborhood of 75 rooms upon the ground-plan, with some uncertainty existing as to many of the subdivisions on the right hand in the vicinity of d and e; but in the cave-built portion every apartment was distinctly marked. Midway in the town is a circular room of heavily and solidly built masonry, that was probably intended for an estufa or council-hall; that is, if we can reasonably assume any similarity in the methods of building or worship to those of the Pueblos of New Mexico. Starting from this estufa is a narrow passage running back of the line of houses on the left to the two-story group, a, where it ends abruptly, further access being had through the back row of rooms, or over the roofs of the lower front row, probably the latter, for it is likely that these roofs served as a platform from which to enter the rooms back of it. At the extreme end a still higher ledge occurs, with the overhanging wall coming down close over it, its outer edge inclosed by a wall, and a little store-room in its farther corner; it was reserved, probably, as an out-door working room. All the buildings of this half are of one story, with the exception of the group a, the residence probably of the chief or of some other important family in the community. The rooms just back of it are the store-rooms of the family, where the corn and squashes were put away for the winter's consumption. At the place marked b, near these store-rooms, there are two half-round inclosures of stone work, that are very likely the remains of small reservoirs or springs. The rock back of them is dug out beneath, and had, even in the dry season, when we were there, a damp appearance, as though water was not far removed, and might easily be coaxed to the surface. The front line of wall of this left side of the town is built upon a steep angle of smooth rock, with the interior of the apartments filled up with earth so as to make their floors level, bringing them a little

below the passage way. In two or three instances, as shown in the plan, the front wall has given way, precipitating all but the back wall to the bottom of the cliffs. Holes have been drilled into the rock in a few places beneath the walls, evidently to assist in retaining them in their places.

The whole front of this portion of the town is without an aperture, save very small windows, and is perfectly inaccessible, both from the solidity of the wall and the precipitous nature of the foundation-rock beneath it. Admittance was probably gained from near the circular building in the center, by ladders or any other well-guarded approach over the rocks.

In Fig. 3 of the same plate we have a view of the other half of the town. From the estufa we have to climb up about 8 feet, reaching a narrow ledge that starts out from the bluff ; from here to the farther end the buildings are built irregularly over the uneven surface of rocky debris, each house conforming to the irregularities by which it is surrounded, but all presenting the general arrangement of a cluster about a central court, as at d and e, that served, in all probability, as corrals for their domestic animals. In some places near these corrals the under surface has broken away, disclosing a solidly packed bed of old manure, very nearly resolved back into dust again, and through which were scattered twigs of willow and sticks of cedar. Some of the rooms are quite large, from 15 to 25 feet in length ; the very small rooms surrounding them were probably for storage, and in some cases seem to have answered the purpose of fire-places, as at f, for baking pottery, very likely. None of these buildings, as far as we could discover, were of more than one story in height. All the door-ways or windows opened from within the courts or corrals, and were unusually large, reaching in some cases the whole height of the wall. The front line was so broken down that it was impossible to tell to what extent it was accessible, although we may reasonably infer that it was not so, with the exception, perhaps, of a way for themselves and their animals. The bluff itself was easy to ascend, being composed of large rocks tilled up with smaller debris.

In their construction these buildings differ from any we have yet met, in the thickness, or rather thinness, of their walls, being very seldom more than a foot, and more frequently between that and six inches thick. The stones of which they are built are in long thin slabs, trimmed down roughly to the required size and laid in an abundance of adobe mortar. In most of the rooms, both the inside and outside have been smoothly plastered over with clay, and, where protected overhead, still retain that coating in fair preservation.

A few rods to the right is another smaller recessed bench, upon which are built two small houses, each about ten feet square, and one with its roof still entire. The approach from below is a smooth, rocky surface, so steep as to be almost impossible to ascend, with no remains of any other easier method of getting up.

At the foot of the bluff beneath that portion of the ruin marked d, in the ground-plan, a low bench rises about ten feet above the surround ing valley, upon which are indications of old buildings and of other remains—our so-called burial-places. Chipped flint-work was plentiful, as we found a number of very beautiful specimens of arrow-points, perforators, knives, and some domestic utensils. In a mass of debris at the foot of the two-story tower, seven large earthen pots of rough indented ware were found imbedded in the soil and filled with earth. They were too fragile to admit of transportation upon pack-animals, so

we put them carefully by for future investigators. In the rubbish at the extreme right, a handsome little jug or vase (see Fig. 5, Plate 21) was found, lacking only its handle. A careful search through the very thick deposits of debris would undoubtedly reveal many treasures, and we felt many regrets that we could not consistently devote a number of days to the pleasant undertaking. We can only expect to skim the surface, leaving to others hereafter the more satisfactory duty of exhausting each subject in detail.

In progressing southward we find it again necessary to climb the steep bluffs bordering the Chelly, here so tortuous and walled up as to be impracticable, if not impassable. Once on top, however, we made our way with comparative ease over great dunes of a very fine, yellowish-white sand, packed so solidly as to inconvenience the animals but very little ; but much the greater part of the way is over a solid floor of bare, nearly white, sandstone, rising into occasional dome-shaped hillocks, and furrowed by shallow ravines. Sage-brush, juniper, and pinion trees were scattered plentifully over the whole region, affording the only relief to an otherwise perfectly barren desert. Traveling thus over this trackless waste, we reach in about fifteen or twenty miles the bare red plains of the famous so-called diamond-fields of Arizona. Beautiful garnets were found scattered plentifully over the whole region, but they could not tempt us to linger, for the sun beat down upon its arid surface with such an intensity that but for the extreme dryness and salubrity of the atmosphere it would have prostrated anything but a salamander.

After crossing this plain we came suddenly upon a side canon running across our course, seemingly a mere gash in the rocky plateau, down into which we were fortunate enough to find a practicable way for ourselves and animals. But what a contrast! A smooth sward of grass, and thick patches of the tall reedy kind peculiar to damp localities, made a change grateful to both man and beast. Continuing down this canyon— which has, in consequence of its inviting appearance, been called the Canyon Bonito Chiquito—a couple of miles brings us to the wash of the Chelly again, bordered with groves of fine old cotton woods, but its bed, in which were pools of clear water, was so deep as to be almost inaccessible. A band of wandering Navajos just before us, with large flocks of sheep, had made a way down, however, that we found practicable.

An after investigation revealed the presence of water in large artificial reservoirs, or tanks, in the canyon Bonito, just above where we entered it first, about which are grouped a number of old ruins. This has been a favorite Indian wintering-ground, so that the ruins here have been much modified by their occupation.

Two miles down the canon of the Chelly we found the house shown in Fig. 1, Plate 15. Its situation is very similar to that of the town shown in Plate 19, but is overhung by a much less height of the impending bluff. It was reached from the valley by a series of steps cut into the rock, but now so weathered away as to be impracticable. It is accessible now by way of the ledge running to the left from the house, some JO or 12 rods in length, but affording a very narrow and precarious footing. At the time of occupancy this was walled across, with possibly a way for getting over or around, for this ledge communicated directly with the plateau above, where there are remains of what was possibly a corral.

The house consists of two stories, 20 feet in height, built against the sloping back wall of the bluff; the lower story is 18 by 10 feet square, divided into two rooms, one slightly smaller than the other, with a com-

municating door between, and a large door opening outward from the larger one. The upper floor appears to have been all in one room, with one large window facing outward, and much smaller ones in the side walls. Extensions existed upon either side, and also some kind of structure in front, probably a sort of platform-house, covering the lower doorway. To the right the ledge grows narrower, and gradually merges into the perpendicular bluff; 40 feet from the house, on this ledge, are the remains of a wall across it. About 20 rods above, at the foot of the bluff, there is a deep natural reservoir of water, formed by the accumulated rains upon the plateau above pouring over the rocks and scooping out a basin 30 feet in diameter and fully as deep, that seems to retain a perpetual supply of water.

Near our camp, just at the junction of the two canons, and on the flat surface of the sage-covered valley, were a row of small squares marked out by large stones set upright, such as have been already described. In this case they were of such careful construction and size as to encourage us to dig into them to a considerable depth. Beyond the scattered bits of charcoal, very sparingly deposited in this instance, however, nothing was found.

Five miles above the canyon Bonito, the Chelly expands into a wide valley that extends, with slight interruptions, to the foot of the canyon De Chelly, at the northern end of the Tunicha Mountains. It is bordered by low but abrupt sandstone bluffs, which have been broken into isolated monuments in some places, and stand like huge sentinels upon either hand, as if to warn the traveler from the desolation surrounding him. Although the bluffs contain numerous great circular caves, favorite building-places of the ancient builders, yet we find only two or three ruins of that kind, and only in the lower end of the valley, the last we noticed being about eight miles above the canyon Bonito. This was the largest and most important one in this vicinity, occupying a large circular cave very similar to the one of the San Juan, divided into twelve or fifteen rooms, with a large corral or court, and an elevated bench to one side, with a low wall running around its front edge. This had been occupied by the Navajos for corraling their sheep.

Over the broad, flat valley, sage-covered, sandy, and monotonous, and through which the wide shallow wash meandered from side to side, we found frequent indications of its former occupancy by the old people whom we have been following up, extending southward until lost in the cultivated region about the head of the valley. There were no more remains of stone-built houses, nor the slightest sign of one; all were probably of adobe, the only clew in many cases being simply a slight mound with considerable quantities of broken pottery surrounding it. Eight miles up the canon De Chelly are the ruins of a cave-town very much like the one described (Plate II), but much smaller, and with a ruined mass of houses at the foot of the bluffs below the cave like bench.* About the head of the valley the Navajo Indians have several hundred acres, in the aggregate, of corn, pumpkins, and melons under cultivation, taking advantage of the water which comes down thus far from the mountains to the east. From here our trail to the Moqui settlements branched off in a southwesterly direction to a low divide under the southern end of the Mesa Vaca, where it turned nearly south and hardly deviated from a bee-line for a distance of nearly 40 miles to Tegua, the nearest of the Moquis towns.

We will not now stop to discuss the question as to what connection may have existed between the ancient builders of the San Juan and the

* Simpson's Report.

present semi civilized people known as the Moquis, but return to the mouth of Epsom Creek and describe the many curious remains found north of the San Juan, all of which bear some relationship to those of the Hovenweep already noticed.

Fifteen miles up Epsom Greek a side canon comes in from the left, down which trickles a scanty stream of brackish water with the peculiarity of taste and action which has given the name to the whole valley. Camping here, we extended our observations up this lateral canyon some 8 or 10 miles in quest of ruins, and found them numerous enough to satisfy our most earnest desire, although not of the importance of the greater ones of the San Juan and De Chelly. All were of the small cave kind, mostly mere " cubby-holes," but so smoke-blackened inside and showing other marks of use as to convince us they had been long occupied but not during any comparatively recent period. In the generality of cases they were on small benches or in shallow caves situated near the bed of the stream, but the farther up we went the higher they were built. In one instance a bluff several hundred feet in height contains half a dozen small houses sandwiched in its various strata, the highest being up 150 feet, each of but one room, and one of them a perfect specimen of adobe-plastered masonry, hardly a crack appearing upon its smoothly-stuccoed surface. A short distance up from the entrance to the canon a square tower (Fig. 2, Plate 18) has been built upon a commanding point of the mesa, and in a position, so far as any means at our command are concerned, perfectly inaccessible. The stones of which it is composed are of a very nearly uniform size, more so than in any of the buildings we have seen west of the Hovenweep.

Upon the opposite side of the main Epsom Creek Valley, and on top of the high bluffs of sandstone which border it for nearly its whole length, we found some cave-houses in a most singularly out-of-the-way place—in the very last place in the world where one would expect to find them. Scaling the bluff at the very imminent risk of our necks, we came suddenly upon a broad open cave, near the top, containing the usual style of stone-built and mud-plastered houses, divided into four or five apartments, of just the size and number that would be required by an ordinary family of eight or ten persons. Farther up, on top of the bluff, we found the remains of a circular tower 40 feet in diameter, and very old, the stones all crumbled, rounded, and moss covered. Near by were remains of two other cave-habitations.

A few miles farther up the Epsom Valley, passing a number of old ruins hardly worthy of mention, we came upon an important group that was evidently the center of the surrounding population—a place of worship or of general congregation—an aboriginal shire-town.

It lay upon both sides of a small, dry ravine, some 20 or 30 rods back from the bed of the creek, and consisted of a main rectangular mass 60 by 100 feet square, occupying quite an elevation, dominating all the others. Just below it, and close upon the edge of the ravine, was a round tower 25 feet in diameter; and 75 feet below that, and also close to the ravine, was a square building, 20 feet across, nearly obscured by a thicket of pinon trees growing about it. On the opposite bank were two small round towers, each 15 feet in diameter, with two oblong structures between, 12 by 15 feet square; at right angles to these four, which were arranged in a straight line, another square building occurred, the same size as the one just opposite on the other bank. Portions of the walls of the towers remained, and a few courses of stone in the walls of the smaller square buildings, but in the large ruin the walls were merely indicated by great mounds of crumbling rock, with the subdivisions distinctly marked, however, into four rectangular apartments. A short

distance above, plenty of water was found in the bed of the creek ,• fine large cotton woods bordered the stream, and the broad fertile valley seemed a far more desirable place of residence than the forbidding desolation of the Chelly.

About thirty miles from the San Juan we left Epsom Creek, and stopped for a night at the head of the canons which run between it and the Montezuma. We were in the midst of quite a thickly-settled, au-cient population, the ruins of their habitations consisting almost entirely of the kind just described—low, rectangular mounds, so completely destroyed as not to leave one stone upon another, yet accompanied always by an abundance of the same kind of pottery we have found so universally distributed over other localities. Not the slightest difference can be detected in its general quality, nor can any one style of manufacture or ornamentation be said to be peculiar to any one district or group of ancient habitations. It is the same with arrow-points and like work, and with the similarities in building; although covering two different periods of their existence, it carries the conviction to us that they were all one and the same people, scattered in families and communities throughout the valleys and canons.

After leaving this last group of ruins, all traces of them suddenly ceased, and in the four or five days spent in the examination of the country upon the southern, eastern, and northern flank of the Sierra Abajo, not a single vestige was found ; and this in, without exception, the most pleasant spot we have touched since leaving La Plata. Clear and cold mountain-streams ripple down through ravines overhung by groves of willow, maple, and quaking-asp, with splendid oaks, and stately pines scattered over the uplands, and an abundance of rich, nutritious grass everywhere, that our poor, half-starved animals knew well how to appreciate. The black-tail deer and grouse were in goodly numbers, starting up from under our very noses, and leading our hunters many a long chase.

Leaving half of our little party of six, and all the animals but those we rode and the trusty Mex. with the apparatus, we made our way down through the deep and narrow canons that lead from the plateau country into the great basin that lies between the Sierra Abajo and the Sierra La Sal, and spent two days in the examination of its arid surface, covered with monumental rocks and ridges, but without coming across so much as a piece of pottery or an arrow-point.

Turning our backs upon the Abajo Peaks, we struck out northeasterly over the plateau, but soon finding a trail bearing southeast, followed it until we saw that it was likely to continue some time upon the plateau, when we branched off to the left, and in a short distance came upon the very brink of the deep canon of the Montezuma, one of the far-reaching arms of the main wash and valley farther east. Winding our way among rocks and scrubby pinons, we almost literally tumbled down the precipitous descent of 1,500 feet, to a narrow bottom, walled in first by a broad belt of massive white sandstone, rising almost perpendicularly from 20 to 50 feet above the valley ; above that the dark red and shaly sand-rocks rose up in receding benches 1,000 feet to a broad tablet of white sandstone on top, so high up that it seemed to shut out all the world and to leave us as engulfed in the bosom of the earth. A narrow but deep "wash" meandered from side to side, containing just a few scattered pools of stagnant water, while dense thickets of oak brush, thickly interwoven with vines, rendered progress anything but pleasant.

We had gone but a few rods before we commenced picking up pieces of pottery and meeting other evidences of occupation ; within three

miles a cave shelter appeared, and then as the valley widened it was dotted in many places with mounds thickly strewn over with the ever-accompanying ceramic handiwork of the ancient people in whose footsteps we are following, and occurring so frequently and of such extent as to excite astonishment at the numbers this narrow valley supported. The line is so sharply drawn that in an hour's ride all traces of any ruins are lost; and there is not so much as a piece of pottery to show that these people had ever extended their residence beyond the limits of their canon.

Soon other cave-dwellings appear, most of them little walled-up circular orifices in the rock, generally inaccessible, but many were approached by steps, or rather small holes cut in in such a manner as to enable the climber to ascend the rock as by a ladder. Examples of these kinds of ruins are shown in Figs. 1 and 4 of Plate 17, each about 40 feet above the valley, the first perfectly inaccessible and without the least sign of the original method of reaching it; in the other one the walls once closing it have been pushed down so that only traces of them remain; the steps leading up, however, show it to have been considerably used ; they are now so worn down by the disintegrating influences of time as to no longer answer their purpose.

Throughout this canon we find frequent examples of the footsteps cut in the rock, in the generality of cases being simply a way of scaling the smooth, nearly perpendicular wall of sandstone, which hems in the canyon on both sides for twelve or fifteen miles; probably a ready mode of escape up the bluff should enemies appear.

The cliff and cave dwellings, very small habitations, seldom larger than the one in Fig. 2, Plate 15, appear to occur in groups, not always in connection with the old valley ruins, but rather to alternate in succession as we progress down the canyon.

In one of the cave-dwellings, Fig. 3, Plate 17, perfectly black with long-continued smokes inside, and bearing other marks of long use, we found the complete skeleton of a human being; the remains, as afterward determined, of a young man somewhat under a medium size. The excrement of small animals, dust, and other rubbish filled the floor of the little house a foot deep, nearly burying the scattered bones; with them are the shreds of a woolen blanket, woven in long stripes of black and white, just such as the Navajos and Moquis make at the present time. It is likely that the remains are those of a Navajo, a people who occupied all this country up to within a short time, within the remembrance of the older persons, and who were driven beyond the San Juan by the onslaughts of the aggressive Utes.

After traveling about 20 miles from our starting-point at the foot of the mountains, half of the way in the canyon, we camped at the intersection of a large canon coming in from the west, traversed by a large, well traveled Indian trail, that continued on down, probably the same one we had crossed earlier in the day. At this point the bottoms widened out to 200 to 300 yards in width, and are literally covered with ruins, evidently those of an extensive settlement or community, although at the present time water was so scarce—not being able to find a drop within a radius of six miles—that we were compelled to make a dry camp. The ruins consist entirely of great solid mounds of rocky debris, piled up in rectangular masses, covered with earth and a brush-growth, bearing every indication of extreme age; just how old is about as impossible to tell as to say how old the rocks of this canon are. This group is a mile in length, in the middle of the valley-space, and upon both sides of the wash. Each separate building would cover

a space, generally, of 100 feet square; they are seldom subdivided into more than two or four apartments. Relics were abundant, broken pottery and arrow-points being especially plenty, and of excellent quality; at one place, where the wash had partially undermined the foundations of one of the large buildings, it exposed a well of regularly-laid masonry extending down six feet beneath the superincumbent debris to the old floor-level, covered with ashes and the remains of half-charred sticks of juniper. From this rubbish a fine example of a stone ax, about the size of one's hand, was found, with a smooth and sharp-cutting edge, formed by grinding it down to an acute angle; its head was roughly chipped to the required shape for binding on a handle. At another point a small earthen bowl, of the superior ware characteristic of the people, was found entire. No special burial places were observed, but a number of bones of the lower extremities were unearthed at the edge of the wash, without any stone-work above them. There were no cave-dwellings in the neighborhood of this group, but two or three miles below several occurred, one of which is built in a huge niche in the solid wall of the canyon, with its floor level with the valley. (See Fig. 2, Plate 17.)

