This Atlas comprises a set of maps that display the locations of most ancient Maya archaeological sites currently known to science. At its heart are maps covering the whole of the Maya region with the archaeological sites marked on them, along with other archaeological features, such as causeways, fortifications, and inscriptions. The maps allow the reader to visualize the spatial relationships among the sites, the features, and the natural landscape.

We created the Atlas both for the general public and for scholars and scientists. The public’s fascination with ancient Maya civilization seems insatiable. We empathize with the allure so many feel because it long ago ensorcelled us. But unlike most, we have had the privilege of discovering new sites. We can testify that there are few emotions that can match the excitement and wonder of climbing a hill deep in the forest and seeing a temple or palace emerge for the first time through the dappled gloom. It is a pleasure well alloyed with melancholy, for one cannot but help to grieve for the lost genius and past loves buried among the fallen stones. We hope to share that exhilaration in some small measure, albeit vicariously, with the reader of this work.

We also designed this Atlas to serve scientists and scholars who need data such as these to build and test models of ancient Maya demography, political structure, economic systems, and social organization. We already share these data with colleagues routinely, and by publishing them we hope to inspire even more collaborations.

We have documented the locations of over 10,000 sites, large and small, during the 25 years we have been working on this project. To be clear, we did not discover the sites, nor did we visit all of them to record their locations. With few exceptions, we compiled the information from published sources, each of which has its own degree of precision, validity, and completeness. A minority of site locations do derive from personal observations, usually GPS waypoints, made by us or contributed by our colleagues.

Archaeology is quintessentially a spatial science because the spatial relationships we document—among sites, features, artifacts, and other remains—supply the crucial contextual information we use to infer ancient behavior and belief. Archaeologists are, understandably, obsessed with maps. In fact, when you clean up after a party, if all the cocktail napkins have maps sketched on them, you know archaeologists were there. The archaeological fascination with maps has a long history, and we present two earlier attempts to map ancient Maya sites. The first is by John Lloyd Stephens, the indefatigable and incomparable explorer who played a central role in rediscovering ancient Maya civilization. His maps date from the 1840s but are still not obsolete. A few years ago, a colleague of ours relocated one of the sites Stephens first documented, one which had remained unexplored for over 150 years, and undertook a remarkably fruitful project there. Without Stephens’s tantalizing description, the huge site would probably have remained nothing more than an obscure landmark for local farmers. The second is the 1940 edition of the Middle American Research Institute’s map of “Archaeological Sites in the Maya Area” (Kramer and Lowe 1940). Tulane’s 1940 map itself was a revised and enlarged version of Frans Blom and Oliver Ricketson’s 1924 “Ruins in the Maya Area”.

Our Atlas project began, as best as we can recall, in a late-night discussion in Walter’s driveway, after a party in which alcohol was served, in the late 1980s or early 1990s. In the warm and fragrant air of New Orleans, on a quiet residential street, we looked up at the stars made invisible by the city’s glow and discussed an apparent gap in the distribution of Maya archaeological sites. Cliff was trying to figure out whether there were any sites known south of Punto Put, which is the point, the proper location of which has long been disputed, where the borders of the states of Campeche, Quintana Roo, and Yucatán meet. Seen from low-Earth orbit, or on a large-scale map, the boundaries between the three Mexican states form a rough letter “Y”; Punto Put is the spot in the center where the three lines, and the states, meet. The question was whether there were sites in a broad north-south strip along the border between Campeche and Quintana Roo, south of Put and north of the east-west highway between Escársega, Campeche, and Chetumal, Quintana Roo. He had already consulted the obvious (to him) sources, such as the existing archaeological atlases of the three states (Garza and Kurjack 1980; Müller 1959, 1960), Ruppert and Denison’s (1943) survey, and other survey and excavation reports from surrounding areas (e.g., Cortes de Brásdefer 1984, Pollock 1980; Potter 1977, Thomas 1981). They showed no sites south of Put and, roughly, north of Zoh Laguna. Therefore, the question became whether there really were no sites or, instead, whether no sites were known, perhaps because of lack of archaeological reconnaissance and prospection. If the answer was that no sites were known but neither had anyone looked, then perhaps a survey would be worthwhile, especially because the same blank area on the map encompassed apparent divisions between Late and Terminal Classic period ceramic styles and regional styles of architecture, which needed investigating in themselves.

Walter suggested that the best way to address the issue would be to put all the data into a geographical information system (GIS) to allow us to visualize the spatial patterning of the sites. In theory, we could also map the boundaries of survey projects in the GIS to differentiate “nobody’s looked” from “someone looked and there’s definitely nothing there”. Plotting the surveys ultimately turned out to be impractical because many of the primary sources we relied upon, such as the atlases, were not proper surveys themselves, but rather compendia of data from varied sources. Nevertheless, Walter’s (superficially) simple idea stayed in our minds.

Although the thought lay dormant for a few years, we eventually resurrected it and decided to start building the GIS. We began with the archaeological atlas of Yucatán (Garza and Kurjack 1980) because it had Universal Transverse Mercator (UTM) coordinates for the sites, some of them from an early GPS receiver. We scanned the tables with the coordinates and used optical character recognition (OCR) software to turn them into letters and numbers. The procedure did not work well. The original font, the quality of the printing, the quality of the photocopy, the quality of the scan of the photocopy, and the abilities of the OCR software resulted in many errors. It is possible that the proofreading required to catch all the errors ended up being more work than typing the data by hand might have been.

Next, we added the archaeological atlases of Campeche and Quintana Roo (Müller 1959, 1960), which did not provide coordinates explicitly, by georeferencing the maps and then picking the site locations off them to create point coverages. That did not work too well either because the base maps in those works, which at best dated from the 1950s, seem to have been relatively inaccurate by current standards.

We cannot pinpoint exactly when we decided to expand the project to the rest of the Maya area. It seemed like a good idea at the time, but we cannot remember what we were thinking. We deny categorically that any psychoactive compounds influenced the decision.

Both authors contributed equally to the work, although in different and complementary ways. Walter’s vast experience in computing made him the natural leader in the GIS design and programming, while Cliff’s extensive knowledge of the literature allowed him to contribute more of the data. It was a labor of love, by which we mean that we had no external funding. Therefore, they both believe that any errors are the other guy’s fault.