Scattered across the Diquís Delta in southern Costa Rica, hundreds of carved stone spheres, some weighing as much as 16 tons, sit in arrangements that no living person can fully explain. The objects first entered the archaeological record in 1943 when Doris Stone published her survey of the Río Grande de Térraba floodplain. Two decades later, Samuel Kirkland Lothrop produced the most detailed catalog of their locations and dimensions through Harvard’s Peabody Museum. Yet neither researcher recovered the kind of field data that would settle the central question: why did pre-Columbian communities invest enormous labor in shaping near-perfect spheres from hard gabbro and granodiorite?
Why the Diquís spheres still resist explanation
The stone spheres are not simply curiosities. They represent one of the few major pre-Columbian artifact classes in the Americas for which no functional consensus exists among archaeologists. Proposals range from territorial markers to astronomical instruments to symbols of chiefly authority, but each hypothesis runs into the same obstacle: the original excavation records are thin. Doris Stone’s 1943 paper in American Antiquity documented the floodplain sites where spheres were found but did not include provenience logs, stratigraphic profiles, or associated artifact inventories that could tie the objects to a specific cultural practice.
Lothrop’s 1963 monograph, published as Volume 51 of the Papers of the Peabody Museum, advanced the record by mapping sphere locations across the delta and noting the consistency of their craftsmanship. His distribution maps remain the most cited spatial data for the spheres. Still, the monograph contains no raw measurement tables, no excavation field notes with soil contexts, and no recorded testimony from local communities about the objects’ role. The result is that the two foundational studies provide location data but almost nothing about meaning.
One testable idea has gained traction among spatial archaeologists. If the original find-spot coordinates logged by Stone and Lothrop were overlaid on modern lidar-derived topography using geographic information system tools, researchers could check whether the spheres cluster along sight-lines to specific horizon points, such as mountain passes or solstice sunrise positions. Non-random clustering would support an astronomical or territorial-boundary function. Random scatter would weaken both theories. The data needed for this test already exist in the two mid-century publications, but no published study has yet carried out the overlay with current elevation and azimuth models.
Stone and Lothrop: the two records that define what is known
Almost everything scholars can say with confidence about the spheres traces back to two documents separated by 20 years. Doris Stone’s survey, published in American Antiquity by the Society for American Archaeology, was the first peer-reviewed account to describe the spheres in their landscape setting along the Térraba floodplain. Stone recorded their presence alongside other artifacts but did not attempt a systematic catalog of sizes, materials, or spatial relationships.
Samuel Kirkland Lothrop, working under the auspices of the Peabody Museum of Archaeology and Ethnology at Harvard University, filled part of that gap. His 1963 monograph on the archaeology of the Diquís Delta remains the single most detailed published record. Lothrop documented sphere diameters, noted the types of stone used, and produced maps showing where individual spheres had been found or relocated. His work established that the spheres were not limited to a single site but spread across a wide area of the delta, suggesting they belonged to a regional tradition rather than a single settlement.
The strength of these two sources is also their limitation. Both were products of mid-20th-century field methods that prioritized surface survey and typological description over stratigraphic excavation. Neither author published the kind of contextual data, such as radiocarbon dates from associated deposits, ceramic sequences keyed to sphere placements, or soil chemistry profiles, that modern archaeologists would require to assign function or chronology with confidence. No institutional update or new primary survey building on Lothrop’s spatial data has appeared in the sources available through the Cambridge Core system or the Peabody Museum’s publication record.
Gaps in the record and what a GIS test could settle
The absence of post-Lothrop primary fieldwork leaves several questions open. First, how many spheres survive in their original positions? Banana plantation clearing in the mid-20th century displaced many spheres, and some were moved to private gardens, government buildings, and museums. Without a modern GPS-based inventory, the number of in-situ spheres is unknown. Second, what dates bracket the tradition? Estimates based on associated pottery styles range across several centuries, but no direct dating of the spheres themselves has been published in the available record. Third, did the spheres serve a single purpose or multiple purposes across time? The size range alone, from small enough to lift by hand to specimens too massive to move without rollers and ramps, hints at varied uses that could have shifted with changing social needs.
