HomeBiographyWritingsPhotosContactAncient Maya Settlement Patterns and Environment at Tikal, Guatemala: Implications for Subsistence Models Dennis Edward Puleston A Dissertation in Anthropology, University of Pennsylvania, 1973

Chapter Two: Field Procedure of the Sustaining Area Project


The Limitations of Aerial Survey
The Brechas: Backbone of the Ground Survey
Mapping Techniques


The task before us was really unknown. It was out intention to map as many square kilometers of terrain beyond the nine km2 area covered by the 1:2000 maps presented in Tikal Report 11 (Carr and Hazard 1961) as we could handle. The difficulties of survey in the Maya Lowlands, along with limitations on time, energy, and money forced us to devise new techniques for mapping and for ceramic survey.

The Limitations of Aerial Survey

The use of aerial photographs has become a mainstay of settlement pattern surveys all over the world in recent years. They wer essential to the surveys carried out in Viru Valley (Willey 1953), Teotihuacan (Sanders 1965, 1970; Millon 1967, 1970), Texcoco (Parsons 1971), the Ixtapalapa region (Blanton 1972), the Candelaria Basin (Siemens and Puleston 1972), and now Kaminaljuyu (Sanders, personal communication) [the valley survey they were used in is referred to by Sanders and Michels (1969:61)]. Photographmetric Techniques were applied to ecological and land use surveys at San Lorenzo (M. Coe 1969) but not to prehistoric settlement patterns. All these sites, apart from San lorenzo, are located in relatively open areas, in direct contrast to the southern lowlands, where the dense forest canopy stands more than 30 meters above ground level. It is this canopy which makes the use of aerial photographs impossible for detailed survey in this region.

Aviateca in Guatemala City being repaired prior to take off for Tikal

This limitation does not mean that aerial survey cannot be useful in other ways. The 25 and 50 meter contour intervals that can be derived from the aerial photographs and which are shown on the maps prepared by the Instituto Geografico Nacional were useful for roughly checking the 10 meter contours wer obtained from surface field data. In many specific instances it has provided valuable leads to the location of "minor ceremonial centers." The Tikal "Barringer Group," located on the Perdido sheet of Tikal Report 11, was discoverd in this way from an oil company helicaopter (E. M> Shook, personal communication). Careful study of aerial photographs by Earle Hindley of the FAO-FYDEP Project has located a number of probable sites in other parts of the Peten, some of which are being checked out on the ground by Ian Graham.

At least one artificial aguada has been recognized in aerial photographs and subsequently investigated on the ground, again by Ian Graham (personal communication). The causeways of Coba were first recognized in aerial photographs because of the difference in the vegetation that grew on them where they creossed the larger bajos (Madeira 1931). The ridged fields and canals of El Tigre in Campeche were discovered in a similar fashion (Siemens and Puleston 1972). Thjese instances, however, are exceptions. The identification of house ruins often only a foot or two high, through use of visible light aerial photography at least, is clearly out of the question.

In spite of this limitation, the possiblity of studying  settlement patterns through vegetation ddifferences does offer a certain potential. Certain tree species, including mahogany (Lamb 1966:37), can be recognized from the air at certain times of year as a result of change in leaf coloration. Survey has revealed, however, that the association of what is probably the best indicator species, the ramon (Brosimum alicastrum), with setlement is more general than specific might be revealed by an aerial photographic survey of ramon trees, the configuration of the settlement groupings, as well as the as the relative sizes and numbers of the individual structures could not be discerned.

In spite of the problems, the approximately 1:50, 000 scale "satellite sites" by means  fo their aguadas and the paths left by the oil company survey trails or brechas which are visible in some photographs. They were also used to check the orientations fo the survey strips by comparing the configurations of major transitional  zones in the vegetation such as occur at the edges of tintal  bajo. The survey itself, however, was conducted entirely on the gound with ground controls.

The Brechas: Backbone of the Ground Survey

Thus without aerial photographs for location, a groung level datum had to be used. For this, straight, narrow trails were used, hereafter called brechas. These were carefully measured and staked. Fortunately, the transit-controlled delimitation in 1963-64 of the 576 km2 Tikal National Park by FYDEP (La Empresa Nacional de Fomento y Desarrollo Economico de El Peter) left us with serviceable trails extending 12 kilometers north, south, east and west from the center of Tikal, and around the four sides of the Park. The existence of these brecha trails was a major factor and in the formulation of the original plan to undertake in the success the sustaining area survey. The brechas were about a meter wide and only slightly grown over, so that a minimal amount of labor was necessary to clear and stake them. [In the one or two years since the time of physical establishment of the park limits, the original stakes were already decayed beyond legibility in most cases.  Where they were legible, on the south brecha, they served as a useful check on our own measurements. By 1966, all but one or two stakes on the other brechas had completely decomposed.]