From the last camp the canon expanded into occasional valleys from 500 to 800 yards across, and then contracting again to a mere narrow passage, but still all shut in by the high escarpment of the mesa. From either side long narrow tongues or promontories extended out 100 yards, and from 20 to 100 feet high, sometimes connected with the main wall by a mere comb or wall of rock, its extremity, however, spreading out to an irregularly oval shape. In the valleys are occasional isolated mesas, the remnants, probably, of former promontories, left here by the great erosive powers which channeled out these canyons. Within a distance of 15 miles there are some sixteen or eighteen of these promontories and isolated mesas, varying in height, and every one of them covered with ruins of old and massive stone-built structures. They will average in size from 100 by 200 feet square down to 30 by 50 feet, always in a solid block, and, with one exception, so nearly similar that a description of one will fairly represent all. This exceptional instance is explained in the sketch (Fig. 1, Plate 20), and the ground-plan (Fig. 1, Plate 16), The peculiarity here consists principally in the size and shape of the stones employed, as well as in the design of its ground-plan. The ruin occupies one of the small isolated mesas, whose floor is composed of a distinctly laminated sandstone, breaking into regular slabs from 18 to 24 inches in thickness; these have been broken again into long blocks, and then placed in the wall upright, the largest standing five feet above the soil in which they are planted. The sketch in Plate (20) is a view along the line a a a, looking toward the round tower. Very nearly the entire length of this wall is made up of the large upright blocks of even thickness, fitting close together, with only occasional spaces filled up with smaller rocks. In one place the long blocks have been pushed outward by the weight of the debris back of it. One side of the large square apartment in the rear is made of the same kind of rocks, standing in a solid row. The walls throughout the rest of the building are composed of ordinary-sized rocks, with an occasional large upright one. Judging from the debris, the walls could not have been more than 8 or 10 feet in height. The foundation line was well preserved, enabling us to measure accurately its dimensions. The large square room was de-jjressed in the center, and its three outside walls contained less material than in the rest of the building. No sign of any aperture, either of window or door, could be detected.

The far more numerous class of ruins occupying like mesas and the promontory points consist of a solid mass of small rectangular rooms, arranged without any appearance of order, conforming to the irregularities of the surface upon which they are built, and covering, usually, all of the available space chosen for their site. All are extremely old, tumbled into indefinite ridges 5 or 6 feet highland as broad, with the stones partially crumbled into sand, and all covered with sage-brush, greasewood, and junipers. They occupied every commanding and available spur of the mesas, usually so placed in the bends as to afford a clear lookout for considerable distances up and down the canon. They resemble in this respect the sites chosen by the Moquis in building their villages; but we are not able to trace the resemblance further, from the extremely aged and ruinous state in which these remains are found. Between these fortresses and on the level bottom-lands, generally close up to the bluff upon either side, are occasional smaller ruins, resembling those at the dry camp. In connection with these a peculiar feature is shown in Plate 20, Fig. 2. At the foot of one of the promontory towns a low bench, tongue-shaped, and only about 10 feet above the valley, runs out from the mesa 200 feet in length and half as broad, through the center of which runs a wall its entire length; a portion of it is composed of the large upright rocks shown in the sketch, the largest standing seven feet above the surface and evidently extending some distance below, in order to be retained so firmly in their places. There are only seven of these standing, placed about 5 feet apart, the rest of the wall-line being composed of a low ridge of loose rock extending up to a mass of old ruins at the foot of the bluff. One side of the space divided by this wall is filled with a great pile of rocks arranged in irregular lines inclosing areas from 20 to 50 feet in diameter, the whole indicating a very considerable structure.

Grouped among the lower end of these towns were a number of the small cliff-houses; a regular colony of them occurring at the first bend of the West Montezuma, about a dozen miles above its junction with the east fork. An exceedingly well preserved and peculiar one is shown in Fig. 1 of Plate 18. A block of sandstone setting on the edge of a mesa bench 50 feet above the valley has a deep oval hole worn in it, probably by natural agencies, which is nearly entirely occupied by a very neatly-built little house 10 feet long, 6 high, and 5 deep. A space at one end is reserved just large enough to serve as a platform to enter from.

Below the bend in which these cliff-houses occur, the Montezuma loses its canon character and spreads out into a wide, barren valley, thickly covered with tall sage-brush, and the wash lined with large cot-tonwoods. The mesas upon either hand dwindle down considerably in height and abruptness, and seldom contain ruins. The large square buildings on the bottom-lands, however, are important features, and were it not for their great number, and the exceeding indefiniteness of their outlines, we might linger longer and describe each in detail. Over all are found immense quantities of broken pottery, many examples of which are shown in the accompanying plates. Arrow-points and like chipped work were especially numerous, and a great many of small size and great delicacy of finish were found.

It should have been mentioned that running water occurs in the Montezuma at the bend spoken of, and a band of Weminuche Utes, who now occupy these canons, have considerable corn planted there. It is not impossible that formerly water was constant throughout the whole length of the region occupied by these ruins. Below the junction of the east and west forks of the Montezuma, the valley must have been

always hopelessly barren and dry, for not a vestige of any ruin occurs. At its mouth and along the San Juan, as we have noticed, they appear at once again in considerable numbers.

Our investigations closed with a side trip up into some of the sterile gorges between the two forks of the Montezuma, but without any results worth especial mention. A few small cliff-houses occurred, and a few scattered remains about the open lowlands. Upon the tops of the mesas in this vicinity, as well as upon those between the Montezuma and the Hovenweep, were old remains of towns. A glance at the accompanying map will give some idea of their distribution, although those about the head of the West Montezuma are only located approximately, in the absence of any precise notes of its topography.

POTTERY.

On the opposite page, Plate 21, we have grouped a few of the most striking examples of the pottery collected upon the trip, and which forms by far the most interesting of all the relics of the ancient people of the San Juan Valley. All who have ever visited this region, which extends from the Rio Grande to the Colorado, and southward to the Gila, have been impressed with the vast quantities of shattered pottery scattered over the whole land, sometimes where not even a ruin now remains, its more enduring nature enabling it to long outlive all other specimens of their handiwork. It is especially interesting as enabling us to see at a glance the proficiency they had attained in its manufacture and ornamentation, displaying an appreciation of proportion, and a fertility of invention in decoration, that makes us almost doubt their ante-Columbian origin; but nevertheless, without going into the details, we believe them to antedate the Spanish occupancy of this country, and to owe none of their excellence to European influences, being very likely an indigenous product.

Fig. 1 is a jar from the valley of Epsom Creek, of dark gray and rather coarse material, without color or glaze, of the indented and banded ware peculiar to the ancient artificers only. It is made by drawing the clay into ropes, and then, commencing at the bottom, building up by a continuous spiral course, each layer overlapping the one under it, the indentation being produced by a pressure with the end of the thumb, and by a slight doubling up of the cord of clay. The design is varied by running several courses around quite plain. Its diameter was 18 inches, with the same height, and 9 inches across the mouth. For so large a vessel it was very thin, not more than one-fourth of an inch. Inside, the surface was rubbed perfectly smooth.

Figs. 2, 3, 11, are restorations from well-preserved fragments of mugs or cups, each elaborately ornamented in black on a white glazed ground ; the last one, especially, is of firm, excellent ware, and the design put on with great precision. The first two are 3 1/2 inches in diameter and 4 inches high, and the last one 4 1/2 inches in diameter by 5 inches in height.

Fig. 4 is a flat disk of pottery for covering a jar.

Fig. 5 is the small jug found at the great cave ruin on the De Chelly (Plate 19), 3 1/2 inches in diameter, of dark gray ware, perfectly round, and very neatly painted. The handle has been broken off, but leaving the marks where it had been attached.

Fig. 6. A slightly oval-shaped jar, 10 inches in diameter, and a mouth 5 inches wide, with the lip rolling over sufficiently to attach a cord to carry it by.

Fig. 8. A small jug with side-handles and narrow neck, 4 1/2 inches in diameter and 1 3/4 inches across mouth.

Fig. 9. A cup or dipper from Montezuma Canyon; bowl, 3 1/2 inches diameter; handle, 4 inches long.

Fig. 12. A pitcher, taken from a grave on the banks of the San Juan, near the mouth of the Mancos, by Captain Moss. In the same find were other similar vessels, some polished stone implements, and a human jaw-bone. The ware of this pitcher is a coarse, gray material, somewhat roughly modeled, but of fine form and tasteful decoration.

Fig. 10 is a peculiar vessel, found among the Moquis of Tegua. They could give no account as to where it came from or who made it. It is probably of Zuni manufacture. The material is rather soft, being easily cut with a knife. The upper portion is painted or glazed white, and the lower red; the figures are painted in red and black. The tallest portion is six inches in height.

Fig. 7 is an example of the modern work of the Moquis of Tegua. The material and workmanship are far below any of the preceding examples; approaching them only in its ornamentation, which is strictly inventioual, but somewhat bizarre.

Plate 22.

This plate is intended to represent some of the most striking instances of taste and ingenuity in the ceramic decoration of the nameless potters, all the examples selected being, with one exception, from vessels of the general form of Fig. 1. With but very few exceptions the ornamentation is on the inner surface, generally in the form of a band, from 1 inch to 4 or 5 inches in breadth, but in many cases it covers the entire inner surface. If the outside of the dish is painted it is in the form of a simple narrow band, like Fig. 15. These dishes or bowls vary in size from 13 inches in diameter (Fig. 9) to mere cups of only 5 inches (Figs. 18, 19).

The ware is dark gray and nearly white; hard and firm, giving a clear ringing sound when struck. It varies in thickness from 3/16 to 3/8 of an inch. Many of the specimens have a fine glossy glaze upon which the black design lies without any perceptible wearing away. Figs. 7, 8, 10, and 16 are good examples, all the others but a trifle less so. This is the more noteworthy from the fact that all have been exposed in open places to all the disintegrating influences of soil and climate for probably hundreds of years. Fig. 2 appears to have suffered the most, but the white ground has worn out, leaving the black design in relief. In some, as in Fig. 2, the design is jet black, running through intermediate shades of a reddish black when the color has run thin, down to where the design is quite faint; whether from the washing away of the paint or whether it was originally so, would be hard to tell.

Fig. 4 is a design that occurs frequently in the bottom of the dishes; we found half a dozen of the same general form. Circles with many radiating points, like a delineation of the sun, also occur quite often.

In Fig. 21 we have the only example yet found from among the ancient pottery of any attempt at imitational ornament. As to what the figure is intended to represent would be rather difficult to decide satisfactorily. This fragment came from the upper canyon of the Montezuma, and represents the inner surface of the bowl ; it is reduced to about one-third in the plate.

Fig. 20 is evidently a portion of the neck of a jug or like vessel of the rough gray ware, of which Figs. 1 and 12 of Plate 21 are composed. It is another rude attempt in plastic material at the imitation of animal life, and evidently is intended for a frog. This comes also from the Montezuma Canon.

LONG BONES.


	There are eight long bones among the remains, viz: 1 humerus, right. 1 ulna, right. 1 radius, left. . 1 femur, left. 1 tibia, left. 1 fibula, left. evidently belonging with the skull marked No 1. Besides these are, 1 femur, right. B. 1 femur left. which should undoubtedly be correlative with skull No. 2.

Description and measurements of bones A.

a. Humerus,—If we may judge from what is known in general about the proportions of the human body, the long bones combined under the head A belonged to a person about 5.6 feet high. The greatest length of the humerus, measured from the apex of the head to the inner and lower border of the trochlea 32^cm.2; greatest diameter of head near the anatomical neck 4^cm.8. Muscular insertions very marked. An-tero-posterior diameter of shaft 7^cm.5 below insertion of supra spinatus 2cm.2. Transverse diameter 2^cm.4; circumference 7^cm.3. The same measurements in the middle of the bone l^cm.7, 2^cm.3, and 6^cm.6, respectively, and l^cm.5 above the olecranon depression l^cm.8, 3^cm.6, and 9^cm.8. Extreme distance between outermost point of external and innermost of internal condyle 6^cm.l.

b. Ulna.—Distance from olecranon to articular part of the head of the lower extremity (styloid process broken) =26^cm.3 ; muscular insertions everywhere strongly pronounced. Greatest circumference of shalt measured 3^cm.7 below the highest point of coronoid process, 5^cm.3; antero-posterior and lateral diameters, in the same plane l^cm.8 and l^cm.6, respectively.

c. Radius-—Distance from highest point of head to lowermost of styloid process 24^cm.6; diameter of head 2^cm.2; depression of its upper surface rather deep; ridge for attachment of portion of supinator brevis uncommonly pronounced. The shaft does not offer anything extraordinary. Autero-posterior and transverse diameters of lower extremity 2^cm.3^cm and 3.4, respectively.

d. Femur.—The length of the femur, measured from the upper surface of the head to the inner condyle, is 45^cm.5, while the greatest distance between the great trochanter and the outer condyle is 43^cm.4. Greatest width measured between the depression for the ligamentum teres and the insertion of the gluteus medius 9^cm.9; greatest diameter of head 4^cm.6; depression for the attachment of the ligamentum teres is less ovoid than usual, and more of a reniform shape, its two diameters being 2^cm.0 and l^cm.5, respectively. Neck very strong; greatest horizontal diameter 2^cm.6; greatest vertical diameter 3^cm.2; anterior surface without any vascular foramina and concave, while the posterior surface is more convex than usual. Antero-posterior diameter of shaft l^cm.5 below the base of the lesser trochanter 2^cm.5; transverse diameter measured in the same plane 3^cm.4. The same measurements taken in the middle of shaft 2^cm.4 and 2'-^m.3, and 3^cm.4, above the upper and external border of the trochlear surface 3^cm.0 and 4^cm.2, respectively. The three cir-

cumferences measured successively in the same planes 9^cm.5, 7^cm.9, and ll^cm.9. In proportion to the massiveness of the bone the linea aspera not very prominent; arch for the femoral vessels not indicated. Distance from outer to inner tuberosity of lower extremity 8^cm.2.

e. Tibia and fibula.—The length of the tibia, belonging with the femur above mentioned, is 38^cm.l, as measured from the highest point of the spine of the head to the most distant point of the internal malleolus. Greatest antero-posterior and transverse diameters of head 5^cm.0 and 7^cm.9, respectively. Shaft rather platycnemic; antero-posterior diameter 6^cm below the attachment of the ligamentum patellae 3^cm.7 ; transverse diameter at the same place, 2^cm.5; circumference 9^cm.8. The same measurements taken in the middle of shaft 3^cm.3, 2^cm.l, and 8^cm.5; and repeated, 7^cm above the lowermost point of the internal mal-leolus=2^cm.7, 2^cm.8, and 2^cm.8. The length of the fibula, from the sty-loid process to the external malleolus, 37^cm.9 ; the latter projecting 2^cm.7 below the inferior surface of the tibia at place of articulation. Antero-posterior diameter of shaft measured in the middle l^cm.7 ; transverse diameter l^cm.0, and circumference 4^cm.8. If the femur and tibia be placed in natural connection the former appears to be much curved, as if it had been attached to a broad, woman like pelvis ; the skull belonging with the bones in question rather exhibits male characteristics than female.

For the sake of comparison, we shall add the following table, containing some measurements of long bones from Kentucky mounds taken by the late Jeffries Wyman.*


	Indians from Kentucky mounds. H. U. R. F. T. Humerus = 1.000 28 21 18 34 28 Ulna =0.816 Maximum............................. 337 284 270 479 397 Radius =0.758 Minimum............................ 283 214 215 383 317 Femur —1.000 Mean................................ 310 253 235 438 363 Tibia —0.829

Description and measurements of bones B.

The bones designated by B are two femora belonging together, consequently it will be sufficient to consider only one of them, and we shall choose the right one for this purpose. Length from upper surface of head to lowermost point of inner condyle 35^cm.6. Distance between uppermost point of great trochauter and lowermost of outer condyle 33^cm.l. Distance from outermost point of the head to outermost point of trochanter major 7^cm.2. Greatest horizontal diameter of the neck l^cm.8; greatest vertical diameter 2^cm.5; diagonal line of the outer surface of the great trochanter not prominent; triangular surface for the attachment of the tendon of the gluteus medius smooth. Antero-posterior diameter of shaft 2^cm.3; below the summit of lesser trochanter l^cm.8; transverse diameter 2^cm.4. Same measurement in the middle of bone 2^cm.0 by 2^cm.2, and 2^cm,9 above the upper and external border of trochlear surface 2^cm.l by 3^cm.0. The three circumferences measured successively in the same planes, 6^cm.7, 6^cm.2, and 9^cm.1. Arch for the reception of the femoral vessels very deep.

* Fourth annual report of the trustees of the Peabody Museum of American Archaeology and Ethnology. Boston, 1871, p. 19.

4 X

DESCRIPTION OF SKULLS.

The three skulls forming the more interesting part of the remains are designated by Nos. 1, 2, and 3, respectively.

Skull No. 1.—Plates 23, 24, and 25, (Fig. 5.)

This is almost perfect, with the exception of the right styloid process, which is missing, and the two condyles articulating with the atlas, which, according to all appearance, were broken off quite recently.

Viewed from above, the outline of the massive skull is somewhat heart-shaped, and but little irregular, the irregularity being chiefly due to the depression of the posterior portion of the left parietal bone. This depression extends also, but less marked, a short distance over the right parietal, inclosing a somewhat triangular space, and likewise over the occipital, down to the protuberance. Serration of sagittal suture coarse. A small Wormian bone in the coronal near the left temporal ridge. Parietal tubers very pronounced.

In profile the considerable height of the skull is very striking, as is also the depression, if the left side be turned toward the observer. Occiput, from a plane laid horizontally through the tubera parietalia downward, straight, almost perpendicular, the central portion of the occipital only bulging out slightly. Temporal ridges pronounced ; mas-toid process heavy; meatus rather elongated. Upper portion of frontal, from the region of the tubera, receding; supra-orbital arches moderately convex. Naso-frontal angle considerable, indicating aquiline features ; anterior nasal spine prominent. There is a slight tendency to progna-thism.

A frontal view shows a moderately broad forehead ; distance between the orbits rather considerable; left supra orbital foramen exceedingly small; the right supra-orbital arch shows 3 foramina. Malars massive; incisive fossae deep; canine eminences considerable. Muscular insertions of the lower jaw invariably very pronounced, but the exterior oblique line comparatively weakly developed. Height of ramus 7^cm.l. Teeth normal, their crowns slightly worn down, especially those of the incisors, but in excellent state of preservation.

In a base view the deformation shows almost as much as in profile. The left mastoid process, being more obtuse than the right one, is pushed forward, and, in connection with this the foramen magnum is somewhat asymmetrical and distorted. Left posterior condyloid foramen wanting; the spot where it ought to be is, however, very thin ; and, as, in similar cases, the size of the existing foramen is greater than usual. The effect of distortion may be noticed along the entire left portion of the base, on which side the eruption of the last molar never took place, although the right one cut through.