A carefully designed GIS study could address at least part of this uncertainty. Researchers could begin by georeferencing the maps in Stone’s article and Lothrop’s monograph, assigning approximate coordinates to each recorded sphere. These points could then be adjusted using known landmarks and modern satellite imagery to refine their positions. Even if some spheres have since been relocated, the original mapped locations would still capture the pattern that existed when the delta was first surveyed.
Once a working coordinate set is established, analysts could overlay it on high-resolution digital elevation models derived from lidar. From there, a series of tests becomes possible. Line-of-sight analyses could determine whether groups of spheres align with prominent peaks, river bends, or coastal features. Viewshed models could show what portions of the landscape would have been visible from each sphere cluster, testing the idea that they marked boundaries or controlled routes. Circular statistics could be applied to any detected alignments to evaluate whether they differ significantly from what random placement would produce.
Such a project would not, by itself, reveal the cosmological beliefs or political structures of the people who carved the spheres. It would, however, move the debate from speculation to falsifiable propositions. If no meaningful pattern emerges from the spatial analysis, then functions that depend on precise placement, such as horizon astronomy, become less likely. If strong clustering along particular azimuths or landscape features appears, those models gain empirical support and can be refined with targeted excavation.
What new fieldwork would need to look like
GIS analysis, though powerful, cannot replace excavation. To move beyond pattern recognition, archaeologists would need to reopen some of the sphere sites under modern standards. That would mean controlled stratigraphic digging around spheres that remain in place, with systematic sampling for radiocarbon dating, microbotanical remains, and soil chemistry. Detailed recording of associated ceramics, lithics, and organic materials could anchor the spheres within local cultural sequences and clarify whether they were erected in single episodes or revisited over generations.
Equally important would be documenting the negative evidence: places where spheres are absent. Comparing sphere-bearing sites with nearby locations that share similar topography but lack spheres could reveal social or ritual criteria that pure environmental analysis would miss. For instance, if spheres consistently appear near certain types of architecture, burial grounds, or water features, that association might point toward commemorative or ceremonial roles.
Any new campaign would also have to account for the spheres’ modern histories. Interviews with local communities, landowners, and workers on former plantation lands could help reconstruct movements of spheres over the last century. These oral histories would not substitute for archaeological context, but they could distinguish long-standing placements from recent relocations and identify spheres that have been removed altogether.
Accessing and extending the legacy data
A practical challenge for any future project is simply assembling the legacy documentation. Stone’s article is accessible through the American Antiquity holdings on Cambridge Core support, while Lothrop’s monograph is distributed through the Peabody Museum’s publication series. Digitizing their maps at high resolution, extracting coordinate information, and reconciling discrepancies between the two sources would be a substantial but manageable first step.
From there, an open-access database of sphere locations, dimensions, and contexts could be built to serve both researchers and local stewards. Such a resource would not only facilitate GIS-based tests but also help track conservation status, document vandalism or loss, and support heritage initiatives in the Diquís region. By tying each data point back to its archival source, the database would make explicit the degrees of uncertainty involved-whether a sphere’s location is known from a precise survey, a sketch map, or a secondhand report.
For now, the Diquís stone spheres remain among the most enigmatic monuments in the pre-Columbian Americas. Their makers invested extraordinary skill in creating geometric perfection from resistant stone, yet left no written explanation and few contextual clues. The mid-20th-century surveys that first brought them to scholarly attention answered basic questions of presence and distribution but stopped short of the analytical depth that current methods can offer. Closing that gap will require both a return to the delta with new tools and a patient reworking of the records left by Stone and Lothrop. Only then will the spheres begin to shift from archaeological mystery toward historically grounded understanding.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
