^^JUVA^^-l'wi^i f,^^V-.-v<>Jt-^*_,N"^~''2^'»ic)Ji'iji.tAij^ lOS.-i THE OLD INDIAN SETTLEMENTS AND ARCHITECI'ORAL STRLTTORES IN NORTHERN CENTRAL AMERICA. DR. CARL SAPRER. FRO.\r THK SMITHSONIAN REPORT FOR 1895, PAGES 537-555 (WITH I'LATKS XXIX-XXXIV). WASHINGTON : GOVERNMENT PRINTING OFFICE. 1896. Huntington Free Library Native American Collection CORNELL UNIVERSITY LIBRARY MV^EVMoFTHEAnEH-lCAN INDIAN MARSHALL H. SAVILLE COLLECTION CORNELL UNrVERSITY LIBRARY 7 939 Cornell University Library The original of tiiis book is in the Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924089417939 loss THE OLD INDIAN SETTLEMENTS AND ARCHITECTURAL STRUCTURES IN NORTHERN CENTRAL AMERICA. DR. OARL SAPPER. FROM THE SMITHSONIAN REPORT FOR 1895, PAGES 537-555 (WITH PLATES XXIX-XXXIV). WASHINGTON : GOVERNMENT PRINTING OFFICE. 1896. THE OLD INDIAN SETTLEMENTS AND AECHITEOTURAL STEUOTUEES IN NORTHEEN OENTEAL AMEEIOA.1 By Dr. Oarl Sapper. The ruins of northern Central America have for some time past enlisted the attention of large numbers of scholars, their scientific investigation having, in fact, begun more than a century ago. (Antonio del Eio in Palenque, 1787.) Nevertheless, we possess but few accurate accounts of old Indian towns and edifices, and a complete series of important new studies can not probably be expected for several years to come. Such are the accounts of ruins in Yucatan by B. Thompson and T. Maler, the thorough exploration of the ruins of Palenque by A. Maudslay, and of the ruins of Copan by an American commission, the plans of Comalcalco and Mench6 Tenamit, drawn by engineers of the Mexican Boundary Commission, and others, of the ruins on the table-land of Guatemala and Chiapas, although deserving as much interest as the majority of the lowland ruins; only very few have as yet been examined more thoroughly. I can recall here, besides Stephens's ■■' descriptions, only the examination of Iximch6 by Dr. Gus- tav Briihl,^ and thus I am compelled in my statements mainly to rely on my own observations. Now, although these are generally nothing more than the result of hasty visits and of rough sketches of the single places where ruins are found, I can not but hope that they may be of some interest, since I have made myself personally familiar with some examples of old Indian towns — settlements and edifices in almost every one of the separate ethnographic districts."* I must mention here that I have not examined these ruins with the eye of the artist or the archi- tect, but as a geographer, desiring to establish the characteristic pecu- liarities of the shapes adopted in building towns and rearing houses, as 'Translated from Globus, Vol. LXVIII, Nos. 11 and 12, 1895. ^J. Stephens, Incidents of Travel in Central America, Chiapas, and Yucatan, Lon- don, 1854, page 313 ff, 331, 365, 383 ff. ■'Globus, LXVI, page 213 ff. ■■In the territory of the Mije and Xinca tribes of Aztec and Zapoteo origin I have observed only a few unimportant building ruins, and shall therefore pay little atten- tion to them in this essay. 537 538 INDIAK SETTLEMENTS IN CENTRAL AMERICA. they differed in the various tribes, hoping that I might thus, if possi- ble, secure some points of contact with the prehistoric migrations and the ethnographic connection of the tribes. I have also tried to show how far the architecture of men depends on the physical and orographic nature of the land and the character of the building material found in the vicinity wherever this could be shown. Under the influence of the above-mentioned views, I have paid but little attention to the ruins already investigated bymore competent men, and in examining unknown or only imperfectly known ruins of settlements, I have confined myself to simple, rough measurements by the aid of the compass. The ground plans and sectional plans given here must, therefore, not be considered as accurate, but are merely intended to give a generally correct view of the disposition and the structure of the single buildings — more did not seem to be required for my purpose. As I have spoken elsewhere of the old Indian settlements of Guatemala and Chiapas, I need not return to them here. As, however, many ruins have not yet been inves- tigated (like those of Ohiapa, Tonala, and Agua Escondida in Chiapas; of Piedras Negras, Taxche, and Jolomax in Peten ; of Benque Viejo in British Honduras; of S. Jorge, Aguacatan, Sacapulas, Mixco, Ohajul, Canilla, Mita, etc., in Guatemala), and as, moreover, many other ruins have probably never yet been discovered, my material must needs be very imperfect, and the results I have obtained will have to be com- pleted and reexamined, I shall therefore limit myself here to what is most important. 1. THE DISPOSITION OF OLD INDIAN BUILDINGS WITHIN THE SETTLEMENTS. All the Indians of northern Central America have in earlier days lived, as they still do in our day, in wooden huts covered with grass or jjalm leaves, and more durable edifices were raised only for purposes of worship or of warfare, perhaps also to serve as homes for the highest spiritual and other dignitaries, in which cases earth and stone, and in times of higher refinement, even mortar, were employed. Of such only ruins are in existence, and they will therefore form the principal subject of this work. The fact that in our day the majority of Indians live in remote regions, far from Spanish influences and scattered in separate homes, or in small clusters of houses, would seem to justify the presumption that a similar system may have prevailed in pre-Columbian times. Nevertheless, the Indians had besides, in those days, larger centers of population, such as surrounded their places of worship, within carefully fortified localities, or their royal residences, the salines, gold washings, and the like. It must be borne in mind, however, that these centers of population held only at fixed times larger numbers, as at times of religious festivals and devotional meetings ; the fortresses only in times of war, the salines in the dry seasons, when alone salt could be made, Smithsonian Report, 1895. Plate XXIX. 95* SO* X Coiri-mon /ndia.n 'Ru'fis. a Tfo/'nS of Torti^/ca^a/js im Distribution of Styles of Building in Middle America. 1. Maya style; la. North Yucateo type: lb. South Yucateo type; Ic. Peten tj-pe; 3. Choi style; .3. Chorti style; 4. Vera Paz style; S. Quiche style; 6. JIame style; 7. Tzendal style; 8. Chiapanec style ; 9. Motozintla style. INDIAN SETTLEMENTS IN CENTRAL AMERICA. 