Stakes on each brecha datum were placed at 25-meter intervals and marked on both sides with the distance in kilometers and meters from the center of Tikal (i.e., 5+275, 5+300, 5+325, etc.). For the north border brecha, these measurements were made from the end of the north brecha, while for the Uaxactun brecha they began approximately at the center of the Ricketson housemound survey. Unfortunately we were unable to get a precise fix because the Uaxactun Aguada which served as our datum m point is incorrectly located on the Ricketson's map. Furthermore the distortions and inaccuracies of the Ricketson's' map are so gross that even if this mislocation could be taken into consideration, an accurate alignment correct for the entire map, would be impossible (see Puleston 1974).

The center of the Tikal National Park unfortunately bears little relation to the map of Tikal. The park center, which forms the beginning of the north, south, east, and west brecha, was represented by a wooden stake near the center of the Great Plaza which I located in 1965 but was never able to find again. Measurements taken at the time showed this stake to have a position 141 meters south and 325 meters west in Square 5D of the Tikal map. The beginning of the Uaxactun brecha is at a point 60 meters west and 87 meters south of an Aero Service Corporation Shoran station marked as "9". According to an inscription on the side of this monument, the station was erected first in 1959 and again in 1960 by Ing. E. Roeich L. for the Argus Petroleum Corp. The starting point of the brecha was determined in 1967 by measuring from the SE corner of the Uaxactun Aguada as mentioned above to the center of the Ricketson housemound survey( Ricketson and Ricketson 1937: Fig. 2).

Aviateca plane landing onto the airstrip in Tikal in 1968

The orientation of the brechas, except in the case of the Uaxactun brecha, followed that of the National Park as it was laid out. This turned out to lie rather enigmatically between true north and magnetic north, approximately 1o30' east of the N-S orientation of the maps of Tikal Report 11 (Carr and Hazard 1961).

At two points on the west brecha and one on the south, the brecha trail made by the surveyors shifted several meters to the left or right following a new course parallel to the old one. If one is going east to west on the west brecha, there is a displacement of 87 meters to the north at 5892 meters and two small displacements of 5 meters to the north at 6022 and 7220 meters. These displacements are shown on the plates. The location of the 4-5 meter displacement to the east on the south brecha was not recorded. The Uaxactun brecha was oriented to magnetic north, but this orientation too may be slightly off, as the transit we used did not line up exactly on a backsite when used in certain directions. Since this problem was discovered in the field and in order to keep the brecha straight, The error was kept constant by consistently setting up the tripod in the same way. The error involved is very small, amounting to less than a degree, and so is insignificant for the scale at which the survey is presented.

Mapping Techniques

Once the brechas had been cleared and staked, mapping could begin. Of the four radial brechas, all except the south began at the edge of the central nine km2 scale map so that the peripheral squares presented in the 1:6250 scale map of Tikal Report 11 (Carr and Hazard 1961) could be remapped. As can be seen by comparing the maps presented here with that map, the general distribution of the mounds and their numbers is in agreement.

The Evolution of a Method
The problem of how to get the best coverage was basically a problem of how to control and organize manpower. This stage of the work, which involved working out an effective procedure, was accomplished in 1965 with the help of Francis P. Bowles. We began by sending out reconnaissance teams controlled from the brecha. Our first attempts, however, with more than one assistant were disastrous. Not only was coverage inefficient, but our reconnaissance men, with no idea of how far they were from the brecha, wandered off into adjoining squares and sometime became lost. Once beyond earshot, which Was considerably less than 250 meters (the limits of our survey to either side of the brecha trails), even experienced woodsmen often became disoriented, and much time was lost while they circled back from as far as half a kilometer away, or in their enthusiasm followed a promising ridge far beyond the limits of the intended strip.

By this method, once a structure group was located, the reconnaissance man returned to the brechas and then, going out again cutting a trail perpendicular to the brecha, he tried to relocate the house platforms. This was often difficult and frequently our carefully laid out secondary brechas missed the intended feature and further searching had to be done to relocate the structure. Taping from the structure to the brecha on a compass line was more efficient but still time-consuming since new trails had to be cut for every group and we were never sure how complete the coverage of the survey area by our assistants really was. This technique was soon abandoned.

In an attempt to control our search teams, I assigned myself the task of pacing around the outside of each 250-meter square, keeping the reconnaissance men inside the square and reporting the features as they found them. Even with this arrangement, we often lost contact and frequently I had to back-track large distance to map platforms found or announced after I had  passed them. Coverage was still erratic. WE finally settled down to keeping the search teams within 25 or 50 meters, and rather than trying to cover a 250-meter square in one circuit, doing it in two or three. At this point we also began pacing distances to the brecha when we realized how close we could come to the measured distance.