Viewed from behind, the outline of the skull is hexagonal. Parietals sloping considerably from the sagittal suture to the tubera, from which they form an almost perpendicular line to the mastoid portions of the squamosals. Lower outline of occipital but slightly convex. The internal occipital protuberance is but faintly indicated, especially if compared with other muscular insertions, but the ridge is strongly marked. While the right depression for the insertion of the rectus capitis posti-cus is deep and well developed, the opposite is but shallow.

Before giving the conventional measurements it may be well to state that the indices of height and breadth were intentionally omitted, as, on account of the deformation of the skull, they would be of little value.


	Measurements. c. cm. Capacity............................................-........................ 1325 mm. Length..................................................................... 168 Breadth.................................................................... 144 Breadth of frontal........................................................... 119 Height.............................-....................................... 135 Frontal arch ............................................................... 319 Parietal arch................................................................ 346 Occipital arch............................................................... 234 Longitudinal arch........................................................... 359 Circumference.............................................................. 508 Length of frontal........................................................... 123 Length of parietal..........................................................., 134 Length of occipital.......................................................... 160 Zygomatic diameter......................................................... 144

Skull No. 2.—Plates 25 (Fig. 2), 26, and 27.

According to its features, the skull designated by No. 2 is that of a woman, and is, with the exception of the zygomatic bones, which are broken, well preserved. If, among the Indians, the eruption of the last molars takes place at the same time as in our race, the skull in question cannot have belonged to a person much younger or older than seventeen.* In general, the skull is delicately built, being rather small, showing smooth surfaces and weak muscular insertions.

Viewed from above, it presents outlines similar to the skull previously described, but while the latter is deformed on its left side, the one in question is compressed on its right and is asymmetrical in a far higher degree.

In a profile view we notice above all, highly developed prognathism, a rather low forehead, and the outline of the occiput to be nearly straight. This latter fact is brought to view more strikingly if we look upon the right side of the skull, where we can lay almost a straight line from near the tuber over the surface of the parietal to the mastoid process, touching within a few millimeters the lambdoidal suture. If we measure the distance from the glabella to the left superior curved line of the occiput, it will be l^cm.3 greater than if measured between the same points on the opposite side, thus showing clearly the asymmetry of the skull.

A view from behind exhibits this less than one from the base, the latter revealing a distortion extending from the right side of the occipital to the narrow and elongated palate process of the maxilla, which is pressed forward. As in almost every instance of prognathism, the foramen magnum is thrown backward. This is, however, not only due to the protruding position of the upper maxilla, but also to a great extent to the flattening of the occiput, causing the occipital bone to turn upward almost abruptly within about half a centimeter from the posterior margin of the foramen magnum, that is on the right side of the skull, there being a little more space on the left. Both posterior condyloid foramina are well developed, the right one being larger than the left. The muscular insertions on the left side of the occipital are more strongly marked than those of the right, especially the places of insertion of the complexus and the rectus capitis. The protuberance is

* As puberty takes place sooner among Indians than among white men in general, we may conclude that the eruption of the last molars occurs earlier with the former races than with the latter.

scarcely indicated. The lambdoidal suture shows two Wormian bones where it joins the sagittal, one being situated to the right, the other to the left of the latter, measuring 2^cm.2 and 3^cm.2, respectively. Another but smaller Wormian bone may be noticed, where the left lambdoidal joins the squamous suture. The rnastoid processes are rather obtuse and small, the digastric fossa of the left being longer and deeper than that of the right, while the right mastoid foramen is larger and situated somewhat higher than the one opposite. The parietal eminences are not very distinct, although they may be recognized by viewing the skull from above; the temporal ridges are not indicated at all, and the right parietal foramen is situated before the left one and is about twice as large. Coronal suture rather straight, without many serrations. The frontal bone is without any median ridge or trace of suture, the eminences are but slightly developed, the superciliary ridges scarcely indicated, and the temporal ridges only well marked in the vicinity of the supra-orbital arches. The angle formed by the nasal bones with the frontal, and the curvature of the nasal bones are rather flat, nasal meatus rather round, anterior nasal spine moderately protruding. The number of teeth in the upper jaw complete, the last molars on each side in the act of eruption. Teeth rather small, with the exception of the incisors, which are directed much more obliquely forward than downward; canines smaller than the incisors; only the crowns of second and third molars ground flat; the other teeth of the upper jaw scarcely showing any wear, while the incisors of the lower bear the marks of extensive use ; not so, however, the molars, which are almost intact, at least those on the left side. Symphysis of the lower jaw scarcely visible, mental process tolerably developed, and incisive fossa of the right deeper than that of the left. External oblique line only well indicated near the anterior border of the ramus. Coronal process rather pointed; insertion of the internal pterygoid muscle less than usually rough.


	Measurements. *c. cm.* Capacity..................................................................... 1020 *mm.* Length..................................................................... 150 Breadth.................................................................... 131 Breadth of frontal........................................................... 106 Height..................................................................... 130 Frontal arch.....................................-............................ 278 Parietal arch................................................................ 325 Occipital arch............................................................... 226 Longitudinal arch........................................................... 342 Circumference.............................................................. 450 Length of frontal........................................................... 118 Length of parietal.......................................................... 117 Length of occipital ......................................................... 150 Zygomatic diameter *....................................................... 116

Skull No. 3.—Plates 28 and 29 (Fig. 13).

The skull designated by this number is rather thin, and was broken into a number of small fragments when brought from the field, and the pieces themselves were very fragile. It could, however, be sufficiently well restored to convey a good idea of its general features, which are decidedly masculine and even animal-like.

* The zygomatic bones being broken, the zygomatic diameter could only be ascertained approximately. Probable errors ± l^mm.5.

Although the occipital bone is almost entirely wanting, it may still readily be seen that the skull had been deformed in a manner analogous to the two others, and as it seems, chiefly on its right side, as a view from above will demonstrate, although the depression of the occipital region does not seem to have been as great as in either of the cases above mentioned. The outline of the cranium is more oval than that of No. 1 and No. 2, and appears somewhat flattened near the median line along the parietals. The superciliary ridges protrude very considerably beyond the general outline of the frontal, and in the same view, from above the zygomatic bones, the nasals and the upper maxillae may be seen to project. There are no traces left of the sagittal suture; the direction of the latter cannot even be detected if the skull be held against a strong light; but the coronal is plainly visible. As already stated, the occipital bone is almost entirely broken, there being left only a narrow piece of its upper portion, which measures about 9 centimeters in length. The right side of the skull is of a dark brown color, which is frequently noticed on bones exhumed from peat-bogs. An examination proved the coloring matter to be chiefly iron. Like the rest of the surface the right side, besides being colored, is covered by numerous fine cracks and grooves, either produced by running water that dissolved the carbonate of lime of the osseous tissue, or by the roots of plants that assimilated the salts. The side under consideration has the exterior portion of the mastoid process, which is rather massive, preserved. Temporal ridge extremely developed ; zygomatic process strong. Frontal bone receding almost as much as in some ancient Peruvian skulls that lie before me, and which are artificially deformed. Superciliary ridges very prominent; nasals strongly curved, aquiliue; upper maxilla prognathic in a considerable degree.

In a front view the heavy superciliary ridges appear very marked; supra-orbital arches strong; foramina large; anterior nasal spine projecting; septum oblique, its posterior portion shifted toward the right from the median line. First left incisor partly decayed; crowns of the teeth, especially on that of the molars, ground down. Insertions of muscles of lower jaw strong; those of the left more developed than the right; mental process prominent. Height of ramus from angle to highest point of condyle, 6^cm.8; sigmoid notch deep ; distance between the outermost points of right and left condyles, 12^cm.2.


	Measurements. c.cm. Capacity..................................................................... -----mm. Length..................................................................... -----Breadth..................................................................... 136 Breadth of frontal............................................................ 119 Height...................................................................... -----Frontal arch................................................................. 297 Parietal arch................................................................. 338 Occipital arch................................................................ -----Longitudinal arch............................................................-----Circumference.................................................,..............—— Length of frontal............................................................. 126 Length of parietal............................................................ 132 Length of occipital............................................................ -----Zygoniatie diameter.......................................................... 144

Thanks to the kindness of Dr. A. Otis, who, with his usual liberality, placed the osteological collection of the United States Army Medical Museum at my disposal, I could examine the two skulls collected by

Doctor Yarrow in the vicinity of the rains near Abiquiu, to which I took occasion to refer on one of the preceding pages. One of them, bearing the number 1178 in the catalogue, is that of a child about ten years old. The other, marked 1179, is that of an adult, and exhibits the general characters of a female.

We shall first consider No. 1178, which is distorted about as much as No. 2 ; but while the latter is deformed on its right side, the former is compressed on its left, and is asymmetrical through its entire extent, including the lower jaw. Viewed from behind it is of almost quiuque-lateral shape, the two upper sides of the pentagon being formed by the parietals sloping down slightly from the coronal suture to the tubers; the two lateral sides are formed partly by the parietals and the posterior portions of the squamosals, sloping from the parietal tubers to the mastoid processes, which are very small and obtuse, while the base of the polygon is formed by a line running obliquely from the left zygomatic process across the occipital to within two centimeters of the right mastoid ; protuberance of occiput scarcely developed ; superior and inferior curved lines wanting; muscular insertions weak.

Owing to the deformation of the skull, the two profile views are rather different from each other. The right side of the cranial portion does not present any extraordinary features, except toward the frontal region and the adjoining facial portion. The left frontal tuber projects about one centimeter above the right one, the whole left half of the frontal being pressed forward, in consequence of which the right side seems to recede to a considerable extent; ridge of the nasals slightly curved inward. There is slight tendency to prognathism. The left profile shows the occipital depressions in almost its whole extent, the lower and posterior portion of the parietal being flat, almost concave instead of convex; this is also the case with the left portion of the occipital. The asymmetry of the skull becomes not less striking in a facial view, in which case the frontal bone recedes in an oblique direction from the left tuber to a short distance behind the right one. The lower jaw partakes, in a similar manner, in the distortion, its left half being also pressed forward, but less than the corresponding half of the frontal bone. The lower border curved inward and upward rather deeply near the symphyses, the curve descending lower on the right side than on the left.

Viewed from above, the skull is of an irregular oval form, the left frontal tuber projecting beyond the general outline, and the posterior portion of the parietal of the same side being compressed so that the right parietal tuber appears to be very prominent.

Unfortunately the base is not complete; but, although the lower and anterior portion of the occipital and the basilar process are missing, there is evidence enough of the general distortion. The left mastoid process is about one centimeter beyond the right one, and although to a less extent, we notice both the palate bone and the palate process of the left upper maxilla to be pressed forward.


	Measurements. c. m. Capacity......................................................................1213 mm. Length........................................................................ 151 Breadth...................................................................... 138 Breadth of frontal............................................................. 110 Height *......................................................................-----Frontal arch.................................................................. 271

* The anterior margin of the foramen magnum being defective, the height could not he ascertained.


	*mm.* Parietal arch.................................................................. 349 Occipital arch................................................................. 195 Longitudinal arch............................................................. 341 Circumference........,............................-............................ 357 Length of frontal.............................................................. 115 Length of parietal ............................................................ 126 Length of occipital........................................................... 153 Zygomatic diameter........................................................... 100

Skull No. 1179 rather delicate; is deformed in a more regular manner than either of the others above mentioned, and might very well be taken for that of an ancient Peruvian. The frontal does not recede as much as it usually does in Peruvian skulls, owing to their mode of deformation, but in this instance the lower anterior portion of the frontal ascends very straight, almost perpendicular, and more so than in any of the other skulls.

Viewed from above, it is broadly heart-shaped, having its greatest breadth about one centimeter above the insertion of the retrahens au-rem in a plane projected perpendicular through the posterior portions of the zygomatic processes. The nasal bones and the alveolar processes of the upper jaw protrude to a considerable extent beyond the general outline of the cranium, while the malars and zygomatic processes stand out but slightly, the former, however, more than the latter. In profile the occipital portion of the skull appears perpendicular within a few degrees; only the left portion of the occipital bone protrudes a little near the superior curved line, which on this side is more developed than on the other. Left zygomatic process slender; the right one defective, partly broken. Nasal bones very projecting, indicating a Roman nose.

Facial portion comparatively small; forehead narrow; frontal tubers well developed; superciliary ridges but slightly indicated. Malars slender; anterior nasal spine prominent; maxillae very nearly orthog-nathic. Teeth very irregular; left incisors of upper jaw missing; the first right one very large, while the second is unnaturally small. Canines standing very oblique, their crowns inclined to a considerable extent toward the symphysis. Second right bicuspid temporary; permanent one in the act of eruption, last molars not yet cut, their position somewhat abnormal. The teeth of the lower jaw are more regular. Both false molars of the left side and second on the right missing, last molar not yet erupted. Mental process well developed; exterior oblique line slightly indicated; sigmoid notch flat; ridges for the insertion of the masseter slight; anterior border of ramus almost straight.

The base of the skull is nearly triangular, and only slightly asymmetrical. Posterior outline of occipital almost straight, and approaching within half a centimeter toward the posterior margin of the foramen magnum. Muscular insertions highly developed; occipital protuberance very much elongated. Condyles long and narrow; their surfaces quite uneven. Only the right posterior condyloid foramen developed and very small; fossa on this side shallow; the one opposite deep.


	Measurements. *c. cm.* Capacity..................................................................... 1380 *mm.* Length...................................................................... 153 Breadth...................................................................... 156 Breadth of frontal........................................................... 128 Height....................................................................... 142 Frontal arch................................................................. 316 Parietal arch................................................................. 370 Occipital arch................................................................ 240


	Longitudinal arch............................................................ 338 Circumference............................................................... 495 Length of frontal.............................................................. 115 Length of parietal............................................................. 130 Length of occipital........................................................... 163 Zygoinatic diameter.......................................................... 123

In order to show the resemblance between the skulls from Southern Colorado and New Mexico, described on the preceding pages, and those of the ancient Peruvians, the diagram, Fig. 14 on Pl. 29, was drawn. Before referring to the latter, we may be allowed to recall the fact that there are two different types of Peruvian skulls, one kind being lengthened similarly to those of the Macrocephali of Hippocrates,* and found chiefly in the chulpas near Lake Titicaca, while those from other localities, and not met with under the "burial towers," are shortened by compression of the occipital region. We hardly need to state that the type alluded to here is the latter one. [2] The skull, whose profile is represented by the less heavy lines, belongs to the collection of the United States Army Medical Museum. It bears the number 250 of the catalogue, and is designated as "Pachacamac skull, from Peru." Viewed from above it is similar in outline to skull 1179, previously described (represented in heavy lines on the diagram). Its frontal bone only proves to be more receding and elongated than that of the latter, it being, besides, a little longer, and in the norma verticalis we perceive the zygomatic bones, the nasals, and the upper maxillae, to project more beyond the general outlines of the cranium than it is the case with No. 1179. It is rather thick and heavy, apparently male, with pronounced muscular insertions. Those of its measurements that may be of interest in our case are as follows :


	*mm.* Length....................................................................... 162 Breadth...................................................................... 156 Breadth of frontal............................................................ 118 Length of frontal............................................................. 119 Circumference................................................................ 505

The dotted profile line of the diagram is that of skull No. 276 of the catalogue above mentioned, and designated as "Cranium of pure Indian. Coban, Guatemala." Viewed from above it appears slightly asymmetrical, the left parietal being compressed at its posterior portion, between the lower third of the sagittal suture, the tuber, and the squamosal. The zygomatic bones and the malars project beyond the general outline of the cranium in this position, and likewise the super-ciliar ridges and the nasals, the greatest breadth being found a little below the parietal tubers. Apparently male; muscular insertions very pronounced ; skull heavy and thick.

As measurements of interest I shall give:


	mm. Length...................................................................... 162 Breadth...................................................................... 142 Breadth of frontal............................................................ 112 Length of frontal............................................................. 122 Circumference.............................................................. 481

*Magni Hippocratis medicorum omnium facile principis opera omnia quse exstant, nunc denuo latina interpretatioue et annotationibus illustrata, Anutio Foesio Medio-inatrico medico authore. Francofurti, MDCXX1V, page 289.

[2] Mariano Eduardo de Rivero y Juan Diego de Tschudi, Antiguedades Peruanas, Vienna, 1851, p. 25, where the type in question is represented, with two others, the authors distinguishing three types on rather trivial grounds. A large view of the type under consideration is also figured in the atlas accompanying the volume, Plate V.

In order not to crowd and confuse the diagram we omitted the outline of skull No. 3. the frontal portion of which (even to the superciliary ridges) would almost overlap that of the Peruvian. It will be seen that the general character of the profile views of the three skulls is not very different, the Abiquiu cranium showing the greatest height, the steepest forehead, and an almost straight occipital outline, while that from Coban is intermediate between the two others. The cranium having the greatest length is that of the Peruvian, viz, 162^mm, differing only l^mm from that of the Coban Indian. In the diagram this difference seems to be greater, as the occipital outline of the Peruvian projects considerably beyond that of the former; but if we come to examine the receding frontal of the latter and remember that the greatest length lies between the glabella and the most prominent point of the occiput, then the features of the diagram appear to be quite normal. A comparison of the breadth of the crania bears out the fact that the measurements of the Peruvian and Abiquiu Indian coincide exactly, they being both 156^mm, while that of the Coban Indian gives 12^mm less. In regard to the breadth of the frontal bone the Abiquiu skull exhibits the maximum, viz, 128^mm; next to this is the Peruvian, namely, 118^mm; and finally the Coban Indian, measuring 112^mm only. The length of the same bone is greatest in the Coban Indian, its measurement giving 122^mm, that of the Peruvian being 3^mm, and that of the skull from Abiquiu 7^mm shorter. The Peruvian skull being the longest among the three, has also the most considerable circumference, namely, 505^mm, while the Abiquiu Indian measures 495^mm around, and the skull from Coban 461^mm only.

After this paper had been placed in the hands of the printer, the United States National Museum at the Smithsonian Institution received a box with relics exhumed from mounds in Tennessee, which happened to be opened in my presence. Besides several cubic feet of bone fragments it contained two crania, showing exactly the same mode of deformation as the skulls obtained in the vicinity of the ruins. And indeed the resemblance between the crania from those two different localities is so great that they might readily be confounded.

It being too late to present cuts of these specimens, I shall limit myself by describing the same, designating one by A and the other by B.

Cranium A.

Very light and smooth, and apparently that of a woman. As the eruption of the last molars had taken place, the individual must have been at least twenty-one years of age when it died; that is, if we are justified in using the same Criterion in judging the age of an Indian as we do in judging that of a white man. Taking into consideration that the skull was taken from a mound, it is in a tolerable state of preservation; the left zygomatic bone and the styloid processes being broken, however, and the upper jaw containing but six teeth. The condition of the lower is less good, there being only three teeth left, the rami being rather detective and the condyloid and coronoid processes wanting.

Viewed from above, the skull is irregularly heart-shaped; the irregularity being produced by the compression of the left parietal. The compression extends from a plane laid horizontally through the tubers, from the right tuber down to the squamous and occipital sutures, so that the left bone and tuberosity seem to protrude considerably. Serration of the sagittal suture strong, the two parietals grooved at their line of junction. The zygomatic bones project but slightly beyond the general

outline of the cranium; the malars, the maxillae and the nasals more so, especially the latter.