539 etc. During the larger part of tlie year it is probable the majority of Indians, and even those who owned a house of their own in the towns, lived in the country in simple huts, surrounded by their cornfields, as the case at this day in parts of Alta Verapaz. It was only the Spaniards who led the Indians to congregate in real towns and villages; and as an evidence of the strangeness which this mode of life had for them, it may be mentioned that many tribes of the Maya family never had a word in their language for this idea, and hence adopted for it the Mexi- can designation "tenamit". Only in Yucatan a stronger tendency to concentrate seems to have prevailed from of old, as the people there were forced to do so by the small number of permanent ponds (aguadas), of caves through which rivers were passing (cenotes), and of real springs. The true nature of old Indian population centers can only be guessed at, since not a trace has survived of what constituted the principal part of a town, or that which was inhabited by the poorer classes who dwelt in mere huts. It is true the Spanish conquerors tell us much of streets and squares, but the actually existing ruins only show that squares, often very extensive, did exist; that in many towns they were exceedingly numerous and beautiful, but streets, in the modern sense^ of the word, I have never been able to find. Only at Iximche and in a few religious structures (Sajacabaja, Pasajon, S. Isidoro), have I found indications of such a design. Generally it is noticed that the ruins of the principal buildings (tumuli and stone structures) show no definite arrangement. A similar state of things exists even now in many Indian villages which have never been subjected to the Spanish rule of straight streets, intersecting each other rectangularly. They show nothing but a confused conglomeration of separate houses, with crooked and much intersected ways between them, but with no streets in our sense of the word. As the church, with its open square around it, now forms the center of these villages, it may be that the groups of public buildings may formerly have also formed, as it were, the kernel of similarly shaped settlements. The old Indian towns of Guatemala and Chiapas have certainly, in ordinary times, harbored no very considerable population, for the space lying within the fortification line is generally very confined, and it is highly improbable that outside of this line other parts of the town should have been added, since such a proceeding would in war times have been as diastrous for those who lived outside as for the fortress itself. It maybe argued, on the other hand, that the older Spanish writers have left us very minute descriptions of many old Indian towns, but I must confess that I am very skeptical as to their accuracy in such mat- ters. They seem to have indulged with great pleasure in large numbers, feeling sure of not being contradicted. Thus, Fuentes tells us that the chief commanders of the Quiches, Tecum Umau, in the year 1524, had 540 INDIAN SETTLEMENTS IN CENTRAL AMERICA. drawn 72,000 warriors from the capital, Gumarcah (Utatlan), alone, and j[^ / ^ j /; the royal palace of Utatlan was said to have been 728 steps long and 376 "" ' '^'•'VN steps wide. If we now recall these astonishing numbers in the light of the ruins of Utatlan, we can hardly keep from smiling, for the habitable surface of the real table-land of Utatlan is not quite 100 meters long, and would, therefore, at best not be able to contain an edifice of dimen- sions like the above mentioned. In order to be quite certain about this question, I measured the plateau by walking along the edges, when I visited the ruins in August, 1894, with my brother Eichard, but I had not the time necessary to measure the buildings also. We agreed, however upon the fact that the principal court of the so-called palacio is only 100 yards long and 60 yards wide, and that the disposition of the surrounding buildings is almost identical with that of the so-called Eesguardo. (I have inserted in the plan of the site (fig. 10) the place of the most important buildings from memory, because I subsequently saw that the plan in Stephens's Incidents in Travel, page 235, gives an erroneous impression.) These ruins have, moreover, since Steijhens and Catherwood visited them, suffered much from dilapidation, mainly by the fault of diggers for treasures, who foolishly turned the whole plateau upside down. It might, to be sure, be assumed that the table-land of Utatlan con- tained only the palace of the ruler with the accessory buildings and the temples, while the rest of the city might have covered the surrounding plain. In fact, there are at some distance from Utatlan a few tumuli rising in the plain which might be considered detached forts, built to protect the parts of the town in which the jioorer people lived. But the Spanish writers say nothing of such an outer town, and the surface of Utatlan is no smaller than that of many other Indian fortified places like Saculeu, Comitaneillo, Iximch6, and others. The nature of old Indian centers of population differed, of course, according as the settlement was made principally for defense or for the performance of religious worship, or merely for the maintenance of a king or a prince. In the table-lands of Guatemala and Chiapas, where a number of warlike peoples and independent hostile tribes of one and the same nation dwelt in close proximity to each other, the fortified character of their buildings naturally prevails, and they usually con- tained also the palaces of their rulers, and the temples of the deities.' Nature here offered in abrupt eminences, which were entirely or par- tially severed from the adjoining table-land by deep ravines, or on mountains with a level plain on their summit, places that could easily be defended and were really used for that purpose by the Indians. In building towns here, where nature had limited the space, the builders 'At times the princes also lived in oi^en towns which they abandoned at the out- break of war, withdrawing into near fortified places, as in the year 1525 Caibil- Balam, King of the Mames, retired at the approach of the Spaniards under Gonzalo de Alvarado from his capital, Chinabhul (Huehuetenango), to the fortress Saculeii. INDIAN SETTLEMENTS IN CENTRAL AMEEICA. 541 saw themselves compelled to crowd the structures as much as could be done, and this closely compressed disposition is hence characteristic of the ground plans of towns ou the table-land of Chiapas and Guatemala. As an example, I may here mention the well-known ruins of Toning (that is, stone house), of which, however, I have only drawn the upper part (fig. 8a). The mass of the ruins lies upon a narrow ridge of hills, which in the direction of Tonina Creek, terminates the principal build- ings (shown in fig. 8a) at the eastern end of the same. Already, below in the plain, considerable artificial hills are found. Then we ascend four distinct terraces, also quite high and artificially produced, the second of which is tolerably wide and bears some cross tumuli; and thus we reach at last the fifth, with the palacio (stone house E), which since Stephens's visit has evidently suffered much. Higher up still stand the two great pyramids. All the buildings are closely crowded to save space and evidently mainly intended for defense. In Yucatan, however, where the supply of water was always a seri- ous question, and where nature, moreover, had not provided such easily defensive localities, the principal buildings are much more freely scat- tered about, and some of them might have served for defense. But the whole arrangement is such and the decoration of the outer walls so pro- fuse, that these towns must rather be looked upon as places of residence for their princes and high priests than as fortified places. It is true that I know only a very few such settlements in Yucatan, but if I may judge from Charnay's statements and from the still unpublished plans, drawings, and photographs of Mr. Tompson, in Merida, a like scattered