Aviateca Plane that landed in Tikal and slid of the runway in 1966

How much were we missing? Experience soon revealed that most platforms 0.15 to 2 meters high could be spotted from a distance of 0-10, and sometimes up to 15 meters. At greater distances, even structures a meter high were effectively screened from view by the vegetation. A distance of 20-30 meters was often sufficient to obscure even the largest structures. Since most plaza groups are 15-20 meters wide, a distance of 25 meters was finally settled on as a good distance to maintain between the searchers (see detailed description of the techniques below). Any platform or plaza group located between the paths  of two searchers then had a good chance of being spotted by one or both searchers. This system worked well once we got going; and in a dry run on an already mapped portion of the Camp Quadrangle (Carr and Hazard 1961), we were able to pick up every structure as well as an unmapped chultun. Where we overlapped the outer perimeter of the 0.5 km2 blocks mapped by Gregerson, we were able to locate many new structures and chultuns. Thus it is likely that we missed relatively few structures. Discounting "invisible" and "barely visible" situations ("platforms" 0-0.15 m high), I would suggest that the completeness of the survey was in the range of 95% as far as structures are concerned. The same cannot be said for chultuns, however; and we undoubtedly missed many throughout the survey.

Once we were able to control our manpower by keeping searchers 25 meters to either side of the archaeologist doing the mapping, we found we were able to increase the size of our team as well as its efficiency. With a few other changes we eventually evolved the system described below, which was used with great success for the next three years.

The Description of Techniques
The mapping tean consisted of five individuals. At the beggining of a particular sweep, the team lined up on the brecha at 25-meter intervals, with the archaeologist-mapper at the center. The line faced into the 250-meter square to be mapped. The equipment carried by the archaeologist, who was the navigator and recorder for the team, included a Brunton compass, a clipboard on which a strip of waterproof Mylar was taped over a piece of graph paper, and a 9H pencil. At a sign the team would begin to advance into the forest, being careful to maintain approximately the 25-meter spacing. The orientation of the team as a whole was determined by the man at the center with the compass. The hand-held Brunton compass was found to be quite accurate if used carefully. Long sightings on distant trees brought the best results. The distance covered was also kept track of by the center man who counted his paces carefully. When a calculated distance of 250 meters had been covered and the team stood at the outer edge of the survey strip, all shifted down the edge 125 meters, turned back towards the brecha, and began the return sweep to complete the mapping of the 250-meter square. If all went well, the team members emerged on the central brecha on the correct count at numbered stakes 125 meters down from where they had gone in. Accuracy within 15 meters in each direction could be expected, and with experience our sweeps became more and more precise. Often the team emerged at exactly the correct point. During a sweep, the center man had the task of plotting rough contour lines so as to have a record of the lay of the land, the relative steepness of the hills, and the location and direction of arroyos. An interval of approximately five meters was used. These impressionistic contours were later tied into transit and altimeter readings taken on the central brecha, as will be explained further on. These contours were translated into line separated by 10-meter intervals on the final map. Major change in the vegetation were also recorded along with trails and chiclero camps when they were found. The amount of extra material of this nature we decided to record increased with each brecha.

When the remains of a chultun, a house ruin or a larger structure was spotted by someone of the team, a relayed cry of "monticulo" or "chultun" brought the entire team to a halt. The center man then recorded his paced position on the map with an "X," and after marking the ground and blazing a nearby tree, paced out perpendicular to the course in the direction of the indicated feature where the discoverer waited.

If the center man did not arrive directly at the feature, he paced until he was abreast of it and then turning 90 degrees, paced off the intervening distance. Lengths and widths of structures were paced. When the orientations of structures, particularly smaller ones, could not be determined, they were assumed to be oriented to a magnetic north-south axis. While mapping was going on, the discoverer of the group was sent out to scour the periphery of the group for chultuns and other small structures. The three other members of the team held their positions. Once a group was mapped, the center man returned to his position at the center of the team and all started forward again. If, at the end of a full sweep the team emerged more than 10 or 15 meters off, the most recently mapped groups were shifted proportionately to place them nearer their actual locations on the map. A constant rate of error was assumed in most cases.

Using this system a single team was able to map up to 375,000 m2 in a day, not counting the groups off the edges of the strip, which we mapped whenever possible. The completeness of coverage, with exception of chultuns, is certainly very near comparable to that of the map of the central nine km2 of Tikal. For the peripheral squares mapped by Gregerson and included on the 1:6250 scale map in Tikal Report 11 (Carr and Hazard 1961), we were able to make many corrections and additions. The accuracy of group locations, and structure orientations inevitably is not so precise but is partially compensated for by our use of a scale that is 2.5 times less than that used by Carr and Hazard. In certain instances, groups at the outer edges of each strip may have accumulated a total error in placement, relative to the center 9 km2 of Tikal, of as much as 20-25 meters. Tests of the technique in plane-table mapped areas on the east brecha indicated that error north and south was about equal to error east and west.


Go Back to Top of Page

PagesNotesPrefaceIntroductionCh 1Ch 2Ch 3Ch 4Ch 5Ch 6Ch 7Ch 8Ch 9ContactVoivode Multimedia