The two profiles show somewhat different features. While the outline of the occipital region, when viewed from the right, is almost perpendicular, that of the left slopes at a slight angle, and the left parietal tuber appears more prominent in this view than the right. Zigoma and malar thin; forehead straight; frontal tubers not projecting; angle between frontal and nasals sharp ; anterior nasal spine prominent; upper maxillae but slightly prognathous.

A facial view discloses a rather narrow forehead, frontal tubers, which are very slight, standing close together, no superciliary ridges, supra-orbital ridges very slight, the right one showing two foramina. Orbits rather oval and nose broad. Maxillae smooth; canine eminence very slight. Remaining teeth showing but little wear; the crowns of the two first molars being worn down most.

Viewed from behind, the outline represents an irregular hexagon. Pa-rietals sloping roof-like from the apex to the tubers, from there almost straight to mastoid portion of the squainosal. From the mastoids the slightly curved outlines converge toward the posterior margin of the great foramen. The left portion of the lambdoidal suture contains three Wormian bones, one fourteen millimeters long and about five wide, being situated in its upper third, while the two others of smaller dimensions, at least in regard to length, are close together, almost touching the lower extremity of the suture. Superior curved line well pronounced, the same being the case in regard to the inferior. Muscular insertions of the left side more marked than those of the right, especially the depression for the rectus capitis. Mastoid processes obtuse.

In a base view the compression of the left side of the skull shows hardly as much as from above, and the asymmetry does not extend over the whole base as it did in several of the other instances.

In taking the dimensions, the zygomatic diameter could not be ascertained very accurately; the probable error of measurement is, however, supposed to be not more than ± l^mm.5. As the skulls are not yet cleaned, the internal capacity was not measured.


	Measurements. mm. Length..................................................................... 156 Breadth..........................................___....................... 141 Breadth of frontal........................................................... 109 Height...................................................................... 142 Frontal arch................................................................ 277 Parietal arch............................................................... 330 Occipital arch............................................................... 227 Longitudinal arch........................................................... 335 Circumference............................................................... 470 Length of frontal........................................................... 115 Length of parietal........................................................... 128 Length of occipital.......................................................... 160 Zigomatic diameter......................................................... 124

Cranium B.

The whole facial portion being broken off, only the cranial portion of the skull is left, and this is rather defective, as the left squamosal and the anterior part of the occipital are wanting. Cranium very light, showing smooth surfaces; age doubtful, but the individual was evidently younger than that from which skull A was derived.

Viewed from above, it appears like a rounded oval, occipital region flattened, right side more than left, asymmetry slightly less than in other specimens. Parietal tubers prominent. Serration of sutures fine.

The right profile, which is more complete than the left one, shows an almost perpendicular but slightly outward curved outline of the occipital region, while in the left view the occipital is seen to bulge out to considerable extent. Temporal ridge narrow and sharp. Mastoid process narrow and small.

A front view shows nothing but the narrow and straight forehead, with tolerably developed tubera, and very slight superciliary ridges.

Viewed from behind, the roof of the skull appears rather flat, with only a slight curvature of the parietals from the sagittal suture to the tubera, while the rest of the outline is similar to that of skull A. Serration of lambdoidal suture complicated. Muscular insertions of the lower portion of the occipital tolerably developed.

Owing to the defective condition of the skull only a limited number of measurements could be obtained. In taking those of the frontal and parietal arches and of the circumference only one-half of the actual distance was measured, and in consequence of this the numbers thus derived had to be multiplied by 2 in order to get the whole dimensions.


	Measurements. mm. Length___.................................................................. 153 Breadth..................................................................... 140 Breadth of frontal......................,............-........................ 118 Frontal arch................................................................. 280 Parietal arch................................................................. 222 Occipital arch................................................................ 212 Circumference............................................................... 460 Length of frontal............................................................. 129 Length of parietal............................................................ 131

GENERAL CONSIDERATIONS.

Disregarding the two skulls obtained from the Tennessee mounds, we notice, as characteristics of the others, the great height and breadth of the cranium, which is most decidedly brachycephalic. As the skulls are deformed we refrain from calculating the indices of height and breadth.

Denoting the excess of the longitudinal diameter as compared with the transverse one as positive, and that of the transverse diameter as compared with the longitudinal one as negative, we obtain the following values:


	Designation of skull. L. B. No. 1.................................................................... + 24^mm. No. 3.................................................................... +19 No. 1778.................................................................. +13 No. 1779................................................................. + 3

It will readily be seen that the adult Abiquiu skull (1779) is more broad than long, while the others are somewhat longer than they are broad. If we were to calculate the index of breadth we would notice it to decrease in the order adopted in the table, from skull No. 1779 upward. How far the brachycephalic character is due to deformation cannot be decided by means of the limited number of specimens on hand, but according to the general features of the skulls, we consider ourselves justified in saying that originally they were brachycephalic, and that in consequence of the deformation of the occiput the brachycephalic character is brought to view more strikingly than it would be if the skulls had not been

compressed. We do not hesitate to express this opinion, although it appears from Welcker's extensive investigations* that brachycephalic skulls are mostly orthognathous, while those described above are all more or less prognathous.

Now the question arises, was the deformation to be found among the ancient inhabitants of the ruins generally, or only in some instances ? Was it practiced as a rule, or is the compression only more or less accidental ? As the five skulls before us show the most unmistakable signs of deformation, it is more than probable that the compression of the skull cannot have been of rare occurrence, for it is scarcely possible that, by a mere accident only, the deformed crania should have been found by the collectors, while others, showing normal features, should have escaped their observation.

If deformations of this kind are practiced, as a rule, among savages or civilized people, there is always a guiding idea, [2] either emanating from .aesthetical feelings or with reference to some practical purpose. For instance, among the ancient Peruvians the flattened skulls were considered a sign of aristocracy and high breeding, while in other cases, as among the inhabitants of the northwestern corner of Europe, [3] the skulls of the children were deformed in early youth in such a manner as to produce a more or less straight line of profile, which we are accustomed to admire in the Greek ideal of human beauty.

That the idea of the beautiful was developed to an uncommon extent among the inhabitants of the ruins is borne out by the fact that the fragments of pottery found show simple but tasteful ornamentation, that great regularity prevails in the structure of their buildings, both in the square and round ones, and that other pieces of their handiwork would even gratify our own aesthetical feelings. As symmetry is found everywhere among their buildings, even where an educated eye would require it, there is no reason whatever to suppose that the taste should comply with these requirements in one instance and not in the other. Why should people who bestow so much care upon a door, in order to make it symmetrical, or upon a piece of pottery, to make it perfect in shape, try, in case they practice deformation of the skull, to produce an asymmetrical form, such as is found in the greater number of our crania ?[4] Wherever deformation of the skull is made a practice, the mother, or whoever may be in charge of the child, performs this operation with the utmost care, in order to produce the conventional shape, which is almost symmetrical in every instance; and some time ago, when we had occasion to examine about 200 flatheads, we found but few that were not symmetrical.

Under such circumstances we are justified in believing that the deformation of the skulls in question is more or less accidental, and, taking the mode of depression into consideration, there cannot remain any doubt in regard to the way the skulls were distorted. Evidently the ancient inhabitants of the ruins were in the habit of strapping their children against cradle-boards, as a great many modern Indians do, and hence resulted the flattening of the occiput.||

* Untersuchungen ueber Wachstbuin und Bau des menschlichen Schiidels von Hermann Welcker. Erster Theil. Leipzig, Wilhelm Engelmaun, 1862, pp. 63 and 90.

[2] Andrese Vesalii Opera Ornnia Anatomica et Chirurgica, cura Herrnanni Boerhaave et Bernhardii Siegfried Albini. Lugdnni Batavorum, MDCCXXV, Tom. I, p. 16.

[3] Hippocrates: loc. cit., De aere aquis et loeis, p. 289.

[4] As it seems, asymmetry is not met with very seldom among the ancient Peruvian skulls. Compare Morton's Crania Americana, PL X.

[5] As Major Powell kindly informs me, the different Indian tribes he is familiar with keep their children strapped to the cradle-board from between one and a half to two years.

With reference to this question, we examined a considerable number of Indian skulls contained in the collection of the United States Army Medical Museum, but we must confess that among the modern ones we did not meet with any deformation that could reasonably be attributed to the influence alluded to. Among a series of skulls exhumed from mounds in Kentucky, No. 723 is deformed very slightly, while No. 730 shows a pronounced depression of the lower and posterior portion of the left parietal bone, and in No. 242 we meet with an almost perpendicular occiput, perceptible especially when viewed from the right side. A considerable number of crania from mounds near Yicksburgh, Miss., could not be used to strengthen our argument, as they are all, with scarcely any exception, deformed intentionally and in every possible direction.

Having obtained these rather negative results, we examined the collection of Indian cradles contained in the National Museum, trying to establish some relation between the construction of the cradle-board and the mode of deformation of the skull. If we were to judge from the specimens as they present themselves at the Smithsonian Institution, we might under the circumstances arrive at erroneous conclusions, as most likely the majority of these cradles are incomplete, consisting in most cases of scarcely anything but the frame-work. Probably, some tribes prepare the head-rests for the children more carefully than others, and the deformation of the skull will be found accordingly.

Unfortunately, there are no skulls of Indian tribes that now inhabit the vicinity of the region in question within our reach, and therefore we are unable to draw any comparisons or to point out cranial affinities that might exist between the ancient inhabitants of the ruins and the people that now live near those deserted dwellings.

According to other evidence, however, there is not much room left to doubt that the present Pueblo Indians are the direct descendants of the ancient inhabitants of Southern Colorado and New Mexico, although there are either no traditions preserved pointing toward this direction, or the few that do exist are of too vague a nature to be relied upon. As one of the arguments in favor of this view, we may take the mode of constructing the houses. The structures erected by the present Pueblos do not differ materially from the majority of dwellings built by the ancient architects. The whole modifications the former offer are merely due to the different topographical features of the ground on which they stand, and to the influence of the white race, as can readily be proven. In the present mode of constructing the dwellings, there is, however, one detail worthy of attention and further investigation. We mean the arched building,* as there are but two tribes inhabiting this continent whose architectural skill proved efficient enough for this purpose, namely, the Peruvians and the Eskimos. As the dome-like structure is only found among this race (besides the people just mentioned), it would be worth while to investigate whether this style of architecture is original or adopted from the Spaniards, and whether it is of rare occurrence or to be met with frequently.

As another evidence in favor of the view that the present Pueblos are the descendants, or rather the remnant, of the ancient cliff-house builders, we may take the pottery found near the ruins. The shape of the pots now in use among the Pueblos is the same as that of the few pieces brought east; so are the ladles with their characteristic hollow handles, and the ancients were in the habit of burying their water-jugs in the ground in order to keep them cool, just as we see it practiced

* " Descending to the chamber by a ladder, it was found probably 25 or 30 feet in diameter, arched above, and about 20 feet high."—Report of Chief of Engineers, p. 1067.

now. If we were to take the pattern of the primitive ornaments as a criterion—which is, however, not permissible, as certain figures are found in all parts of our planet, wherever people are far enough advanced to invent and to use ornaments—we would meet with the same designs in both instances, only that in the majority of cases the ancient ornamentation is more careful, and exhibits besides more artistic feeling on the part of the workmen than that of the modern.

The most striking peculiarity of a large number of fragments of pottery found near the ruins, and hitherto never noticed in America, is the existence of a regular glazing of great hardness. How this glazing was produced cannot be ascertained without considerable difficulty. In some instances it is rather opaque, whitish, and covers the surface of the fragment in a thick layer, while in others it is perfectly transparent or shows a slight bluish hue. The colors used in painting the pottery are the same as now, namely, a reddish-brown, brown, and black ; the former being produced by iron and manganese, the latter consisting of an organic substance, the character of which could not be determined. Until we shall have some accurate analysis, which will be rather difficult to make, we may perhaps call this organic substance charcoal, mixed intimately with fine clay before it was used for painting, and prevented by the siliceous cover from being destroyed by the heat while the vessel was being baked. A closer examination of the different patterns demonstrates that the painting was done in a very careful, in some instances hesitating, manner, and it seems that a great deal of time was devoted to this process. There are several fragments on which we may easily recognize the method followed by the artist, and notice how he drew his outlines and how he afterward painted the rest of the square or triangle.

The facts bearing upon the probable age of the human remains are rather contradictory, and it is difficult to make them agree and to draw some satisfactory conclusions. Evidently, the buildings near which they were found represent different ages of perhaps considerable periods of time. How far back the earliest ones may date can scarcely be surmised, but there are some instances, resting however on rather unsatisfactory bases, that might give us some clew as to the approximate time the buildings were abandoned.

Mr. Jackson states that he noticed several places that appeared to have been used as corrals, and from Dr. Endlich I learn that he discovered a stratum ten feet below the surface of the soil, containing, besides fragments of pottery, numerous bones of sheep. Now, nothing would be more tempting than to combine these two facts, and to conclude that the people in question domesticated sheep, which could not have been done during the pre-Columbian period, as this animal does not belong to the American fauna.

Among a collection of arrow-heads and minor stone chips I noticed a glass bead, which is, beyond doubt, of Venetian workmanship, and very similar to beads exhumed lately from the grave-mounds at Santa Barbara, Lower California, and preserved at the National Museum. If the bead in question did not find its way accidentally to the ruins through some Indian of a later period, we might consider it as a proof in favor of our view ; but at the same time we must not forget that no trace of any metal was found in the graves, as we might have reasonably expected, since the cliff-house builders communicated directly or indirectly with the inhabitants of the Pacific coast. This latter fact is supported by the presence of a considerable quantity of shells of Olivella gracilis

used as beads, and by a large spine-fragment of a recent Echinus, which I noticed among some pieces collected by Mr. Jackson.

It would exceed the limits of these pages, to the preparation of which we could devote but a few days, to make any further remarks which would so readily suggest themselves. Evidently, the ruins in question bear testimony of one of the earliest centers of civilization in this country, and would well repay the investigations of a critical anthropologist. Let us hope that ere long such investigation may be undertaken.

N. B.—All the drawings contained in the following plates are orthographic figures, made by means of a diopter, and reduced with a pantograph from natural size.

PLATE 23.

Fig. 1.—Lateral view of skull No. 1. Fig. 2.—Frontal view of the same.

PLATE 24.

Fig. 3.—Occipital view of skull No. 1. Fig. 4.—Lateral view of the same.

PLATE 25.

Fig. 5.—Basal view of skull No. 1. Fig. 6.—Lateral view of skull No. 2.

PLATE 26.

Fig. 7.—Frontal view of skull No. 2. Fig. 8.—Occipital view of the same.

PLATE 27.

Fig. 9.—Vertical view of skull No. 2.

Fig. 10.—Basal view of the same. Owing to the faulty shading of the occipital region, the occipital bone does not appear to turn upward as abruptly from the posterior margin of the great foramen as it actually does.

PLATE 28.

Fig. 11.—Frontal view of skull No. 3. Fig. 12.—Vertical view of the same.

PLATE 29.

Fig. 13.—Lateral view of skull No. 3.

Fig. 14.—The heaviest outline of the diagram represents a profile-view of skull 1179 (Catalogue U. S. A. Med. Museum) from Abiquiu. The less heavy line represents the same view of a Guatemala Indian, No. 276 of the catalogue, and referred to on page 56, while the dotted line represents the profile of a Peruvian, No. 250 of the catalogue, and mentioned on the same page.



	DEPARTMENT OF THE INTERIOR. UNITED STATES GEOLOGICAL AND GEOGRAPHICAL SURVEY OF THE TERRITORIES. F. V. HAYDEN, U. S. Geologist-in-Charge. I.-A NOTICE OF THE ANCIENT REMAINS OF SOUTHWESTERN COLORADO EXAMINED DURING THE SUMMER OF 1875. By W. H. HOLMES. II.—A NOTICE OF THE ANCIENT RUINS IN ARIZONA AND UTAH LYING ABOUT THE RIO SAN JUAN. By W. H. JACKSON. III.—THE HUMAN REMAINS FOUND NEAR THE ANCIENT RUINS OF SOUTHWESTERN COLORADO AND NEW MEXICO. By Dr. EMIL BESSELS. EXTRACTED FROM BULLETIN OF THE GEOLOGICAL AND GEOGRAPHICAL SURVEY OF THE TERRITORIES, Vol. II, No. I. Washington, March 21,1876.



	4 what extent wood and adobe were used can hardly be determined. It is evident, however, that the greater portion of the villages and dwellings of the lowlands have been of material other than stone, frequently doubtless of rubble and adobe combined. As to situation, they may be classified very properly under three heads: (1) lowland or agricultural settlements; (2) cave-dwellings ; and (3) cliff-houses or fortresses. Those of the first class are chiefly on the river-bottoms, in close proximity to water, in the very midst of the most fertile lands, and located without reference to security or means of defense. Those of the second are in the vicinity of agricultural lands, but built in excavations in low-bluff faces of the Middle Cretaceous shales. The sites are chosen also, I imagine, with reference to security; while the situation of the cliff-houses is chosen totally with reference to security and defense, built high up in the steep and inaccessible cliffs, and having the least possible degree of convenience to field or water. As to use, the position for the most part determines that. The lowland ruins are the remains of agricultural settlements, built and occupied much as similar villages and dwellings would be occupied by peaceable and unmolested peoples of today. The cave-dwellers, although they may have been of the same tribe and contemporaneous, probably built with reference to their peaceable occupations as well as to defense, and it is impossible to say whether or not they made these houses their constant dwelling-places. The cliff-houses could only have been used as places of refuge and defense. During seasons of invasion and war, families were probably sent to them for security, while the warriors defended their property or went forth to battle; and one can readily imagine that when the hour of total defeat came, they served as a last resort for a desperate and disheartened people. In form, the parallelogram and circle predominate, and a considerable degree of architectural skill is displayed. Where the conformation of the ground permits, the squares are perfect squares and the circles perfect circles. A greater part of the ordinary structures are square or rectangular; while attached to each group, and sometimes without indications of contiguous buildings, are the circular ruins frequently resembling towers. These are the most pretentious structures, being often as much as forty feet in diameter, and in many cases having double or triple walls. They are solidly built of hewn stone, dressed on the outside to the curve, neatly jointed, and laid in mortar. The space between the outer walls is invariably divided by heavy partition-walls into a number of apartments, while a circular depression, or estufa,* occupies the center of the inclosure. It seems evident, from the extraordinary form of these structures and the unusual care shown in their construction, that they were not designed for the ordinary uses of dwelling or defense. It has been observed that, among nearly all the ancient tribes of North America, the grandest and most elaborate works of art were the offspring of their superstitions, and it does not seem at all improbable that these great towers had a religious origin. It is stated that the eternal fire—an essential of their worship—has always been kept in circular inclosures, and that the circle symbolizes the sun, their deity. The occurrence, therefore, of one or more of these circular inclosures in each of their settlements can be rationally accounted for; but it is with less certainty we arrive at conclusions in regard to the triple walls and the cell-like apartments. In the inhabited -------------------------------------------------------------------------------- * A Spanish word signifying "sweat-house" or council-house.