disposition seems to prevail elsewhere as well as inUxmal or Tzibinocac. The southernmost Maya buildings, that is to say, the town ruins in Peten, like S. Clemen te, and especially grandiose Tiewl, show, on the ^^^ other hand, clearly that they were intended for fortifications. The crowded position, the variety of isolated buildings, and the arrange- ment of many around a court, each one of which formed a new center for defense, prove this beyond all doubt. In spite of Mr. Maudslay's careful researches, we have as yet no really complete account of Ti#«l, hcL and I was unfortunately unable to trace the ground plan of these grand old town ruins, which are slumbering here in the shade of primitive forests. I can only say that here may be seen a whole series of easily defensible courtyards, which in part lie in the form of terraces one above the other, while in the vicinity of the principal court, surrounded by magnificent buildings of stone, a number of steep, defiant pyramids arise, each bearing a grand stone building on the summit. Much simpler and less important, and on that account also much ■^^^'^^ more easily understood, are the ruins of S. Clemente (fig. 9), which -Ojui'vrxi^^^ had long remained unknown, concealed as they were in the forest, although they were within 200 yards of the riding path from Peten to Belize. The ruins cover a somewhat long hill, over which the buildings were so scattered that they formed a number of courtyards or squares. 542 INDIAN SETTLEMENTS IN CENTRAL AMERICA. SV^^^_ i almost everyone of which might have been most easily defended by "^^*voi-TAt«J ^*®®^^- ^^^® courts B and G are on the same level, while the square :' marked A is one story higher, and the square D, separated from A by ;' a narrow ravine, lies perhaps 4 meters higher. The stone houses I and II show on the outside only a smooth wall, at the foot of which begins J a sheer precipice. The rooms in I and II are accessible from it, but the stone house III has its doors of entrance on the south side, now sadly in ruins, from whence they can easily be reached across a steep parapet. The upper plateau of III continues eastward at the same height, so that the continuation (Ilia) viewed from B looks two-storied, until, at the end of the edifice, it becomes once more one-storied. A narrow passage leads from the southwestern corner of B through Ilia to 0. The western termination of B is formed by a wall, which rises as high as the square A; the eastern and northern ends consist simply of stone walls, and in the same way the eastern and western termina- tions of G and E. The walls 4 and 5 are built of cut stone and 3 to 4 meters high. On top of the rampart 9 a small much-decayed stone house is standing. Between and D two tall, strongly- built stone houses are seen, each of which contains but a single room, open to the north, upon high artificially modified eminences. What is very remark- able is that at the foot of the hill, looking toward the northwest, a round hole has been found, barely large' enough to let a man pass through. This leads to a subterranean story below, which I, however, did not dare to examine, as I had neither a rope nor suflflcient light. In the former Choi territory also similar connections of houses seem to exist, built in the shape of terraces side by side (e. g.. Las Quebra- das). Copan, also, otherwise in its disposition perhaps the most remarkable creation of Indian architecture, shows certain features of the same system. The ruins in southern Yucatan are inferior in extent to those of the Maya territory. They often display the clearly pronounced character of fortifications, walled-in courtyards on high hills (as in Ixtinta, fig. 2) or extensive stone walls, or buildings on high passes, as upon the height of Caca de Xkaoja, which may have served for the defenses, but may also have been used by travelers for the offering of prayers and of sacrifices (fig. 3). At all events, the type of fortified places is less pro- nounced here than in Peten. The buildings are less crowded, and the houses, built of stone, show much more careful, almost artistic, treat- ment of the outer walls. On the other hand, the structures are still not quite as much scattered as in the towns of northern Yticatan, and they lack the ornamental sculptures of the latter, so that the ruins of southern Yucatan occupy an intermediate position between the edifices of northern Yucatan and those of Peten. In like manner we find in Mench6 Tenamit certain features which connect Ti(fal with Palenque, and Tonina recalls in its buildings the towns of the lowlands, but follows in the arrangement entirely the habits Smithsonian Report, 1895. Plate XXX. ^ r^ •^\•.^■-^,.^■^^\■^-•^ •■^•;^ JD iSM^^^^^^^^^^^^ I E w Tl n f'JiSl JUJ rar J LTT J L RJ 1 ■c] '"' ■ffcx^" lEJ %0 1-' D |Sl ^ z d 6 6 6 Ui M M W ,- 'fS 1 fl d S - — O o j9 _ <"-3 C.d en* ■326 I I. n^ I i (i 6 g d; CJ ti;i fc* pq fcf pq INDIAN SETTLEMENTS IN CENTRAL AMERICA. 543 of the tribes on tbe highlands. There can be uo doubt that the cul- ture of a nation in northern Central America has always had its influence on the neighboring tribes, and thus we notice frequently, especially in frontier districts, features which remind us of the puculi- arities of the architecture of adjoining districts. The isolated westcnu court of Ohama presents very clearly the type of Verapaz, while the eastern buildings remind us more of the Choi buildings, and thus also the ruins of Pneblo Viejo, whilst the nearly adjoining ruins of Chacujal, which probably once formed part of the whole, are entirely original. In Verap iz we know only of small ruins of edifices, which in their simplicity contrast strikingly with the more complicated disposition of buildings in the settlements of both the lowland and the highland dis- tricts, although they share with them the fundamental type of a court- yard walled in all around or only in part, within which ordinarily small terraced pyramids have been standing. We also know of fortifications attempted in Verapaz, such as walls closing a narrow pass (at Las Pacayas), or mountain summits fortified or rendered inaccessible by piled-up masses of stone, e. g. Yaltenamit. As I have not examined the few remains of settlements which I know near Elbarrizal and Gua- temala, I am not able to state whether the ruins within the Pokomam territory bear the Verapaz imprint or the character of the highlands. Old Indian establishments for purposes of worship have been com- paratively rare (e. g., Kalamt6) and they also (as, for instance, Sajacabaja, Copan) were at the same time arranged for defenses, as of course the temple buildings with their walled-in courtyards and their terraced pyramids famished groups of buildings that could easily be defended. Palenque I consider, with Charnay, a city for priestcraft and higher culture, also Q,uirigu4 and the ruins on the Eio de la Pasion, where F. Artes in 1892, commissioned by the Guatemalan Government, obtained photographs of the monoliths and exhibited them at the Chicago Exposition. A survey of the ruins within the Maya territory, as far as they are known to us, convince us that everywhere the fundamental type of inclosed courtyards reappears. In the highlands of Chiapas and Guatemala the disposition of the buildings