	5 pueblos of to-day, there are underground rooms, mostly circular, used as council-chambers as well as for the performance of the mysterious rites of their religion. Similar chambers occur, according to Lieutenant Simpson, in all the ruined cities of New Mexico, but having single walls of no great height or thickness. It is stated by Squier and Davis* that in Mexico the sacred inclosures were also used for defensive purposes, and it certainly seems probable that these curious structures served the double purpose of temples and fortifications, and that the apartments between the walls were the cells of the priesthood or the receptacles of sacred or valuable property. The smaller single-walled towers, which are scattered at intervals along the river-courses and canons, frequently in commanding situations, were probably watch or signal towers. The cave-dwellings are made by digging irregular cavities in the faces of bluffs and cliffs formed of friable rock, and then walling up the front, leaving only a small doorway for entrance and an occasional small window at the side or top. The cliff-houses conform in shape to the floor of the niche or shelf on which they are built. They are of firm, neat masonry, and the manner in which they are attached or cemented to the cliffs is simply marvelous. Their construction has cost a great deal of labor, the rock and mortar of which they are built having been brought for hundreds of feet up the most precipitous places. They have a much more modern look than the valley and cave remains, and are probably in general more recent, belonging rather to the close than to the earlier parts of a long period of occupation. Their position, however, has secured them in a great measure from the hand of the invader as well as from the ordinary effects of age. Of works of art other than architectural that might assist in throwing light upon the grade of civilization reached by these people, but meager discoveries were made; although I imagine that careful search and well-conducted exhumation might develop many things of great interest. A small number of arrow-heads, stone-implements, ornaments, and articles of fictile manufacture, that may fairly be attributed to the age of the cliff-builders, were collected. The greater part of these are figured in plates XIII and XIV. There are no evidences whatever that metals were used. Numerous hieroglyphics were observed, both engraved and painted upon the cliffs. Drawings of a large number were made, and some of the more notable examples are given in plates XI and XII. A great number of burial-places were noted, but of the graves examined few yielded further evidences of occupation than small quantities of charcoal and bits of painted pottery. These burial-places, which are in a number of cases covered by a heavy growth of full-grown pinon pines and cedars, are usually found on the summits of high ridges and promontories, and are still marked by slabs of sandstone set on edge and arranged in circles, and parallelograms of greatly varying dimensions. But that they did not always bury their dead in high places is proven by the frequent discovery of human remains in the arroyos or deep washes in the valleys. Three skeletons were obtained, in the vicinity of ruined villages, from the sides of recent washes. The accompanying plates are, with one exception, reproductions of pen-drawings, and are arranged for convenience, rather than from any method of classification. The plans are not drawn to a uniform scale, because of the inconvenience of such an arrangement ; but measure ------------------------------------------------------------------------------- * Ancient Monuments of the Mississippi Valley, page 102.



	6 merits are so frequently given on the plates themselves that no confusion need occur. Measurements were taken by tape-line in all the more important structures; but in many of the ordinary ruins, where exact dimensions were not considered essential, the distances were estimated. It is to be greatly regretted that extreme haste frequently prevented close and accurate work. The accompanying map will give the location of all the more important-groups of ruins. The fragments of information collected in this notice are given with the hope that they may throw a little additional light upon the very interesting problems of ancient American history. RUINED VILLAGE ON THE RIO LA PLATA. Plate I. The first group of ruins observed is situated on the Rio La Plata, about twenty-five miles above its junction with the San Juan, and three or four miles south of the New Mexican line. It is doubtless the remains of a large, irregular village, and stands on a low terrace, some 20 feet above the river-bed, and near the center of a large, fertile valley. It will be seen, by reference to the plate, which includes only the more important part of the town, that the buildings have been isolated, and, in a measure, independent of each other, differing in this respect from most of the groups of ruins farther south. The forms are chiefly rectangles and circles; one or two seem to have been elliptical, while a number have consisted of irregular groups or clusters of apartments. All that now remains to mark the site of these ancient structures is the low, rounded heaps and lines of debris, composed of earth, water-worn pebbles, and small fragments of sandstone. The walls of four of the main structures are quite distinctly marked. That of the circle c is still 4 feet high on the outside, and incloses a depression, probably an estufa, which, in the center, is 2 or 3 feet below the terrace-level. North of this, about 300 feet, is a truncated rectangular mound, 9 or 10 feet in height and 50 feet in width by 80 in length. On the east end, near one of the angles, is a low, projecting pile of debris that may have been a tower. There is nothing whatever to indicate the use of this structure. Its flat top and height give it more the appearance of one of the sacrificial mounds of the Ohio Yalley than any other observed in this part of the West. " It may have been, however, only a raised foundation, designed to support a superstructure of wood or adobe. North of this again, and 100 feet distant, is a rectangular inclosure about 60 by 100 feet. It is slightly excavated in the center, and the rounded and irregular wall is from 4 to 6 feet in height. The space between this and the last-mentioned structure is filled in to the depth of 2 or 3 feet, and the amount of debris about their bases indicates original walls of considerable height. North of this are scattered a number of inferior ruins, the walls of which are not always distinctly marked. These extend back toward a row of low hills, the remnants of a superior terrace, on the summits of which a number of artificial depressions were found. Such "dug holes ^v are generally quite numerous in the vicinity of these ruins, and have doubtless in many cases been made by throwing up earthworks for defensive purposes. South of the large circle is a mass of ruins covering some 15,000 square feet, but so much reduced that nothing further could be determined than the fact that it had contained a large number of irregular apartments. Next to this is a rect-



	7 angular ruin, containing three well-marked apartments. Its walls are 6 or 7 feet high, and, unlike those of the preceding examples, do not coincide with the cardinal points. South of this, and occupying the extreme southern end of the terrace, are a number of small circles and mounds, while an undetermined number of diminutive mounds are distributed among the other ruins. To the east of the Indian trail, as shown in the plate, are a number of inclosures of lesser importance, which, from want of time, were not closely examined. Nowhere about these ruins are there any considerable indications of defensive works, and the village, which is scattered over an area fully two miles in circuit, has no natural advantages whatever. Neither are there traces of ditches, nor of anything that might throw important light upon the habits or occupations of the people. A few arrow-heads and minute cutting-implements were picked up. Countless chips of jasper, obsidian, and flint were scattered around, and the soil was literally full of fragments of painted and ornamental pottery. On the opposite side of the river, and at intervals above and below, are isolated groups of ruins and heaps of debris—certainly the remains of dwellings. These seem to have been distributed very much as dwelling-houses would be in the rural districts of civilized and peaceable communities. It is possible that there are undiscovered ruins on this stream equally important with that described; for, in pursuing my geologic investigations, I was compelled to take a long detour to the westward from this point, returning to the La Plata again a few miles above its junction with the San Juan. On this occasion, while riding through a desert-like locality, quite naked and barren, much resembling the well-known Mauvaises terres, I was surprised to observe fragments of pottery strewn around, and presently a number of ruins, in a very reduced state and almost covered by the drifting sand, and this six or eight miles from water. On the high, dry table-lands, on all sides, fragments of pottery were picked up. What could have induced people to build and dwell in such a locality it is useless to surmise. GROUP OF CAVE DWELLINGS AND TOWERS ON THE RIO SAN JUAN. Plate II. On the San Juan River, about thirty-five miles below the mouth of the La Plata and ten miles above the Mancos, occurs the group of ruins figured in Plate II. The river is bordered here by low lines of bluffs formed from the more compact portions of the Middle Cretaceous shales. At this particular place, the vertical-bluff face is from 35 to 40 feet in height. I observed, in approaching from above, that a ruined tower stood near the brink of the cliff, at a point where it curves outward toward the river, and in studying it with my glass detected a number of cave-like openings in the cliff-face about half-way up. On examination, I found them to have been shaped by the hand of man, but so weathered out and changed by the slow process of atmospheric erosion that the evidences of art were almost obliterated. The openings are arched irregularly above, and generally quite shallow, being governed very much in contour and depth by the quality of the rock. The work of excavation has not been an extremely great



	8 one, even with the imperfect implements that must have been used, as the shale is for the most part soft and friable. A hard stratum served as a floor, and projecting in many places made a narrow platform by which the inhabitants were enabled to pass along from one house to another. Small fragments of mortar still adhered to the firmer parts of the walls, from which it is inferred that they were at one time plastered. It is also extremely probable that they were walled up in front and furnished with doors and windows, yet no fragment of wall has been preserved. Indeed, so great has been the erosion that many of the caves have been almost obliterated, and are now not deep enough to give shelter to a bird or bat. This circumstance should be considered in reference to its bearing upon the question of antiquity. If we suppose the recess to be destroyed is six feet deep, the entire cliff must recede that number of feet in order to accomplish it. If the rock were all of the friable quality of the middle part, this would indeed be the matter of a very few decades; but it should be remembered that the upper third of the cliff-face is composed of beds of comparatively hard rocks, sandstones, and indurated shales. It should also be noted still further that at the base of the cliff there is an almost total absence of debris, or fallen rock, or even of an ordinary talus of earth, so that the period that has elapsed since these houses were deserted must equal the time taken to undermine and break down the 6 feet of solid rock, plus the time required to reduce this mass of rock to dust; considering also that the erosive agents are here unusually weak, the resulting period would certainly not be inconsiderable. Figure 2 gives a fair representation of the present appearance of these dwellings, while their relations to the group of ruins above will be understood by reference to figure 1. These ruins are three in number— one rectangular and two circular. The rectangular one, as indicated in the plan 0, is placed on the edge of the mesa, over the more northern group of cave-dwellings; it is not of great importance, being only 34 by 40 feet, and scarcely 2 feet high; the walls are 1 1/2 feet thick, and built of stone. The small tower B is situated on the brink of the cliff, directly above one of the principal groups of cave-houses. It is neatly built of stone, which, although not hewn, is so carefully chosen and adjusted to the curve that the wall is quite regular. That the stone was procured from the neighboring cliffs is indicated by the presence of great numbers of characteristic fossils. The wall is 18 inches thick and from 2 to 6 feet in height. Long lines of debris, radiating from all sides, indicate that it has been much higher, and has but recently fallen. This tower is inclosed by a wall, also circular in form, but open toward the cliff, as seen in the drawing; the ends projecting forward and irregular and broken as if portions had fallen. Its construction is like that of the inner wall, but the height is not more than 3 feet at any point. The diameter of the inner circle is 12 feet, that of the outer 22 feet; the distance, therefore, between the walls is a little less than 4 feet. In this space, there are indications of partition-walls that have originally divided it into a number of apartments. About one hundred and fifty yards to the southwest of this ruin are the remains of another similar structure. It has been, however, on a much grander scale. The walls are 26 inches thick, and indicate a diameter in the outer wall of about 140 feet. They are not above 4 feet high at any point, and in the parts toward the cliff can only be traced



	9 by a low ridge of earth. The remaining fragments of wall are at the remoter parts of the circles, and are in every respect like the walls already described. The inner wall, which can be traced but a short distance, is 8 1/2 feet from the outer, and has been connected by partition-walls, as in the other case. The first impression given by this curious inclosure is that it was designed for a "corral," and used for the protection of herds of domestic animals; but since these people are not known to have possessed domestic animals, and when we further consider that inclosures of pickets would have served this purpose as well as such a massive and extraordinary structure, we can hardly avoid assigning it to some other use, which use, doubtless similar to that of the smaller tower, is very naturally suggested by its location and construction. That they both belonged to the community of cave-dwellers, and served as their fortresses, council-chambers, and places of worship, would seem to be natural and reasonable inferences. Being on the border of a low mesa country that rises toward the north, the strong outside walls were doubtless found necessary to prevent incursions from that direction, while the little community by means of ladders would be free to pass from dwelling to temple and fortress without danger of molestation. The original height of these structures must necessarily be a matter of conjecture, and it is true that although there is every evidence of age, both in the cave-dwellings and in the walled inclosures above, the lack of great quantities of crumbling walls and debris, and the general bareness of the ruins, give rise to the notion that they were but meager affairs. If we conclude, however, that the outer walls were constructed for defense, and their thickness and form favor such a hypothesis, their height would probably have been as great as fifteen or twenty feet, while the inner walls, being equally heavy and well built, would be sufficiently high to accommodate two or three stories. With these conclusions in view, I have ventured to present a sketch showing a restoration of the smaller tower (plate III). This sketch illustrates the probable appearance of the dwellings and tower, and the supposed means of communication between them. The manner of walling up the fronts of the cave-dwellings, as here given, was observed frequently on the Rio Mancos, where, in corresponding cliffs of shaly sandstones, there are many well-preserved specimens. A large group situated on this stream, about ten miles above its mouth, was subsequently examined. The walls were in many places quite well preserved and new-looking, while all about, high and low, were others in all stages of decay. In one place in particular, a picturesque outstanding promontory has been full of dwellings, literally honeycombed by this earth-burrowing race, and as one from below views the ragged, window-pierced crags, he is unconsciously led to wonder if they are not the ruins of some ancient castle, behind whose rnoldering walls are hidden the dread secrets of a long-forgotten people; but a nearer approach quickly dispels such fancies, for the windows prove to be only the doorways to shallow and irregular apartments, hardly sufficiently commodious for a race of pigmies. Neither the outer openings nor the apertures that communicate between the caves are large enough to allow a person of large stature to pass, and one is led to suspect that these nests were not the dwellings proper of these people, but occasional resorts for women and children, and that the somewhat extensive ruins in the valley below were their ordinary dwelling-places. On the brink of the promontory above stands the ruin of a tower, still twelve feet high, and similar in most respects to those already described. These



	10 round towers are very numerous in the valley of the Mancos. From this point alone at least three others are in view, some on the higher promontories, others quite low, within twenty or thirty feet of the riverbed. I visited and measured seven along the lower fifteen miles of the course of this stream. In dimensions they range from ten to sixteen feet in diameter and from five to fifteen feet in height, while the walls are from one to two feet in thickness. They are in nearly every case connected with other structures, mostly rectangular in form. At the mouth of the Mancos, however, a double circle occurs, the smaller one having been the tower proper. It is fifteen feet in diameter, and from eight to ten in height. The larger circular wall is forty feet in diameter and from two to four feet high, and is built tangent to the smaller. This ruin is at the point where the Mancos reaches the alluvial bottom bordering the Rio San Juan, and about one mile above its junction with that river. On the opposite or south side of the river are traces of somewhat extensive ruins, but so indistinct that the character of the original structures cannot be made out, and indeed no single mile of the lower fifty of the Man cos is without such remains. CANYON OF THE RIO MANGOS. Fifteen miles from its junction with the San Juan, this stream emerges from the southwest border of the Mesa Verde, through which it has cut its way, producing a most remarkable canyon. This canon has already been once or twice described; but in order to make my subsequent descriptions better understood, I shall give here an outline. The Mesa Verde is a somewhat irregular table-land, comprising an area of about seven hundred square miles, and is formed of a great series of nearly horizontal sedimentary rocks, of which the surrounding country has been denuded. This series of strata consists, in the upper part, of massive sandstones; in the middle part, of alternating sandstones and shales; and in the lower one thousand feet, chiefly of shales and clay. These softer beds are, when once exposed to the erosive agents, carried away with great rapidity, and, as a consequence, the firmer rocks above are undermined and break down in vertical cliffs, and, where soft and hard beds alternate, a series of steps, with intervening slopes, is formed. It will readily be seen that a canyon thus formed would consist in general of a narrow, irregular river-bottom, long steep slopes of debris rising like the arms of a letter V from this, then a succession of steeps and slopes, culminating above in a series of lofty, embattled cliffs. The canyon is nearly thirty miles in length, and ranges from one to two thousand feet in depth. It seems to have been a favorite resort of the cliff-building people, and traces of their industry may be found everywhere, along the bottoms, in the cliffs, and on the high, dry tablelands above. The four following plates will be taken up in the delineation of the more interesting portions of these remains. PLATE IV. Figure 1 illustrates the method of walling up the cave-fronts as described on a preceding page. This sketch was made at the last-mentioned locality on the Rio Man cos. The group occurred in the cliff about thirty feet from the base. The three doorways opened into as many small apartments, and these were connected with each other by very small passage-ways. The farther door could not be reached from the outside, as the platform of rock had broken away.



	11 Figure 2 gives a plan of the double tower near the mouth of the Man-cos; it has already been described. The ruin, of which a plan is given in figure 3, occurs on the left bank of the Mancos about eight miles above the foot of the canyon. It is one of the best preserved specimens of the ruined towers, and seems to have been built with much skill. It is 9 feet in diameter on the inside and about 16 feet high. There are three rectangular apartments attached, the walls of which are almost leveled with the ground. In the side of the tower facing the river is a window, about 8 feet from the ground and 2 feet high by 1 1/2 wide. I had been previously led to the conclusion that these towers were in all cases built without windows or openings of any kind within reach of the ground from without, and it is not improbable that this opening did not communicate with the outside, but served as a doorway between the tower and one of the adjoining apartments. The advantage of such an arrangement in a defensive work, such as we may suppose this to have been, is clearly apparent, and evinces not a little intelligence and forethought on the part of the builders. Being built in connection with dwellings and places of resort, they could, in case of alarm, be reached with ease from within, but be altogether secure from without. This ruin was visited by Mr. Jackson last year, and a sketch of it has been published. Figure 4. The large circular ruin, of which a ground-plan is given in this plate, was also visited by Mr. Jackson, photographs were made, and a brief description given; but I deem it best to give a more detailed description, the result of such observations and measurements as could be made in a period of time entirely too short for a work of such importance. This ruin is situated on a narrow strip of alluvial bottom about midway in the canyon of the Mancos. On first approaching it, one does not observe that it differs greatly from the ordinary fragmentary structures below, as it is much decayed and almost hidden by artemisia and vines. Closer inspection, however, develops the greater part of the outline, and I imagine that a little excavation would bring all the foundations to light. The inner wall can be traced throughout the entire circle, and is in places 6 or 8 feet high. A portion of the outer wall, at the point farthest from the river, is still 12 feet in height and in a fair state of preservation. The space between the walls has been divided into cells_ras in the two examples given in plate II. Four of the cross walls are still a number of feet high, while others can be traced by lines of debris. The diameter of the outer wall is 43 feet; that of the inner, 25 feet. They are faced up with larger stones than usual (the heaviest of which, however, could be lifted with ease by a single workman), and have been filled in with rubble, adobe, and wood. The outside courses have been dressed to the curve, and the implements used, judging from the appearance of the picked surfaces, have been of stone. The main wails are 21 inches in thickness, while the partition walls are somewhat lighter, and seem to have been but slightly built into the circular walls. In order to determine the probable number of these cells, I measured the two having complete walls, and found the inner side of each to be 8 feet. As these were both on one side of the circle, I had but to measure the remaining space to complete the semicircle, and on so doing found that there was just room for three additional cells and the necessary partition-walls; two of these were still traceable. To complete the circle, therefore, ten apartments would be necessary. Being desirous of confirming this conclusion, I took the diameter of the inner circle