is compact, since the build- ings bore mainly the character of fortifications, and on that account localities were chosen which were naturally already confined, such as ravines, sudden precipices, etc. In Peten also the buildings are much crowded, evidently on account of warlike events which then occurred, although on the whole the settlements in the lowlands are more open and without any signs of defensive works in the foreground. In all Maya ruins the buildings are, if not uniformly, at least very generally, built so as to face a certain direction; among the lowland tribes, toward the cardinal points. In the Verapaz tribes and among the Quiches, Tzutuhiles, Uspantecos, Aguacatecos, and other high- land tribes (the Tzendal group, Mame group, and Cakchiquel), the 544 INDIAN SETTLEMENTS IN CENTRAL AMERICA. buildings face more generally in intermediate directions, but always in such a manner that within each town one certain direction prevailed. In eastern Chiapas I have seen at Mazapa and Mbtozintla certain ruins which differ from the Maya type, although the people there now speak Maya languages. Besides other peculiarities, the absence of clearly defined courts must be noticed and the long drawn out charac- ter of the general plan (see fig. 11). Quite near by, at Chimalapal, I saw from a distance old Indian settlements, with clearly defined courts of the Maya type, facing the cardinal points (fig. 13). In the Chiapas territory no such clearly defined courts as the Mayas have can be found. The courts, if at all existing, are not completely walled in, the buildings do not seem to face any one direction decidedly, but to be scattered about without any rule. The choice of locality, however, and walls evidently built for defensive purposes, show clearly that the builders intended to give to the whole the character of fortifications. In western and southern Chiapas, in Soconusco and southern Guate- mala, I found but few old Indian settlements, and those I did see were so completely ruined that I was not able to discern any striking peculiarities. In the territory of the northern Pipiles, in the upper Motagua Yal- ley, and in lower Verapaz, I have frequently seen traces of old Indian settlements, but they were almost completely effaced and beyond rec- ognition. The ruins near S. Agustin Acasaguastlan are long stretched out, resting in one direction on a mountain slope, somewhat like the ruins of Mazapa. They show terraces and half courts and always face the cardinal points. 2. SINGLE BUILDINGS AND GROUPS OF BUILDINGS. I have above called attention to the fact that the Indians of northern Central America lived in the days before Columbus in straw huts, as they do now, and there is no reason to assume that they have changed since in the construction of their houses. There existed, therefore, in those days the same difference in building among the various tribes and groups of tribes as I have been able to see in the present time and to describe elsewhere.' For the larger buildings, however, and especially for the substruc- ture, other materials were employed which promised greater durability. Where civilization was still lagging behind, walls of earth and of stone had to suflBce, or terraced pyramids were erected of the same material, and probably bore on the summit wooden structures adapted to the desired purpose. The most primitive shape of these walls were probably simple walls 'Contributions to the Ethnography of the Republic of Guatemala (Petermann's Mitteil., 1893, p. 12 ff) and contributions to the Ethnography of Southeast Mexico and British Honduras (same journal, 1895, p. 177j. Smithsonian Report, 1 895. Plate XXXI. Fig. 86. Fig. 8a. Principal Parts of the Ruins of Tonina, near Ocosinqo (Chiapas). INDIAN SETTLEMENTS IN CENTRAL AMERICA. 545 of earth ■whicli, as reflnement increased, were incased in a covering of stone. Frequently, however, the whole wall was built of stone, and even the terraced pyramids consisted only of a kernel of earth, some- times containing stone chambers within, while a covering of stone on the outside gaive to the building a suitable outward form and durability. This is the point of development where the majority of the buildings of Chiapas, southern Guatemala, and Verapaz remained stationary, and even among the Maya tribes, who are so much further advanced, similar structures are still met with. At this time the stones that form the outer case were either not cut at all or only roughly; really well- cut stone is very rarely met with in such structures. This depended, naturally, very largely on the nature of the stone found in these dis- tricts. In Alta Yerapaz and parts of central Chiapas the material is an easily split dolomite or limestone; in the Chiapas and Motozintla districts granite prevails; in the districts of the Tzotzil and the southern Pipiles other eruptive stones of more recent origin, which the Indians with their extremely imperfect tools must have found very hard to cut. The same difficulty no doubt also accounts for the fact that we find in the highlands of Guatemala and Chiapas compara- tively few sculptures in stone, and that wherever any are found the nature of the material on hand has been specially favorable. Andesite, of recent origin or disunited, has frequently been used for the purpose, more rarely sandstone or even limestone; for small articles sometimes argillaceous schist, but never dolomite. In proportion as the available stone material was less fit for building and sculpture the love of ceramics increased, and, to jnention but one example, the Indians of Alta Verapaz remained far behind their neighbors in all that con- cerned architecture; but, on the other hand, they furnished pottery of such admirable good taste and true artistic skill that they do not seem to have been surpassed by any of the Maya tribes. This would lead to the conclusion that architecture is not a standard by which we can measure the culture possessed by a people, because the absence of suitable material may easily interfere with their development and force their artistic predilections into other channels. Nor must it be for- gotten that architecture is fostered and improved in proportion as it is favored by a nation politically and financially powerful. This seems to have been much less the case in Alta Yerapaz than in Yucatan or in the highlands of Guatemala. Edifices consisting merely of earth and of stone, simply piled one upon the other, are generally found in a lamentable condition, and it is but rarely possible to trace the outlines of such buildings accurately. The same difficulty applies to the nature of the steps which are almost always found on pyramids, and frequently on ramparts. As far as my researches go, the ground plan is generally square, rarely (by cutting off the corners) of five or more sides, but in spite of these variations, in the upper part of the buildings, square once more. Eound foundations SM 95 35 546 INDIAN SETTLEMENTS IN CENTRAL AMEEICA. I have never been able to trace with any certainty ; a careful inves- tigation led me in almost every case, even in apparently round or rounded-off structures, to trace the originally rectangular lines. The steps in the Maya buildings seem always to have been produced by an alteration