	12 as given in my notes, and, by adding twice the thickness of the wall, obtained a circumference of 89 1/2 feet; just sufficient space to accommodate ten apartments, with an equal number of partition-walls a fraction less than 12 inches in thickness. By adding to the diameter of the inner circle the total thickness of the walls, plus twice the distance between them, I obtained a diameter of 43 feet for the outer circle. The circuit of the structure is, therefore, 135 feet. Although these figures are greater than those previously given (estimated), I am confident that they cannot vary greatly from the truth. There were no indications of windows or doors in the fragment of outer wall, but two nearly rectangular openings in the inner wall seem to have served as doorways between the central inclosure and the cells. We may suppose that each cell had similar means of communication with the interior. The one doorway that remains entire is 6 feet from the ground, and measures 2 feet in width by 3 in height. The stone-work of the facing is very neat and exact, and the lintel is of a single slab of sandstone. It may be fairly presumed that the outer wall had no doorways or windows within reach of the ground, and that entrance was obtained by means of ladders, through high windows or by way of the roof. The central inclosure has doubtless served as an estufa, and there are still evidences of a considerable depression. That this ruin is quite ancient is attested by the advanced stage of decay, and that it has been of considerable height may be inferred from the large quantities of debris. A similar and somewhat more perfect example of double-walled tower is illustrated in plate VIII. There seem to have been no buildings of importance in connection with this ruin, but many in the vicinity. On the point of a low rocky promontory that extends down from the mesa on the west to within a few yards of the circular ruin are some masses of decaying wall, and a large circular depression, not differing in appearance from the usual estufa. It is probable that there are other remains higher up on the rocky slope; indeed, others could be seen fro, the trail, but I found no time to visit them. A few hundred yards below the great tower, and very near the trail, a smaller tower occurs, having other ruins connected with it, and in a weather-worn cavity in a massive crag near by is the cosy little dwelling shown in figure 5. The rude little fire-place illustrated in figure 6 was observed by Mr. Brandegee in connection with a cliff house on the opposite side of the canon, a little farther up. It is remarkable as being the only example discovered by our party. There seem to be no traces whatever of fireplaces, ovens, furnaces, or chimneys in or about any of the ruins described, which is rather remarkable, since fires must have been used in baking pottery and for domestic purposes, and we cannot suppose that a people so well advanced in architectural skill were unable to build fire-places and furnaces. PLATE V.—CLIFF-HOUSES OF THE MANCOS. In this plate are given sketches of two of the more interesting groups of cliff-houses, or fortresses, as they may quite as appropriately be called. Figure 1 represents a portion of a group found about ten miles from the foot of the canon in a subordinate cliff on the west side. This low cliff is of massive sandstone, and is washed by the river, the trail being crowded back against the steep wall. At the height of about



	13 forty feet above the river, a bed of shale occurs in the sandstone, which,being easily disintegrated, has been weathered out and carried away, leaving a sort of horizontal groove some four feet high and from four to six feet deep. In this a row of diminutive houses has been built. Three of these are almost perfect, having a fresh, new look that certainly belies their age. Four others are much more decayed, and fragments of wall only cling to the cliffs. They have been made to occupy the full height and depth of the crevice, so that when one reaches it at the only accessible point, he is between two houses and must pass through these to get at the others. The doorways are quite small and bear no evidence of the fitting or hanging of doors; and the windows, of which a number open to the front, are but a few inches square. The walls are strongly built and are from eight to ten inches thick. The stones are small, dressed roughly on the outside, and laid in mortar. In many places the heavier seams of mortar have been chinked with bits of pottery and small flakes of sandstone. The marks of the mason's pick are as fresh as if made within a few years, and the fine, hard mud-mortar, which has been applied with the bare hands, still retains impressions of the minute markings of the cuticle of the fingers. The house at the left hand in the drawing has two apartments, the farthest of which has a curved wall conforming with the rounded end of the crevice floor, which, beyond this for some distance, is broken down. Specimens of the mortar and of the dressed stone were procured from this house and brought east. Below the middle part of this line of houses, on an irregular projection, are the remains of a number of walls, in such a state of ruin, however, that the character of the original structure could not be made out. In digging among the debris of this ruin, I came upon a bin of charred corn, in which the forms of the ears were quite perfect. It seems to be of a variety similar to that cultivated by the tribes of the neighborhood at the present time. That this corn had been placed there by the ancient occupants seems probable from the fact that it occupied a sort of basement apartment or cellar, and had been buried beneath the fallen walls of the superstructures. Imbedded in this mass of charcoal, I found the very perfect specimen of stone implement figured in plate XIV (figure 3). Many large fragments of the ordinary painted pottery were also picked up here. A certain new look about portions of this group leads one to suspect that it cannot boast of great antiquity ; but it is very difficult to calculate the effects of age upon walls so perfectly protected and in such a climate. The group given in figure 2 is of a much more interesting and remarkable character. It was first observed from the trail, far below and nearly three fourths of a mile away. From this point, by the aid of a field glass, the sketch given in the plate was made. So cleverly are the houses hidden away in the dark recesses, and so very like the surrounding cliffs in color, that I had almost completed the sketch of the upper house before the lower or " sixteen-windowed ^v one was detected. They are at least eight hundred feet above the river. The lower five hundred feet is of rough cliff-broken slope, the remainder of massive bedded sandstone full of wind-worn niches, crevices, and caves. Within one hundred feet of the cliff-top, set deep in a great niche, with arched, overhanging roof, is the upper house, its front wall built along the very brink of a sheer precipice. Thirty feet below, in a similar but less remarkable niche, is the larger house, with its long line of apertures, which I afterward found to be openings intended rather for the insertion of beams than for windows.



	14 PLATE VI. I subsequently climbed the canon-walls to make a closer examination of these ruins, and the plans given in plate VI were obtained. The lower house was easily accessible, and proved to be of a very interesting character. It occupies the entire floor of a niche which is about sixty feet long and fifteen in depth at the deepest part. The front walls are built flush with the precipice, and the partition-walls extend back to the irregular wall of rock behind. Portions of the wall at the left, viewing the house from the front, are greatly reduced; but the main wall, that part which contains the window-like openings, is still thirteen or fourteen feet high. The arrangement of the apartments is quite complicated and curious, and will be more readily understood by a reference to the ground-plan (figure 1). The precipice-line, or front edge of the niche-floor, extends from a to b. From this the broken cliffs and slopes reach down to the trail and river, as shown in the accompanying profile (figure 3). The line bcd represents the deepest part of the recess, against which the walls are built. To the right of b, the shelf ceases, and the vertical face of rock is unbroken. At the left, beyond a, the edge is not so abrupt, and the cliffs below are so broken that one can ascend with ease. Above, the roof comes forward and curves upward, as seen in the profile. The most striking feature of this structure is the round room, which occurs about the middle of the ruin and inside of a large rectangular apartment. The occurrence of this circular chamber in this place is highly significant, and tends greatly to confirm my previously-stated opinion that the circle had a high significance with these people. Their superstitions seem to have been so exacting in this matter that, even when driven to the extremity of building and dwelling in the midst of these desolate cliffs, an inclosure of this form could not be dispensed with; a circular estufa had to be constructed at whatever cost of labor and convenience. Its walls are not high and not entirely regular, and the inside is curiously fashioned with offsets and box-like projections. It is plastered smoothly, and bears considerable evidence of having been used, although I observed no traces of fire. The entrance to this chamber is rather extraordinary, and further attests the peculiar importance attached to it by the builders, and their evident desire to secure it from all possibility of intrusion. A walled and covered passage-way,/,/, of solid masonry, ten feet of which is still intact, leads from an outer chamber through the small intervening apartments into the circular one. It is possible that this originally extended to the outer wall, and was entered from the outside. If so, the person desiring to visit the estufa would have to enter an aperture about twenty-two inches high by thirty wide, and crawl, in the most abject manner possible, through a tube-like passage-way nearly twenty feet in length. My first impression was that this peculiarly-constructed doorway was a precaution against enemies, and that it was probably the only means of entrance to the interior of the house; but I am now inclined to think this hardly probable, and conclude that it was rather designed to render a sacred chamber as free as possible from profane intrusion. The apartments l, k, m, n, do not require any especial description, as they are quite plain and almost empty. The partition-walls have never been built up to the ceiling of the niche, and the inmates, in passing from one apartment to another, have climbed over. The row of apertures indicated in the main front wall are about five feet from the floor, and were doubtless entered for the



	15 insertion of beams, although there is no evidence that a second floor has at any time existed. In that part of the ruin about the covered passage-way, the walls are complicated, and the plan can hardly be made out, while the curved wall inclosing the apartment e is totally overthrown. In digging among the debris with our hammers, we came upon a large earthen vessel at ft, and shortly afterward discovered another near i. They were so situated in a small recess under the sheltering walls that the falling rubbish had not reached them. Roughly-hewn stone lids were fitted carefully over the tops, but both were empty. One had been slightly broken about the rim, while the other had been pierced on the underside by some sharp instrument, and had been mended by laying a small fragment of pottery over the aperture on the inside and cementing it down with clay. They are of the ordinary corrugated pottery, and have a capacity of about three gallons. Beneath the vessels, spread out on the floor, was a large piece of rush-matting, and beneath this a quantity of fine vegetable tissue from the interior bark of some kind of tree. The vessels are illustrated in plate XIII, and the matting in plate XIV. The rock-face between this ruin and the one above is smooth and vertical, but by passing along the ledge a few yards to the left a sloping face was found, up which a stairway of small niches had been cut; by means of these, an active person, unincumbered, could ascend with safety. On reaching the top, one finds himself in the very doorway of the upper house (a, figure 2) without standing-room outside of the wall, and one can imagine that an enemy would stand but little chance of reaching and entering such a fortress if defended, even by women and children alone. The position of this ruin is one of unparalleled security, both from enemies and from the elements. The almost vertical cliff descends abruptly from the front wall, and the immense arched roof of solid stone projects forward 15 or 20 feet beyond the house (see section, figure 3). At the right the ledge ceases, and at the left stops short against a massive vertical wall. The niche-stairway affords the only possible means of approach. The house occupies the entire floor of the niche, which is about 120 feet long by 10 in depth at the deepest part. The front wall to the right and left of the doorway is quite low, portions having doubtless fallen off. The higher wall f g is about 30 feet long, and from 10 to 12 feet high, while a very low rude wall extends along the more inaccessible part of the ledge, and terminates at the extreme right in a small inclosure, as seen in the plan at c. In the first apartment entered, there were evidences of fire, the walls and ceiling being blackened with smoke. In the second, a member of the party,* by digging in the rubbish, obtained a quantity of beans, and in the third a number of grains of corn, hence the names given. There are two small windows in the front wall, and doorways communicate between rooms separated by high partitions. The walls of these houses are built in the usual manner, and average about a foot in thickness. The upper house seems to be in a rather unfinished state, looking as if stone and mortar had run short. And when one considers that these materials must have been brought from far below by means of ropes, or carried in small quantities up the dangerous stairway, the only wonder is that it was ever brought to its present degree of finish. * Mr. Brandegee.



	16 Figure 3 is given for the purpose of making clear the geologic conditions that give shape to the cliffs as well as to show the relations of these houses to the cliffs. The hard and massive beds of rock resist the erosive agents; the soft and friable beds yield, hence the irregularity— the overhanging cliffs, the niches and benches. a is a section of the lower house, b of the upper. It has heretofore been supposed that the occupants of these houses obtained water, either from the river below or from springs on the mesa above; but the immense labor of carrying water up these cliffs, as well as the impossibility of securing a supply in case of siege, made me suspect the existence of springs in the cliffs themselves. In three or four cases these springs have been found, and it is evident that with a climate a very little more moist than the present, a plentiful supply could be expected. Running water was found within a few yards of the group of houses just described, and Mr. Brandegee observed water dripping down the cliffs near a group of small houses on the opposite side of the canon. About one mile farther up the canon, I came upon the ruin photographed by Mr. Jackson in 1871, and minutely described by him as the two-story cliff-house of the Rio Mancos. It is also in the cliffs of the north side, about seven hundred feet above the river, and although not so large or complicated in design as the houses just described, it shows higher skill in construction and is in a better state of preservation. It is also exceedingly difficult of access. It seems hardly necessary for me to enter into a detailed description, as little can be added to what has already been published ;* but for the purpose of having as much of the matter together as possible I present plate VII, illustrating some of the interesting features of this house. Figure 1 gives the ground-plan, and shows the position of the house in relation to the floor of the niche. There are four small apartments only; the front one, a, being 10 feet long by 6 wide. Of the back rooms, one is 9 by 10 and the other 6 by 6 feet, while the apartment with the curved wall is much smaller. The walls are about twelve feet high and reach within from two to three feet of the overhanging roof. They are built in the ordinary manner of stone and adobe mortar, and what is rather remarkable are plastered both inside and out. This plaster does not differ greatly from the common mortar, is lightly spread over the walls, probably with the hands, and in color imitates very closely the hues of the surrounding cliffs, a pleasing variety of red and yellow grays. Whether this was intended to add to the beauty of the dwelling or to add to its security by increasing its resemblance to the surrounding cliffs, I shall not attempt to determine. Another remarkable feature of this house is the consummate skill with which the foundations are laid upon and cemented to the sloping and overhanging faces of the ledge. The buttresses b, b, which have probably at one time supported a superstructure of wood or stone, now totally obliterated, are most striking illustrations of this; and just here is a fact that has an interesting bearing upon the question of the antiquity of this structure. These wall-supports, or buttresses, have originally been four in number, one evidently having falling off, and are built in continuation of the front wall, on a smooth, sloping surface of rock. Now the sandstone of which this rounded slope is composed is rather coarse and soft, and hence easily disintegrated. It is here also not greatly protected from the weather, since the cliffs above do not

	* Bulletin No. 1, second series, p. 20.



	17 overhang to any extent, and must, year by year, yield a little to the elements ; but I observe that since the construction of these foundations no perceptible change has taken place; the thickness of a sheet of paper has hardly been washed from the surface of the rock, and the mortar, which is of almost equal firmness with the rock, lies upon it as if placed there within a dozen years, and the plaster on the outer wall, although somewhat cracked and broken off, does not add greatly to our impressions of antiquity. There is also a fact worthy of notice in regard to the question of occupancy. I have already stated my impressions that these houses were not used as constant dwelling places, but rather as places of occasional resort. I notice that, although the building seems complete, and has had its floors laid and its door-ways and windows conveniently and carefully arranged, the plastering of the interior is almost untouched, that with the exception of three names scratched in the soft, thick coat of adobe by Mr. Jackson's party, there is almost no trace of the presence of man ; yet this plaster may have been applied only shortly before the final desertion, and hence no definite conclusion can be drawn. A sketch of one of the doorways is given in figure 2. The outline is accurately drawn, but there is a little too much regularity in the stonework. It will be seen that the aperture is of very nearly the same width above and below, which is rather unusual, since, in these ruins, as well as in those farther south, the door ways and windows are, as a rule, narrower at the top. This drawing also shows the manner of employing a number of small straight beams of wood as lintels, for the purpose, evidently, of strengthening the masonry above. There are two of these exterior door-ways only, one op3ning into each story of the front room from the unoccupied part of the niche; these are shown in figure 3, a sketch of the interior of the front room taken from the side f. There is only a low wall between this room and the room c, while small door-ways communicate with the other apartments. There is a small rectangular window, 22 inches high by 30 wide, in the front wall, from which a fine view can be had of the deep narrow valley below. Figure 4 is designed to show the extraordinary situation of these houses. Whether viewed from below or from the heights above, the effect is almost startling, and one cannot but feel that no ordinary circumstances could have driven a people to such places of resort. There are no ruins of importance in the canon of the Mancos above the two-story house. Indistinct remains occur on the bottoms in a number of places, and a few small houses were observed in the cliffs. The most interesting of these is built upon a ledge about 40 feet above the trail, and is nearly midway between the two-story house and the head of the canon. It does not differ in any essential point from the ruins already described. I shall therefore pass it by, in order to take up two very interesting groups of ruins that occur about twenty miles to the northwest. Between the Mesa Verde and the Late Mountains, of which Ute peak is the culminating summit, there is a long, deep valley or strip of low land that connects the great lowland of the Lower Mancos with the canon-cut plain that rises toward the Dolores. The southern end of this depressed strip drains into the Mancos, the northern into the McElmo. The latter stream heads along the north base of the Mesa Verde within five miles of the Mancos at the point where it enters the canon, and flows westward, passing along the north base of Ute Mountain, curving around to the southwest and reaching the San Juan nearly ten miles



	18 beyond the Utah line. The large depressed area drained by this stream contains a great number of ruins, many of which have not yet been examined. PLATE VIII.—THE TRIPLE-WALLED TOWER. The group partially illustrated in this plate is situated on a low bench within a mile of the main McElmo, and near a dry wash that enters that stream from the south. It seems to have been a compact village or community-dwelling, consisting of two circular buildings and a great number of rectangular apartments. The circular structures or towers have been built, in the usual manner, of roughly-hewn stone, and rank among the very best specimens of this ancient architecture. The great tower is especially noticeable on account of the occurrence of a third wall, as seen in the drawing and in the plan at a. In dimensions it is almost identical with the great tower of the Rio Mancos. The walls are traceable nearly all the way around, and the space between the two outer ones, which is about five feet in width, contains fourteen apartments or cells. The walls about one of these cells are still standing to the height of 12 feet; but the interior cannot be examined on account of the rubbish which fills it to the top. No openings are noticeable in the circular walls, but door-ways seem to have been made to communicate between the apartments; one is preserved at d. The inner wall has not been as high or strong as the others, and has served simply to inclose the estufa. This tower stands back about one hundred feet from the edge of the mesa and near the border of the village. The smaller tower, b, stands forward on a point that overlooks the shallow gulch; it is 15 feet in diameter; the walls are 3 1/2 feet thick and 5 feet high on the outside. Beneath this ruin, in a little side gulch, are the remains of a wall twelve feet high and twenty inches thick. The remainder of the village is in such a state of decay as to be hardly traceable among the artemisia and rubbish. The apartments number nearly a hundred, and seem, generally, to have been rectangular. They are not, however, of uniform size and certainly not arranged in regular order. The walls are marked by low lines of loose rubble which show no stone in place, and I am inclined to believe that they have never been raised to any great height. It is not impossible that they have been, originally, of a species of rubble-masonry such as is seen in some of the great casas farther south, and that these meager remains are all that is left of an imposing structure, but the total want of regularity both in the form and size of the apartments seems incon-sistent with such a conclusion. In reality they are more like a cluster of pens such as are used by the Moqui tribes for the keeping of sheep and goats. The site of this village can hardly have been chosen on account of its defensive advantages, nor on account of the fertility of the surrounding country. The neighboring plains and mesas are as naked and barren as possible. The nearest water is a mile away, and during the drier part of the season the nearest running water is in the Rio Dolores, nearly fifteen miles away. To suppose an agricultural people existing in such a locality, with the present climate, is manifestly absurd. Yet every isolated rock and bit of mesa within a circle of miles is strewn with remnants of human dwellings. PLATE IX.—RUINS AT "AZTEC SPRINGS." Another very important group of ruins is located in the depression between the Mesa Verde and the Late Mountains, and near the divide



	19 between the McElmo and Lower Mancos drainage. It is stated by Captain Moss and others who have been in this locality that up to within two or three years there has been a living-spring at this place, and the spot has been christened by them Aztec Springs. The site of the spring I found, but without the least appearance of water. The depression formerly occupied by it is near the center of a large mass of ruins, similar to the group last described, but having a rectangular instead of a circular building as the chief and central structure. This I have called the upper house in the plate, and a large walled inclosure a little lower on the slope, I have, for the sake of distinction, called the lower house. These ruins form the most imposing pile of masonry yet found in Colorado. The whole group covers an area of about four hundred and eighty thousand square feet, and has an average depth of from three to four feet. This would give in the vicinity of one million five hundred thousand solid feet of stone-work. The stone used is chiefly of the fos-siliferous limestone that outcrops along the base of the Mesa Verde a mile or more away, and its transportation to this place has doubtless been a great work for a people so totally without facilities. The upper house is rectangular, measures 80 by 100 feet, and is built with the cardinal points to within five degrees. The pile is from 12 to 15 feet in height, and its massiveness suggests an original height at least twice as great. The plan is somewhat difficult to make out on account of the very great quantity of debris. The walls seem to have been double, with a space of 7 feet between ; a number of cross-walls at regular intervals indicate that this space has been divided into apartments, as seen in the plan. The walls are 26 inches thick, and are built of roughly-dressed stones, which were probably laid in mortar, as in other cases. The inclosed space, which is somewhat depressed, has two lines of debris, probably the remains of partition-walls, separating it into the three apartments, a, b, c. Inclosing this great house is a net-work of fallen walls, so completely reduced that none of the stones seem to remain in place; and I am at a loss to determine whether they mark the site of a cluster of irregular apartments, having low, loosely-built walls, or whether they are the remains of some imposing adobe structure built after the manner of the ruined pueblos of the Rio Chaco. Two well-defined circular inclosures or estufas are situated in the midst of the southern wing of the ruin. The upper one, A, is on the opposite side of the spring from the great house, is 60 feet in diameter, and is surrounded by a low stone wall. West of the house is a small open court, which seems to have had a gate-way opening out to the west, through the surrounding walls. The lower house is 200 feet in length by 180 in width, and its walls vary fifteen degrees from the cardinal points. The northern wall, a, is double, and contains a row of eight apartments about 7 feet in width by 24 in length? The walls of the other sides are low, and seem to have served simply to inclose the great court, near the center of which is a large walled depression, (estufa B.) No other ruins were observed in the neighborhood of these, although small groups are said to exist along the base of the Late Mountains, a few miles to the southwest. PLATE X. —RUIN AT OJO CALIENTE, NEW MEXICO. For the sake of comparison, I present in Plate X the ground-plan of a ruined pueblo found at Ojo Caliente, New Mexico. It occurs on a high,



	20 almost isolated fragment of terrace near Caliente Creek. It has been constructed chiefly of adobe, and has consisted of rows of apartments surrounding a number of large open courts. Individual walls cannot be traced, and the rows of houses are reduced to smooth rounded ridges of earth. These are indicated on the plan, and are often as much as 8 feet high, and 30 feet wide at the base. The courts contain a number of small circles and mounds, a, a, and the single estufa is identical in appearance with those among the ruins of Colorado. A number of openings, b, b, through the walls indicate the location of gate-ways. Metates, arrow-heads, and many fragments of pottery were found. Many other groups of ruins similar to this occur in this as well as in the neighboring valleys. Near Abiquiu, a large pueblo occurs, at which I found a stone axe and a number of arrow-heads and metates. A couple of skeletons were also obtained here. This ruin is described at length by Dr. Yarrow in his report for 1874.* PLATES XI AND XII. Although it is quite impossible to read the curious rock-inscriptions of unknown tribes, or even to conjecture to any extent their meaning, yet it is conceded that in most cases they have a meaning and represent an idea or record an event. Aside from this, however, they are valuable to the historian as records of the grade of civilization reached by the tribes who executed them. That the examples given in the two following plates belong to the age of the cliff-builders, cannot be satisfactorily proved, but, at the same time, evidence that they do, is not wanting. Some are found on the cliffs and in the niches with the cliff-dwellings, while all are in localities that must have been frequently visited by these people. Some are found in the canon of the Mancos, others on the bluffs of the San Juan, and many in the canons farther west. Figures 1, 2, and 3, Plate XI, occur on the Mancos near the group of cliff-houses figured in Plate V. They are chipped into the rock, evidently by some very hard implement, and rudely represent the human figure. They are certainly not attempts to represent nature, but have the appearance rather of arbitrary forms designed to symbolize some imaginary being. Figures 4,5, and 6 were found in the same locality, not engraved, but painted in red and white clay upon the smooth rocks. These were certainly done by the cliff-builders, and probably while the houses were in process of construction, since the material used is identical with the plaster of the houses. The sketches and notes were made by Mr. Brandegee. The reproduction is approximately one-twelfth the size of the original. The examples given in figures 7, 8, 9, 10, and 11, as well as those in Plate XII, occur on the Rio San Juan, about ten miles below the mouth of the Rio La Plata. A low line of bluffs, composed of light colored massive sandstones that break down in great smooth-faced blocks, rises from the river-level and sweeps around toward the north. Each of these great blocks has offered a very tempting tablet to the graver of the primitive artist, and very many of them contain curious and interesting inscriptions. Drawings were made of such of these as the limited time at my disposal would permit. They are all engraved or cut into the face of the rock, and the whole body of each figure has generally been chipped out, frequently to the depth of one-fourth or one-half an inch. The work on some of the larger groups has been one of immense * Report of the Chief of Engineers for 1875, page 1064.



	21 labor, and must owe its completion to strong and enduring motives. With a very few exceptions the engraving bears undoubted evidence of age. Such new figures as occur are quite easily distinguished, both by the freshness of the chipped surfaces and by the designs themselves. Figure 11 gives a specimen of the modern work; it is evidently intended to represent a horse, and is done in the manner of the Navajoes. It will readily be seen that among all the figures given of the ancient work there is no animal that resembles a horse, and we can hardly suppose that artists who could so cleverly delineate birds and deer and men, would fail in an attempt to represent an animal of so marked a character. The curious designs given in figure 10 have a very perceptible resemblance to many of the figures used in the embellishment of pottery. The most striking group observed is given in figure 1, Plate XII. It consists of a great procession of men, birds, beasts, and fanciful figures. The whole picture as placed upon the rock is highly spirited, and the idea of a general movement toward the right, skillfully portrayed. A pair of winged figures hover above the train as if to watch or direct its movements; behind these are a number of odd figures, followed by an antlered animal resembling a deer, which seems to be drawing a notched sledge containing two figures of men. The figures forming the main body of the procession appear to be tied together in a continuous line, and in form resemble one living creature about as little as another. Many of the smaller figures above and below are certainly intended to represent dogs, while a number of men are stationed about, here and there, as if to keep the procession in order. As to the importance of the event recorded in this picture, no conclusions can be drawn; it may represent the migration of a tribe or family or the trophies of a victory. A number of figures are wanting in the drawing at the left, while some of those at the right may not belong properly to the main group. The reduction is, approximately, to one-twelfth. Figures 2 and 3 of the same plate represent only the more distinct portions of two other groups. The complication of figures is so great that a number of hours would have been necessary for their delineation, and an attempt to analyze them here would be fruitless. PLATE XIII.—POTTERY. It is generally conceded that the ancient tribes of the San Juan produced fictile fabrics superior to those of the town-building tribes of to-day. There is, however, great similarity between the ancient and modern work, both in material and execution, and the differences are not greater than could be expected in the manufactures of the same tribe at periods separated by two or three centuries of degeneration. The study of the fragmentary ware found about the ruins is very interesting, and its immense quantity is a constant matter of wonder. On one occasion, while encamped near the foot of the Mancos Canon, I undertook to collect all fragments of vessels of manifestly different designs within a certain space, and by selecting pieces having peculiarly marked rims, I was able to say with certainty that within ten feet square there were fragments of fifty-five different vessels. In shape these vessels have been so varied that few forms known to civilized art could not be found. Fragments of bowls, cups, jugs, pitchers, urns, and vases, in infinite variety, maybe obtained in nearly every heap of debris. The art of ornamentation seems to have been especially cultivated, as



	22 very few specimens are found that are not painted, indented, or covered with raised figures. Indeed, these ornamental designs are often so admirable, and apparently so far in advance of the art-ideas of these people in other respects, that one is led to suspect that they may be of foreign origin. This suspicion is in a measure confirmed when we discover the scroll and the fret struggling for existence among the rude scrawlings of an artisan who seems to have made them recognizable rather by accident than otherwise. It is not improbable, however, that the specimens referred to are but rude copies of models designed by more accomplished artists or procured from some distant tribe. There is certainly no conclusive evidence that these people ever came in contact with Europeans. The material used in the manufacture of pottery is generally a fine clay, (in which the country abounds,) tempered with sand or pulverized shells. The modeling is done almost exclusively with the hand; no wheel has been used, and no implement whatever, except for surface creasings or indentings. The thickness of the ware varies from 1/8 to 1/2 an inch. Lightness has evidently been greatly desired, and vessels having a capacity of many gallons are not more than 1/4 of an inch thick in any part. Nearly all of the vessels and fragments collected have been baked or burned, but not to such a degree as to greatly change the color of the clay. Most, if not all, of the painted pottery is glazed with a very thin vitreous coating that gives a beautiful enamel-like surface of great hardness; upon this the coloring-matter is laid, apparently with a brush. With one or two exceptions the corrugated pottery is without the glazing, and in no instance contains painted figures. The peculiarities of this variety can be described more readily by reference to the examples in the plate. Figure 1 represents a large vessel obtained in one of the Mancos cliff-houses (Plate VI). It is of the corrugated variety; has a capacity of about three gallons, and was probably used for carrying or keeping on hand a supply of water. It is quite light, not weighing more than a common wooden pail, and is made of a light-gray clay tempered with coarse sand, and but slightly burned. The corrugated appearance is given by laying on strips of clay, in somewhat regular succession, and pressing them into place and indenting them with the thumb or a stick. Whether a thin shell of clay is first constructed and the strips laid on and pressed down so as to unite with it, or whether the vessel is built up by the strips alone, cannot be determined, since the inside is perfectly smooth, excepting finger-marks, and the strips are so welded into the general texture of the vessel that individual strips cannot be detected beneath the surface when examined on broken edges. In the specimen figured the workman has begun near the center of the rounded bottom and laid a strip in a continuous but irregular spiral (see Fig. 3) until the rim was reached, indenting the whole surface irregularly with the finger. Two small conical bits of clay have been set in near the rim, as if for ornament. Other specimens have small spirals, while others have scrolls, and still others very graceful festoons of clay (Figs. 2 and 2a). A number of the more distinct styles of indentation are given in connection with this figure (Figs. 3, 3a, 3b, 3c, and 3d). Figure 4 is a bowl restored from a large fragment. It is painted both inside and out, and the designs are applied with rather more than usual care. Figures 5, 5a, and 5b are prominent among the ornamental designs. I have corrected the drawing, but have introduced no new element.



	23 Figure 6 represents a very usual pattern of mug or cup. It is of the ordinary painted ware, and is made to contain about a pint. The specimen is not entire. Figure 7 is apparently a pipe. It was found by Mr. Aldrich, near a ruin on the San Juan, and is made of the ordinary potter's clay; it is 2 inches in length. Figure 8 represents part of an ornamental handle, formed by twisting together three small rolls of clay. Figure 9 represents a small spoon or ladle. Fragments of similar implements are quite numerous. Figure 10 is a portion of the handle of some small vessel. As to whether the manufacture of pottery was carried on in certain favorable localities only, or whether each village had its own skilled workmen or workwomen, I cannot determine, since, as previously stated, no remains of kilns or manufactories were discovered. The forms and styles of ornament are pretty uniform, which is to be expected in either case, since the inhabitants of the various villages must have had constant communication with each other. PLATE XIV. This plate contains drawings of a number of stone implements; arrowheads, ornaments, and other articles manufactured or used by the ancient inhabitants of this region. Nearly all were found so associated with the architectural remains that I do not hesitate to assign them to the same period. Figure 1 represents a small fragment of rush matting, a large piece of which was found on the floor of one of the cliff-houses of the Rio Mancos. (See Plate VI.) It is probably manufactured from a species of rush, Scirpus validus, that grows somewhat plentifully along the Mancos bottoms. Figure 2 represents a bundle of small sticks, probably used in playing some game. They are nearly a foot in length, and have been sharpened at one end by scraping or grinding. They were found in one of the cliff-houses of the Mancos, buried beneath a pile of rubbish. The bit of cord with which they are tied is made of a flax-like fiber, carefully twisted and wrapped with coarse strips of yucca bark; beside this a number of short pieces of rope of different sizes were found, that in beauty and strength would do credit to any people. The fiber is a little coarser and lighter than flax, and was probably obtained from a species of yucca, which grows everywhere in the southwest. Figure 3 is a very perfect specimen of stone implement found buried in a bin of charred corn in one of the Mancos cliff-houses (figure 1, Plate V). It is 8 inches in length and 2 1/2 inches broad at the broadest part; its greatest thickness is only 1/2 an inch. One face is slightly convex, while the other is nearly flat. The sides are neatly and uniformly rounded, and the edge is quite sharp. It is made of a very hard, fine-grained, siliceous slate, is gray in color, and has been ground into shape and polished in a most masterly manner. Although its use is not positively determined, it belongs, in all prob-ability, to a class of implements called " scrapers," which are employed by most savage tribes in the dressing of skins. This specimen may have been used for other purposes, but certainly not for cutting or striking, as the material is very brittle. The most conclusive proof of its use is the appearance of the edge, which shows just such markings as would be produced by rubbing or scraping a tough, sinewy surface. Figure 4 represents a part of a metate or mill-stone. The complete im-



	24 plement consists of two parts—a large block of stone with a concave surface, upon which the maize is placed, and a carefully-dressed but ooarse-grained slab of stone for grinding. This slab is generally from 8 to 12 inches long by 3 to 6 wide, and from 1 to 2 inches thick. The specimen illustrated is made of black cellular basalt, and was found with many others at the ruined pueblo near Ojo Caliente, N. Mex. (Plate X.) Figure 5 is a very much worn specimen of stone axe, which was found at an ancient ruin near Abiquiu, N. Mex. It is made of light-colored chloritic schist, and measures 2 inches in width by 3 in length. Figures 6 and 6a are specimens of ear-ornaments, such as are found in connection with very many of the ruins of Southern Colorado. These are made of fine-grained gray slate, only moderately well polished, and measure an inch and a quarter in length. Figure 7 represents a marine shell of the genus Olivella, obtained probably from the Pacific coast. Large numbers of this and allied shells are found about these ruins. They are generally pierced, and were doubtless used as beads. Figure 8 represents a small carved figure found on the Rio Mancos. It is made of hard gray slate. Its use or meaning cannot be determined. Figure 9 represents a stone ring 5/8 of an inch in diameter, and probably intended for the finger. It is made of hard gray slate; is shaped like the usual plain gold ring, and is quite symmetrical. Figure 10. Arrow-heads were found associated with nearly every ruin examined. They present a great variety of forms ; some of the more striking of these are given in the figure. The materials used in their manufacture are principally the more beautiful varieties of obsidian, jasper, and agate.



	26 canons the ascent is more abrupt; the upper third of the escarpment being generally perpendicular, with the lower two-thirds composed of talus. Their summits and sides are usually clothed with a scrubby growth of pinon and juniper trees, increasing in density and size as we approach the divide on the north, while the valleys below sustain dense masses of sage-brush and greasewood, that, in some places, attain a height of from 10 to 12 feet. Vigorous, fresh-looking cottonwoods line the main channels, and are as deceptive to the thirsty traveler as a mirage. One may travel for miles in the parched bed of the wash at their feet, while overhead their wide-spreading branches cast a grateful shade, and yet not be able to find a drop of water anywhere in their vicinity. West of the Montezuma two or three small tributaries of the San Juan head from the southern face of the Sierra Abajo, and then comes Epsom Greek, rising among the plateaus farther to the west—so called from the water in one portion of its bed having the effect and tasting like that salt. For a distance of some twenty-five miles above its mouth the valley of this creek presents upon its eastern side a remarkable wall, some 400 feet in height, inaccessible throughout its whole length with the exception of one place where the Indians have made a way for themselves. It is caused by an immense fold in the sandstones, running north and south in a semicircular line, for some forty miles, and this valley has been eroded from that portion of it where the strata stood nearly perpendicular. On the west the beds sweep up in graceful curves to a nearly horizontal position, upon which isolated mesas rise up above the general level in bold relief against the sky. The Rio San Juan drains all of this great interior basin, covering over twenty thousand square miles, as well as several great mountain masses bordering it nearly all around. It has at the mouth of the McElmo an average width of fifty yards, and a depth of from 4 to 6 feet; its current moving somewhat sluggishly in great sweeping curves that almost touch upon themselves again. The water is warm, and well freighted with the soil which it is continually undermining—a great contrast to the clear, ice-cold tributaries which give it existence. The bottoms are from three to five miles in width, and, bordering the stream, covered with dense growths of cottonwoods and willows. The broad and fertile alluvial lands, well covered with grass, and the low sagebrush benches bordering them, will undoubtedly prove a rich agricultural possession at no distant day. Back of all, upon either hand, rise up the precipitous sandstone bluffs, picturesque in outline and color, that gradually close down upon the river until it is ingulfed in the great canon commencing just below the mouth of the Rio De Chelly, and is then lost to all knowledge until it re-appears mingling its waters with those of the still more turbid and turbulent Colorado. South of the San Juan, the Rio De Chelly, coming in opposite the mouth of Epsom Creek, does not differ in its canyon character from those of the north. The bordering plateau, however, is more massive and less cut up by side canons. The same aridity prevails throughout nearly its whole length. Having thus superficially surveyed the region on which are to be found a vast number of prehistoric ruins, we will now return to the Hovenweep and examine, in such detail as our rapid reconnaissance will allow, the more prominent of the abundant remains. Starting from the pueblo of the Hovenweep, described on page 30 of Bulletin No. 1, second series, we do not find in the immediate neighborhood any other ruins of importance; but a short distance down the



	27 canon they begin to occur quite frequently. We observe first, on the left, the remains of a tower perched upon a rock, jutting out into the valley, beneath and about which were other ruins, evidently belonging to the tower. In the vicinity are other "rock-shelters," occurring upon either side of the canon, some merely walled-up caves, while others are semicircular walls built out from the rock and protected overhead by an overhanging ledge. Some seven miles from the pueblo, and about three above the McElmo, on the western side of the valley, is a jagged, butte-like promontory, of a brownish-yellow sand-rock, standing out from the mesa, upon the face of which are a number of benches and cave-like recesses. These have been built up and inclosed with neatly-laid walls, making six different houses or sets of rooms upon three benches, one above the other. Access was had from below, first by ascending a steep slope of debris for about 100 feet to the foot of the rock, where we find the first and largest of the houses. It is some 12 feet in length by 5 feet deep, divided midway into two rooms, but rendered somewhat indistinct by the falling-down of a portion of the rock back of it. The second bench was reached in the manner shown in Fig. 2, Plate 15, the little house there seen being the first of three strung along in a row. Above these were two other similar ones, very difficult to reach, the ledge upon which they stand projecting over the one beneath. The perfectly flat floor of the valley at the foot of the rock contained faint indications of having been occupied by buildings; and one of the curves of the wash, here some 10 feet in depth, in cutting away the soil disclosed a thin stratum of charcoal about 6 feet below the surface ; one piece that we picked out being about 3 inches thick, and the earth about the mass in which it occurred was much burnt, as though the fire had been long continued. About a mile farther down we came to an expansion of the valley with a canon opening in from the west. An examination up this for six miles failed in discovering any remains of stone buildings, but very numerous indications of probably adobe structures, or earthen foundations for wooden ones ; in every instance circular, with a diameter of from 15 to 25 feet. A dozen such were found within three miles of each other. Fragments of pottery of excellent quality and neatly ornamented were very abundant. Opposite the mouth of this canon the mesa juts out prominently into the valley. Half-way up its face is a bench-like spur, upon which rests an almost perfectly rectangular block of sandstone fallen from the cliff above. It is 38 by 32 feet square and 20 feet high. The upper surface is entirely covered with the remains of a wall from 3 to 5 feet high, running around its outer edge; a diagonal line divides the interior into two nearly equal spaces, one of which is again subdivided into three smaller rooms, the passage between them formed by the dividing walls overlapping, their opposite ends being set off from each other about 20 inches, thus necessitating a zigzag course in passing from one to the other. At the foot of the south side of the rock, and directly beneath the subdivided half of it, there is a line of stone wall inclosing a space 40 feet square, the rocks forming one side, with the center depressed a couple of feet below the surrounding level. In the right-hand corner of this inclosure, against the rock, are the ruins of another building 20 feet square; 10 feet above the base of the rock, and over this ruin, four holes have been drilled into it, six inches deep and four inches in diameter, serving evidently to support the roof of the building below and to afford a means of access to the rock above, a door-way in the surrounding wall being plainly indicated at that point.