of horizontal with perpendicular or nearly perpendicular planes; they are at the same time usually of the same height and depth. A very striking difference appears, however, in the structures of the Chiapas and Motozintla tribes, as they ascend sideways and leave only a small space horizontally open (iig. 11a). It is possible that this peculiarity betrays a certain dependence on the building material, as the rolling, rounded-off' bowlders of granite which abound in that region can not very easily be piled up peri)endicularly, and hence the building would acquire greater durability by steps ascending in a side wise direction. However this may be, the fact is that the method of building here differs essentially, and I feel justified in concluding from it that the districts of Motozintla and Mazapa, where now Maya idioms are in use, were formerly inhabited by a race of foreign origin. What race of men this may have been I can not even guess; I only believe that tliey could not even have been Chiapas, partly because the build- ings in the Motozintla district seem to be more carefully arranged than among the Chiapas and partly because in front of several tumuli in Masapa and Motozintla (fig. 2, A, B, and C) carefully wrought pave- ments of fiat granite tiles may be seen, such as I have until now never met with in the Chiapas district. When we meet with perpendicular or nearly perpendicular walls of cut stone we may assume that this indicates a higher style of architec- ture, even though these walls may be erected without the aid of mor- tar. Such buildings are met with here and there, as in S. Agustin Acasaguastlan, frequently also in towns, where already stone houses are found standing, e. g., the stone tumuli 4 and 5 in the court C of the ruins of S. Clemente (fig. 9). The most remarkable of such edifices are those of Chacujal (Alta Verapaz), where primitive argillaceous slate, carefully carved on the outside, has been employed without any kind of cement for the purpose of raising perpendicular or very steep walls, and which bear on the upper platform a Itind of parapet. The inner kernel of these walls consists of rounded-off liver shingles. I have never found this same method of building anywhere else. Still greater progress in architecture is seen in those structures on the high table-land of Guatemala, in which the stones forming the walls are held together by an abundant use of mortar. Mortar is, by the way, also found elsewhere (e. g., in Alta Verapaz), but not to such an extent that its use should have essentially influenced architecture. Even in Iximch6, mortar seems to have been but of secondary impor- tance. In Kalamt6 and Comitanuillo, in Utatlan and Saculeu, how- ever, many edifices consist simply of walls, and in order to secure steps these walls were erected perpendicularly; but where higher walls had Smithsonian Report, 1895. Plate XXXII. ^iillililiife \\^ Fig. 9.-RUINS of San Clemente (Peten), 1:1600. INDIAN SETTLEMENTS IN CENTRAL AMERICA. 547 to be raised without steps they were erected with a steep inclination, and separate staircases led up to the platforms. And it was the same with terraced pyramids that had especially high steps. A casing of smooth mortar formed the outer covering of these walls. Similar appli- cations of liquid mortar formed the floor of the more important Y)laces and of the platforms of tumuli. In Utatlan, even now and in spite of the general destructiou of the buildings, traces may be seen of paint- ings on the walls, and on some of the platforms the evidently griev- ously injured casing of mortar has been covered by a second and even a third application. In the highlands of Chiapas, these last-mentioned architectural forms seem to be wanting, and the tribes of the Mam and the Quiche family appear therefore to be in some way technically opposed to the Tzeudal group. Yet the highland tribes of Guatemala and Chiapas show not only in the manner of laying out their towns, but also in the erection of any definite temples, a surprising uniformity. We speak of a temple building consisting of tAvo main structures which are alike, lie parallel to each other, and display on the side that faces the other a small, low terrace resembling a trottoir. Between these two edifices the temple court appears, deeply sunk, but spreading out wider beyond the two main buildings and almost altogether walled in, so that the shape of the court assumes a resemblance to a large letter H or !. From El Sacramento in Chiapas to Sajacabaja and Kalamte these H -shaped temple courts reappear with the same ground plan, but yet each one has its somewhat modified form. I insert a few slight sketches of such courts (figs. 12 and 10a). It is very remarkable that in Iximch6 this kind of temi)le building ^c-v->^f»/u is'altogether wanting, or at least there are hardly any traces of such fundamental ideas to be discovered there, and yet the Cakchiquel have displayed in their architecture, which they have developed with great originality, no small correspondence with the same art among their neighbors. Thus we find in Ixiiuch^ a rectangular, lengthy tumulus ("A" in Bruhl's Plan of Iximche, Globus, vol. 66), the platform of which is walled all around, and this shows a courtyard which, relatively to the edges of the tumulus, is sunk deeply. The same manner of building I found twice in Sajacabaja and once in Saculeu, though here not quite according to the type. It may be presumed that on all the edifices which have so far been discussed, unless they were intended to serve exclusively for defease, wooden huts must have been standing on the uppermost platforms, which either contained the images of their idols or may have served as reception rooms or residences for eminent personages. Among the lowland tribes of the Maya family, however, architecture has taken a step further in advance by substituting for these mere wooden build- ings structures of stone with durable and habitable inner rooms. The fact that in Yucatan, northeast Chiapas, and in Peteu well-stratified^ 548 INDIAN SETTLEMENTS IN CENTRAL AMERICA. level limestone appears, lias evidently caused and favored this improve- ment, as the appearance of finely grained limestone, resembling the slate of Solenhofen at Palenque and Mench6 Tenamit, must have caused the flourishing industry of relievo painting in that district. In the Chorti territory, where, nearOopan, an easily worked building material appears in abundance (a decomposed, eruptive stone), architecture has progressed in a very peculiar manner, but the stone houses, while other- wise apparently of the same construction, appear less large and impos- ing than in the more northern districts. In the low plains of Tabasco (Oomalcalco) the Indians (Chontal) have erected stone houses by the help of an artificially produced building material (bricks), beyond doubt in imitation of the stone houses of their eastern and northeastern neighbors. Stone houses, so far as is now known, have not been built by any but the lowland Maya family m North Central America — that is, by the Mayas and the tribes of the Choi group (Choles, Chontales, and Chor- tles). These are the same tribes which even now difi'er from the other Maya people by certain peculiarities in their house architecture (advanced walls). Such stone structures as the Mam and the Quich6 tribes possess are here entirely wanting, as are also the temple courts resembling the letter H. In the territory of the Chontales only the ruins of Comalcalco, and in that of the Chortles, only those of Copan are known tolerably well. As I do not know the former from personal knowledge, and during my visit to the latter (January, 189|) only found a beginning made of a more careful examination, I can add nothing new to what has been stated before. I limit myself, therefore, in the remarks that follow to the observations which I made in the Choi territory in Peten and Yucatan. The ground plan of almost all stone houses is rectangular, and wherever wings or other additions appear they also are rectangularly added to the main building. In Yucatan I saw several rounded-off edges on tower-like side wings (Ixtinta, Tzibinocac), and I thought it remarkable that these exceptions from the general rule should occur there alone, where the dwelling houses of the Indians uniformly show rounded-off ground plans. These stone houses are iu their simplest form narrow buildings with but one inner apartment, to which access is had from the side of a passage (e. g. the stone houses V, VI, and VII in S. Clemente). Where the buildings, show any progress, the one inner room appears sub- divided by niches, passages, and additions, and is approached by several doors of entrance on one side (fig. 6a, the principal temple of Mench(5 Tenamit), or several separate rooms are found in the same stone house, connected with each other, but each having its own means of entrance from without — e. g. the stone houses I, II, and III iu S. Clemente, fig. 9. If architecture has made still greater progress we Smithsonian Report, 1895. Plate XXXIII. o o o INDIAN SETTLKMENTS IN CENTRAL AMERICA. 549 find two or three rows of rooms, one beMnd the other, which are con- nected with each other and have exits, on one or the other of their longer sides (fig. 7). It is not my intention here to pursue the almost boundless variety of plans of stone houses at the various places of ruins; I only mention here that where architecture has progressed still further, side wings are found — added to the main building (barely indi- I ^"^ cated in fig. 7, Tiie»l, more clearly in the stone houses of Tzibiuocac, fig. 4, and Ixtinta, fig. 1), or the building incloses a court partly (Tie«l, KtK fig. 6) or entirely (Palenque, Uxmal). In Palenque the front wall is sometimes reduced to a number of pillars by the frequency, the height and width of the door openings, so as to change the wide, outer room in its great length into a kind of well-lighted portico. The stone houses of Tonin4,^ the only ones known to exist in the territory of the Highland tribes, are, as far as the ground plan is con- cerned, most nearly related to the structures at Palenque. The outer walls of the stone houses rise either perpendicularly or they are steeply inclined, parts occasionally extend even beyond the foot of the walls. On the whole, however, the horizontal section through the building diminishes in proportion as it is made at a greater height — that is, the building grows smaller from below upward. The outer walls are partly shaped by a smooth layer of mortar (so usual at Peten), partly adorned with stucco (Mench6 Tenamit, a few houses in Ti««l), ^cv partly ornamented with separate tablets, showing images or hiero- glyphics (Palenque), or cased with a smooth covering of stone (South Yucatan), which in North Yucatan is adorned with sculptures. The substance of these buildings inside the casing consists, where no well- stratifled calcareous schists are found, of bowlders and an abundance of mortar. All around the edifice continuous cornices are seen, which produce the appearance of a building of several stories, as they occur at almost equal distances, one above the other. This impression is aided by the fact that the external divisions of the outside occur exactly at the place where the cornices appear. Thus one meets with buildings of one or four stories, though ordinarily they are only of two or three stories. Sometimes, again, certain parts of a stone house are higher than others, and when in this way (as in the stone houses of Ixtinta and Tzibanocac) the main body and the wings seem to be of different height, such structures gain a certain variety of forms which is pleasantly felt in contrast with the general uniformity of all Maya buildings. The inner rooms of stone houses are small and rather badly lighted, since no daylight enters except through the door openings. Only rarely little low windows are found, which are pierced through the ' In Kalamt^ I saw the scanty remains of a small stone house, the thin walls of which, however, caused the presumption that these could not have supported a massive upper story, such as is characteristic of Maya buildings. It seems, therefore, to have heeu a different kind of construction. 550 INDIAN SETTLEMENTS IN CENTRAL, AMERICA. outer walls, if they do not contain several inner rooms of the same house (Tioul, flg. 6). The upper closing of a room is brought about by the gradual approach of the longitudinal walls, till they are near enough to each other to allow the space between to be closed by a few flat, stone tiles. This reduction of space is produced by overreaching stones, each upper row protruding over the one below. The edges thus produced are smoothed over with mortar (e. g., in Toninaflg. 8b) or tliey are entirely concealed. Where more careful work was required, the stones were cut obliquely, so that when laid one upon the other they would show a straight-lined reduction (flg. 7b, in Ticul), and in Uxmal may actually be seen, in a few cases at least, a few slightly curved lines of reduction, convex or concave. Between the two walls which are thus treated to lead to a closing above there are commonly found some cross pieces of wood, generally zapote wood, which were meant to increase the durability of the structure, and perhaps in dwelling rooms, to suspend hammocks. Above the door openings, which are simply covered flat at the top, without any effort to approach the sides, strong cross beams, mostly of zapote wood, serve as supports; in PaleiKiue and Mench^ Tenamit huge slabs of stone. Where the inner rooms are long and narrow, only the long sides are shortened; on the two short sides the walls go up straight and unreduced; but if the four. sides do not differ much (as in Tonin4), all are shortened in the above-mentioned way. In small and narrow passages the closing is brought about by horizontal slabs of stone. In ToninA a peculiar way of forming a ceiling is noticed as shown ia flg. 8b. The inner rooms of a stone house are generally of the same height. Staircases in the interior of houses, I have never seen — excepting the famous tower of Palenque — unless it be in the tower-like raised side wings of Ixtinta, where they only lead on the outside to the upper plat- form.' The inner rooms of Maya stone houses are, as a rule, lacking in orna- ments; only rarely wall paintings are seen (as in OhichenitzS,, Tonin4, Tzibinocac), or stucco ornamentations (Toning), or in separate niches relievo tablets and hieroglyphics (Palenque), or statues (Mench6 Tena- mit). Most structures of this kind show their principal ornamentation on the outside. The outside of stone houses in North Yucatan are specially rich in sculpture adornments ; and here the contrast of the architectural style with that of Mitla in Oaxaca (the Zapotec district) is most startling, for the above-mentioned ediflces, which also difter fundamentally from the Maya buildings in the construction of their roofs and in the introduction of round pillars of stone have their prin- cipal adornment in the interior, while the external walls are left com- paratively plain and unadorned. This great simplicity of the small inner rooms in a building which is on the outside almost too richly adorned, as is the case in the Casa del Grobernador at Uxmal, makes a ' others have also been observed in ISdrth Yucatan and Copan. Smithsonian Report, 1895. PLATE XXXIV, tS i (3 INDIAN SETTLEMENTS IN CENTRAL AMERICA. 551 very peculiar impression upon the beholder and reminds him involun- tarily of the narrowness of views which seems to he a universal weak- ness of the Maya family, who otherwise jjossess so many most attract- ive characteristics. Besides the stone houses, several towns of Yucatan and Peten possess still another remarkable type of ediflce — steep stone pyramids, which on their uppermost platform bear a somewhat long stone house.' I have seen these structures only in Uxmal and Ti««l, and the four stone pyr- l^aj amids at the last-mentioned place were so entirely overgrown with forest trees and dense shrubbery that I could not obtain a clear view,^ although I climbed up to the top of one with great trouble. The stone pyramid of Uxmal rises in two unequal terraces, over which a compara- tively low, perpendicular terrace leads to the ujiper platform which bears the stone house. In the middle of one of the long sides (from the east) a very steep staircase with nearly one hundred steps leads to the platform. This kind of pyramid seems to be peculiar to the Mayas (in Yucatan and Peten) and to Copan, for in the territory of the Choi and the Ohorti we find only the ordinary terraced pyramid, with x)erpendic- ularly ascending steps of equal height and depth. The same applies to the highlands of Guatemala and Chiapas, where*it is true the steps occasionally seem to lose their original purpose and to assume gigantic dimensions (1 to 2 meters in height aild depth), as in Saculeu and Toning. In the two great pyramids of Tonin4 we notice, from the bottom to the uppermost small platform, six to eight lofty steps. These steps, however, being so very high, do not ascend perpendicularly (as at Saculeu), but at a great inclined angle. That Catherwood's Eecon- struction of the Pyramids, in Stephens's Incidents of Travel (p. 384), is incorrect, is best seen from the northwest, where both show their best preserved side. I am sure I do not wish to blame Catherwood, but only to point out the great difficulties which arise at every such attempt at reconstruction. It is extremely difficult to form by means of decay- ing ruins a picture of former proportions, and often a mere glance from an accidentally chosen but favorable standpoint gives a clearer sight than a long study of the ruins themselves. To this must be added that the generally poor state of preservation of all such ruins has led almost everyone to add to this sketch or ground plan a number of reconstructions; my own very trifling sketches must not be considered final in any way. I shall be well content if I should have succeeded in giving my reader a tolerably correct idea of the old Indian structures • in North Central America and of their great variety. An exhaustive description of old Central American architectures, which should enter into all details and state their peculiar manner of forming settlements, must needs be preserved for future generations, and I shall be satisfied if I may state here the conclusions which the available material has enabled me to form. ' Recently also found iu Copan. ^A reconstruction will be found in Maudslay at the proper place, page 18. 552 INDIAN SETTLEMENTS IN CENTEAL AMERICA. 3. COMPREHENSIVE REMARKS xVND CONCLUSIONS. The old Indian edifices of northern Central America display in various parts of this territory an unusually great variety both as regards the disposition and the construction of the single buildings. Upon closer investigation we find, however, that the structures of special districts show certain peculiarities which are common to all of them, but are not found in the structures of adjoining districts. These common pecul- iarities, however, apply only to general features, whilst we never meet with slavish imitation of a definite style of architecture. On the con- trary, even within the boundary lines of a certain style there exists still an almost indefinite variety of disposition and outward formation among the Central American Indians. As the edifices in the neighbor- hood of a style of building often already show hints suggested by the peculiarities of a neighboring style, we may conclude from this that the Indians maintained among themselves a lively intercourse and pos- sessed a great capacity for acquiring knowledge and taste from their neighbors. Everywhere we find the fundamental idea of walls and terraced pyramids, but in their erection many varieties of style at once appear. Unfortunately, I must here limit myself to the structures on the highlands of Guatemala and Chiapas, and to those of East Guate- mala, Peten, Tabasco, and Yucatan. Within this district I think I may define the following styles of architecture : I. The steps of the pyramids and walls are ascending. The build- ings are not distinctly grouped around courtyards (squares). 1. Chiapas style : The buildings of a settlement are rather irregularly arranged. 2. Motozintla style: The buildings of a settlement show a tendency to face a certain direction. In front of many tumuli tile pavements are made. II. The steps of the pyramids and walks are perpendicularly ascend- ing. The buildings of a settlement face any one decided direction. The larger settlements show a part of their buildings arranged entirely or in part around an inclosed courtyard or square. ARCHITECTURAL STYLES OF THE MAYA PEOPLE. A. Yarapaz style. — The settlements are mostly small. The buildings face the four cardinal points. Mortar was but imperfectly used. In Chacujal stone buildings with perpendicular walls, parapets on the platform. B. Architectural styles of the highland tribes. — The settlements show a crowded disposition of their buildings. In the whole district temple courts shaped like H make their appearance. (rt) No mortar is used in the buildings. 1. Tzenal style: The buildings of a settlement are not arranged so as to face the cardinal points, but preferably intermediate directions. INDIAN SETTLEMENTS IN CENTRAL AMERICA. 553 (&) In many buildings mortar is used in erecting stone houses. 2. Mame style : The buildings of a settlement are generally made to face intermediate directions. 3. Quich6 style: The buildings of a settlement face the cardinal points. 0. Architectural styles of the lowland tribes. — In many buildings stone walls, cemented with mortar, are found. Stone houses with habitable inner rooms. The buildings mostly face the cardinal points. 