	28 Two miles farther down the McElmo comes in at nearly right angles from the east, and upon the point of the mesa included within the angle thus formed by the two canons or valleys—we cannot call them streams— are a group of ruins similar to ones above, but much less regularly built. An interesting inscription occurs upon the under face of a large rock that supports a ruin, covering some GO square feet of surface; animals resembling goats, lizards, human figures, and many hieroglyphical signs abound. While sketching these our attention was called to the top of the mesa by Mr. Holmes (who has accompanied us thus far with his division on his way to the San Juan, and who had ascended to the summit for the purpose of sketching), to some very interesting remains he had discovered there. The perpendicular scarp of the mesa ran around very regularly, 50 to 100 feet in height, the talus sloping down at a steep angle. On cave-like benches at the foot of the scarp, is a row of rock shelters, much ruined, in one of which was found a very perfect polished stone implement. Gaining the top with some difficulty, we found a perfectly flat surface, 100 yards in width by about 200 in length, separated from the main plateau by a narrow neck, across which a wall had been thrown, but now nearly leveled. Very nearly the entire space fenced in by this wall was covered by an extended series of small squares, formed by thin slabs of sand-rock set up edgewise. (Fig. 3, Plate 20.) All were uniformly about 3 by 5 feet square, arranged in rows, two and three-deep, and adjusted to various points of the compass, but there were also a few circles disposed irregularly about the inclosed area, each about 20 feet in diameter, and formed of these same squares, leaving a circular space of 10 feet diameter in the center. These squares occur indiscriminately over the whole region that has come under our observation, upon the mesa tops and in the valleys, all of the same general shape and size, very seldom accompanied by even the faintest indication of a mound-like character, but nearly always in groups. We have always supposed them to be graves, but have not as yet found any evidence that would prove them such. Some that we excavated to a depth of 5 and 6 feet, into a solid earth that had never been disturbed, rewarded us with not the faintest vestige of any remains, excepting, in nearly every case, a thin scattered layer of bits of charcoal from 6 to 18 inches beneath the surface. In one instance, near the Mesa Verde, the upright slabs of rock which inclosed such a square were sunk 2 feet into the earth and projected 6 inches above it. In another, was found a mass of charred matter that promised to throw some light upon the subject, but a chemical analysis by Dr. Endlich proved it to be simply charred juniper-wood, without perceptible admixture of animal matter. In the present instance, as the soil was thin and sandy, in some places blown entirely off, leaving the bare bed-rock exposed, we excavated several of these with pick and shovel, there being only from 12 to 18 inches of earth to remove, but in no case finding anything more than the scattered charcoal spoken of above. In some, the earth was calcined, as though a fire had been made within them, while in others there was no vestige of a fire beyond the presence of the charcoal. The question very naturally arises as to whether they might not have been cremationists, a supposition that would have some appearance of likelihood, could we but find any trace of human remains among the bits of charred wood. Scattered over the whole surface of this mesa were a great many flint-chip-pings, from among which we picked up a number of very beautiful arrow-points. As the summit commands a wide sweep of country, it is not unlikely that sentries of old beguiled their tedious watch with arrow-making.



	29 From the camp at the pueblo Mr. Chittenden, of Mr. Holmes's division, rode up the Hovenweep some eight miles, to where it divided equally into two branches, and upon the point between these forks he found the remains of a round tower, commanding an extended view down the main canon. No other ruins were noticed. The parties, under the guidance of Mr. Gardner, camped one night near the head of the Hovenweep, and found there an important group of ruins, described as follows by Mr. Adams: "The first of these we met are situated at the upper edge of the side of the canon, about one-third of the distance from the top, on a ledge about 300 feet long and 50 wide. On this small space were crowded some 40 houses, as well as we could judge from the ruins. The general plan of structure was circular, varying in size, but generally from 10 to 15 feet in diameter. The stone was dressed to three times the size of an ordinary brick and in the same shape. ^# * * The whole arrangement of the little town was for defense ; perched up high above, on the summits of bowlders, were little watch towers, which commanded the plateau above." Between the Montezuma and the Hovenweep is a high plateau running north and south, from the San Juan to the Dolores ; the southern portion a level sage-covered plain, while the northern is more undulating and covered with junipers and pinon pine. Upon this we found the remains of many circular towns, generally occupying slight eminences, and in but one or two cases, as far as we observed, were so entirely demolished that not one stone remained upon another. In one of these exceptions, about half the circumference only of a tower remained, 15 feet in height and of average masonry. Broken pottery was but sparingly scattered about, showing them not to have been occupied as much as the very similar remains in the valleys below. This mesa, averaging 500 feet in height above the surrounding country, does not contain a spring or drop of water, except such as may remain in the holes in the rocks after a shower. The soil is thin and sandy, blown off clean to the rock-bed in places, yet what there is is well grassed, and sage-brush flourishes luxuriantly. As cultivation was out of the question, and permanent residence improbable, it is very likely these towns were lookouts or places of refuge for the shepherds, who brought their sheep or goats up here to graze, just as the Navajos used to, and the Utes do at the present time. Rude huts of a later day are now found scattered over its surface, by the side of the washes where water would be likely to collect. In traveling down the San Juan, from the mouth of the McElmo, there are not within the first ten or twelve miles any ruins that would claim attention upon a rapid reconnaissance. Indistinguishable mounds of earth frequently occur along the bottom lands, surrounded by the ever-present fragments of pottery, showing them to be the sites or the remains of habitations; the quantity of pottery, domestic utensils, and arrow-points helping somewhat to determine the length of time they were occupied. Crossing the mouth of the broad sandy wash of the Montezuma, that is here bordered with groves of brilliantly-green cottonwoods along its arid course, we pass about three miles below, and find camp under a grove of patriarchal trees within a well-grassed bend of the river. A wide gravelly bench, some 50 feet in height, and running back to the bluff line, rises up abruptly from the bottom lands; a few rods below camp, the river in its meandering sweeps close to the foot of this bench, producing an almost perpendicular face. Upon the top of the bench at



	30 this point, overlooking the river, are the ruins of a quadrangular structure of peculiar design. Referring to the ground-plan, as shown in Fig. 2, Plate 16, we see that it is arranged very nearly at right angles to the river, its greatest depth on the left, where it runs back 120 feet; the front sweeps back in a diagonal line, so that the right-hand side is only 32 feet in depth. The back wall is 158 feet long, and at right angles to the two sides. In the center of the building, looking out upon the river, is an open space 75 feet wide, and averaging 40 feet in depth, its depressed center divided nearly equally by a ridge running through it at right angles to the river. We judged it to have been an open court, because there was not the least vestige of a wall in front, or on the ridge through the center, while upon the other three sides they were perfectly distinct; although it is difficult to explain why it should have been hollowed out in the manner shown in the plan. Back of this court is a series of seven apartments of equal size, springing in a perfect arch from the heavy wall facing the court, leaving a semicircular space in the center, 45 feet across its greatest diameter. Each one is 15 feet in length, and the same in width across its center, the walls somewhat irregular in thickness, but averaging 20 inches, compact, and well laid. On the left are three rooms extending across the whole width of the building, each averaging 45 by 40 feet square; on the right only one was discernible. Back of the circle, our impression was that the walls diverged in the manner shown in the plan, although there is so much confusion resulting from the heaping up of the debris that much must be left to conjecture. There is also a slight shadow of doubt in regard to the wall facing the river on the right; it is barely possible that it extended somewhat farther out, although there is here a steep inclination to the brink of the bluff, and that it has become entirely obliterated by its foundations giving way. The remains of the wall above, however, led us to believe that it had been originally built in the way it is shown in the plan. Extreme massiveness is indicated throughout the whole structure by the amount of debris about the line of the walls, forming long rounded mounds, 4 to 5 feet high, with the stone-work cropping out, 20 to 24 inches in thickness. Portions of the outer wall have fallen outward almost in one solid piece, the stones remaining spread out in much the same order they occupied in the standing wall. The stones were of fair size, but yet not so large but that one man could handle the largest of them. They were obtained from the neighboring bluff, and probably undressed, but broken into very nearly rectangular blocks, so that when carefully laid and dressed up with adobe cement they would have all the effect of dressed stone. Their extreme age, which has crumbled a great many into dust and rounded the asperities of all into shapeless bowlders, renders any conjecture upon this point somewhat uncertain. Where portions of the undisturbed wall did appear above the rubbish it showed a solid, well-constructed masonry. No indications whatever could be found of any passage-ways, nor could we expect to find any so near their base, for all of the apartments were probably entered by ladders, the same as in other buildings of this order that we have found in other localities. Upon either side and back of this building were low, indefinite lines of earth, not more than 12 to 18 inches above the surrounding surface, inclosing areas from 40 to 60 feet in diameter, which were probably corrals for domesticated animals, the walls being composed of adobe or turf brought from the valley below, and which would, of course, wash down to a barely perceptible ridge.



	31 In the face of the bluff immediately under this ruin, and upon a recessed bench situated about half-way between top and bottom, is a row of little "rock shelters." A stratum of a rotten shaly sandstone has been weathered or dug out, probably both, for a distance of 300 feet along the bluff, to a depth of about six feet, leaving a firm floor and a projecting ledge overhead, with just room enough to walk along without stooping. A continuous row of buildings occupied this bench, although most of them have tumbled into the river, and none have their front walls remaining. Door-ways through each of the dividing walls afforded access along the whole line. A few rods up stream, and in the same line of the bluff as the preceding, was another little niched cave-house (figure 3, Plate 18), 14 feet in length, 5 feet high at the center and 6 deep, divided into two equal apartments; a small square window, just large enough for one to crawl through, was placed midway in the wall of each half. We well might ask whether these little " cubby-holes" had ever been used as residences, or whether, as seems at first most likely, they might not simply be " caches," or merely temporary places of refuge; and while, no doubt, many of them are such, yet in the greater majority the evidences of use and the presence of long continued fires, indicated by their smoke-blackened interiors, would prove them to have been quite constantly occupied. Among all dwellers in mud-plastered houses it is the practice to freshen up their habitations by repeated applications of clay, moistened to the proper consistency, and spread with the hands, the thickness of the coating depending upon its consistency. Every such application makes a building appear perfectly new, and many of the best-sheltered cave-houses have just this appearance, as though they were but just vacated. A quarter of a mile back over the flat level bench, is a long narrow hill about 100 feet in height, commanding an extended view up and down the valley, upon the summit of which is one of the circular, mound-like inclosures which occur so frequently upon both the highlands and the lowlands. It evidently has some connection with the group below on the river's edge, for there are no other ruins within several miles. Continuing down the river, under the great bluffs which border it closely, we find many ruins of the "rock-shelter " kind occurring frequently in all sorts of positions, from the level of the valley to a height of over 100 feet, and from the smallest kind of a " cache," not larger than a bushel-basket, to buildings that sheltered several families. We illustrate one group in figure 4, of Plate 18, that consists of a row of three small houses built upon a ledge running horizontally along the perpendicular face of the bluff, about 60 feet above the trail immediately below it. The ledge was so narrow that the buildings occupied every available inch of its surface. As near as we could judge from below, each was about 5 feet wide and 10 long, with apertures through their end walls, and in the two first ones windows in the outer wall. No possible means of access were discernible, and if ladders were ever used they were taller than any of the trees available for the purpose that grow in this vicinity. About twelve miles below the Montezuma we discovered, far away upon the opposite side of the river, a great circular cave, occupying very nearly the entire height of the bluff in which it occurred, and in which, by close inspection with the glass, we were enabled to make out a long line of masonry. Fording the river and approaching it, we found that the bluff-line at this place was a little over 200 feet in height, the upper half a light-colored, firm, massive sandstone, and the lower a dark red and shaly variety. The opening of the cave is almost perfectly cir-



	32 cular, 200 feet in diameter, divided equally between the two kinds of locks, reaching, within a few feet, the top of the bluff above and the level of the valley below. It runs back in a semicircular sweep to a depth of 100 feet ; the top is a perfect half dome, and the lower half only less so from the accumulation of debris and the thick brushy foliage, the cool dampness of its shadowed interior, where the sun never touches, favoring a luxuriant growth. A stratum of harder rock across the central line of the cave has left a bench running around its entire half circle, upon which is built the row of buildings which caught our attention half a mile away. In figure 3, Plate 16, we have a plan of a horizontal section of the cave, showing the ledge and the manner of the disposition of the buildings upon it; in the drawing at the top of Plate 17, we have a view of them as seen from the opposite side of the cave. It will be seen that the houses occupy the left hand or eastern half of the cave, for the reason, probably, that the ledge was wider on that side, and the wall back of it receded in such a manner as to give considerable additional room for the second floor, or for the upper part of the one-story rooms. It is about 50 feet from the outer edge in to the first building, a small structure 16 feet long, 3 feet wide at the outer end, and 4 at the opposite end; the walls, standing only 4 feet on the highest remaining corner, were nearly all tumbled in. Then came an open space 11 feet wide and 9 deep, that served probably as a sort of workshop. Four holes were drilled into the smooth rock floor, about 6 feet equidistantly apart, each from 6 to 10 inches deep and 5 in diameter, as perfectly round as though drilled by machinery. We can reasonably assume that these people were familiar with the art of weaving, and that it was here they worked at the loom, the drilled holes supporting its posts. At b, in this open space, are a number of grooves worn into the rock in various places, caused by the artificers of the little town in shaping and polishing their stone implements. The main building comes next, occupying the widest portion of the ledge, which gives an average width of 10 feet inside; it is 48 feet long outside, and 12 high, divided inside into 3 rooms, the first two 13 1/2 feet each in length, and the third 16 feet, divided into two stories, the lower and upper 5 feet in height. The joist holes did not penetrate through the walls, being inserted about six inches, half the thickness. The beams rested upon the sloping back-wall, which receded far enough to make the upper rooms about square. Window-like apertures afforded communication between each room, all through the second story, excepting that which opened out to the back of the cave. There was also one window in each lower room, about 12 inches square, looking out toward the open country, and in the upper rooms several small apertures not more than 3 inches wide were pierced through the wall, hardly more than peep-holes. The walls of the large building continued back in an unbroken line 130 feet farther, with an average height of 8 feet, and divided into 11 apartments, with communicating apertures through all. The first room was 9 1/2 feet wide, the others dwindling down gradually to only 4 feet in width at the other extremity. The rooms were of unequal length, the following being their inside measurements, commencing from the outer end, viz : 12 1/2, 9 1/2, 8, 7 1/2, 9, 10, 8, 7, 7, 8, 31 feet; the ledge then runs along, gradually narrowing, 50 feet farther, where another wall occurs across it, after which it soon merges into the smooth wall of the cave. The first of these rooms had an aperture leading outward large enough to crawl through; the wall around it had been broken away so that its exact size could not be determined; all the others, of which there were about two to each room, were mere peep-holes, about 3 inches in diameter, and



	33 generally pierced through the wall at a downward angle. No sign of either roofing or flooring material could be found in any of the rooms. Everything of that kind has been thoroughly burnt out or removed, so that not a vestige of wood-work remains. We cannot be positively certain that they had ever been roofed, the mild temperature of this region hardly necessitating any other covering than such as the ample dome of the cave itself offered. In the central room of the main building we found a circular basin-like depression (a), 30 inches across and 10 deep, that had served as a fire-place, being still filled with the ashes and cinders of aboriginal fires, the surrounding walls being blackened with smoke and soot. This room was undoubtedly the kitchen of the house. Some of the smaller rooms appear to have been used for the same purpose, the fires having been made in the corner against the back wall, the smoke escaping overhead. The masonry displayed in the construction of the walls is very creditable; a symmetrical curve is preserved throughout the whole line, and every portion perfectly plumb; the subdivisions are at right angles to the front. The stones employed are of the size used in all similar structures, and are roughly broken to a uniform size; more attention seems to have been paid to securing a smooth appearance upon the exterior than the interior surfaces, the clay cement being spread to a perfectly plane surface, something like a modern stucco finish. In many places, of course, this has peeled away, leaving the rough, ragged edges of the stones exposed. Inside some of the subdivisions that appear to have been less used than others, the impressions of the hands and even the delicate lines on. the thumbs and fingers of the builders were plainly retained; in one or two cases a perfect mold of the whole inner surface of the hand was imprinted in the plastic cement. They were considerably smaller than our own hands, and were probably those of women or children. In the mortar between the stones several corn-cobs were found imbedded, and in other places the whole ear of corn had been pressed into the clay, leaving its impression; the ears were quite small, none more than 5 inches long. In the rubbish of the large house some small stone implements, rough indented pottery in fragments, and a few arrow-points were found. It is a wonder that anything is found, for it is more than likely that every house has been ransacked time after time by wandering bands of Utes and Navajos, who would search with keen eyes for any articles of use or ornament left after the first spoliation. The whole appearance of the place and its surroundings indicates that the family or little community who inhabited it were in good circumstances and the lords of the surrounding country. Looking out from one of their houses, with a great dome of solid rock overhead, that echoed and re-echoed every word uttered with marvelous distinctness, and all about a steep descent of 100 feet down to the broad fertile valleys, covered with waving fields of maize, the scattered groves of the majestic cotton wood, and the meanderings of the Rio San Juan, these old people, whom even the imagination can hardly clothe with reality, must have felt a sense of security that even the incursions of their barbarian foes could hardly have disturbed. Soon after leaving the Casa del Eco, as we named the last ruins, our trail bore away to the right upon the plateaus, which now begin to encroach too closely upon the river to permit us to follow its course, and we come under a second line of bluff's, which were gradually surmounted also. The evidences of former occupation continue as numerous as ever, finding shape principally in cave-houses, all too near alike to bear much further repetition. A novel feature at one point is shown in Fig. 5, Plate 18, where we have a smooth bluff of cream-colored sandstone 3x