1. Maya style: At times steep pyramids. The door beams made of zapote wood. la. Peten type : The buildings of a settlement are closely crowded. Formation of many courts (squares). Character of fortification. The walls show a casing of mortar. Mostly unadorned houses. lb. Type of South Yucatan : Transition type. The arrangement of buildings is less crowded. The walls of the stone houses are often incased in stone, carefully cut, but simple. Ic. Type of North Yucatan : The arrangement of the buildings is rather a scattered one. The outer walls of the stone houses are often richly adorned with sculptures. 2. Ghol style : The door openings are generally closed above with level slabs of stone. The ornamentation of stone houses consists in stucco ornaments or in tablets containing images or hieroglyphics. 3. Chorti style : Very peculiar development of the pyramidical struc- ture and of courts (squares). In Copan a steep pyramid. The stone houses of Tonina belong to the Choi style, while the other edifices and the general arrangement follow the Tzenal style. The ruins are now situated in the land of the Tzenal people, but not very far from the line, since the nearest Lacandon and Choi settlements are hardly 30 or 40 kilometers distant, and it can not be absolutely asserted that Tonina may not originally have been a Tonintl or a Choi settlement. However this may be. Toning haS' always shown a mixed style, at all events borrowing from a neighboring style, so that I do not feel justi fied, by the occurrence of a single instance, to attribute the existence of stone houses to the Tzenal style. The Indian edifices of northern Central America very frequently show a striking want of symmetry. The very simplest buildings, to be sure, are almost always symmetrical, because in their very great sim- plicity there was no scope for unsymmetrical arrangement. But the more varied single buildings and central structures, like great temples, show almost always an unequal development on the two sides from a middle line, and although the gradual development of architectural art led step by step to a better observance of symmetry, it is nevertheless curious to observe how, after all, only the very best edifices of Yucatan and Palenque ever attained unto full symmetry. It is true that fre- quently only mere trifles display a want of this kind, but on examining a building of this kind, or even a ground plan, we always feel as if this 554 INDIAN SETTLEMENTS IN CENTRAL AMERICA. unsymmetrictil arrangement were not so inucli the result of negligence as of a set purpose. And how capriciously even the inner rooms of the stone houses may be differently adorned on both sides of the entrance can be seen in the plan of a room in Tifal (fig. 15). Even the richly ornamented inner room of the principal temple of Menche displays in the different positions of the entrances relatively to the most external side apartments constant neglect of symmetry. (42) In this connection it may be noticed that the Indians of the Maya family showed the same tendency to an unsymmetrical cultivation of development of the individual members in their musical melodies also.i (13) These side apartments seem to have been used by the Lacondons mainly for their sacrifices, since it was in these rooms that I found, in 1891, most of their sacrificial vessels made of clay. All the tribes which belong to the Maya family^ have certain i^ecu- liarities in their manner of building, and it is of the utmost impor- tance that within the territory they at present occupy, and according to the limits of the knowledge which we at present possess, no buildings are known to us which betray a foreign style, except only the few at Motozintla, of which I have spoken above. This would justify the same conclusion which I also reached in studying the local names — that the Maya tribes have for a long time already occupied their present homes in northern Central America. A comparison of the architectural features which all the Maya tribes have in common suggests also a certain conclusion as to the degree to which their architectural skill had raised among them before the tribe sepaiated. This is a very low degree — walls and terraced x)yramids of moderate size, facing a fixed direction^ and frequently grouped around a court or square. It appears, however, as if the lowland tribes had already parted with the primitive Maya people at a time when the Verai^az tribes (the Poco- man group) were still in close contact with the highland tribes, since their straw huts (dwelling houses) are in their construction perfectly identical, while the lowland tribes differ in having an advanced wall. At the same time the Choi and Chorti Indians, dwelling near by, still adhere to the rectangular ground plan of the highland huts, while Chontals and Mayas abandon this type in favor of rounded-off ground, plans. While the Verapaz tribes thus remained on a lower grade of archi- tecture, the tribes of the highlands and the lowlands both developed the art in their own peculiar manner. Among the highland tribes, the Quich6 and Mam group made great progress in architecture, 'Compare the New Journal for Music, year XI (1895), Nos. 7 and 8, and year XIII (1892), Nos. 22 and 23. ^I have, unfortunately, here to omit the Huasteos, since I have uo information as to their buildings. K INDIAN SETTLEMENTS IN CENTRAL AMERICA. 555 which the Tzenal group, however, did not share, while the lowland tribes, each in its most original manner, attained high success in the same art, no doubt largely assisted by the favorable stone material which they found at hand. In like manner the peculiar qualities of the outcropping stone no doubt led in the Choi territory to making relievo works, in the Ohorti territory to monolithic sculpture, and iu northern Yucatan to a sculptural adornment of their houses. A very long time must, of course, have elapsed between the time when simxjle buildings were raised by the primitive Maya family and the days when the beautifully developed architecture of temples in Sajacabaja of original Copan pyramids, of excessively adorned stone houses in Yucatan, of defiant Ti«ul structiires, and of harmoniously /<;«<, composed and ornamented edifices in Palenque began to show itself. Hence, we assume with certainty that each one of the Maya tribes here mentioned may have occupied their present homes for a more or less extended period of time, and that their architecture was developed only within this time. The influence of the surrounding stone material on this architecture may here and there become ])erceptible, and the locally limited origin, as well as the locally varying development of this architecture, will clearly show that any influence of Asiatic styles of architecture is absolutely excluded. It is true that so far the study of the architectural ruins has given no clew to the original home and to possible former migrations of the Maya family. I can, therefore, here only express the wish and the hope that future and more extensive studies, made on a broader basis, may succeed in establishing the views here suggested more firmly in determining the exchange of culture by comparing the architecture of neighboring races, and iu thus providing a safe basis for prehistoric research.