A Zooarchaeological Analysis of Turtle Deposition at the Late Postclassic to Protohistoric Maya Site of La Punta, Mensabak, Mexico

Institutional Context and Project Overview

This study represents a Master’s thesis in Archaeology submitted by Áki Acamapichtli Kjarvalsson to the Department of Archaeology, Conservation and History within the Faculty of Humanities at the University of Oslo in Spring $2026$. The research was conducted under the supervision of David K. Wright and fulfills requirements for a Master's degree. The project was made possible through the support of Joel W. Palka, who provided the intellectual framework and access to the Mensabak research context. Collaboration included taxanomic assistance from Yael Sánchez, who provided faunal data and helped construct the subsample dataset, and review collaboration from Ana Brown. Field observations were supplemented by communications with Josué Gómez regarding the sites of La Punta, Los Olores, and El Mirador. The work adheres to ethical standards regarding the use of generative artificial intelligence, noting its use only for linguistic and stylistic revision rather than as a source of factual archaeological interpretation.

Introduction and Research Problem

The thesis explores the unusual density of turtle remains at the site of La Punta, located within the rainforest lakes of Chiapas, Mexico, near the landmark El Mirador Mountain. The central research question asks: what do the spatial, vertical, and taphonomic patterns of turtle remains at La Punta reveal about turtle deposition across five working stratigraphic horizons, and how do these patterns inform our understanding of Postclassic Maya ritual and consumption practices at Lake Mensabak? The study aims to move beyond treating faunal remains as mere subsistence debris, instead situating them within the dynamic socio-ecological ritual practice of the Late Postclassic ($1400$-$1520$ AD) to Protohistoric ($1520$-$1700$ AD) periods. The research is structured around spatial distribution, taphonomic markers (burning, modification, fragmentation), and vertical patterning across excavation levels. A distinct focus is placed on the potential for "ritual deposition," defined as repeated patterns of placement or treatment set apart from routine discard.

Mensabak as a Ritual Landscape

Mensabak is characterized as a long-lived Maya waterscape where lakes, cliffs, rock-shelters, and mountains governed habitation and ceremonial life. The region is positioned at a nexus of trade and pilgrimage linking Maya and Aztec centers. El Mirador Mountain, also known as Chak Aktun (Red Hollow [turtle] Mountain [cave with water]), is the "centre of gravity" for this landscape. It joined aquatic and terrestrial realms and is conceptually linked to the Maize God's emergence from a split turtle carapace in Maya iconography (e.g., Codex-style plate K1892). While the Classic period saw the decline of dynastic systems, the Postclassic at Mensabak represents a period of resilience, migration, and reorganization. La Punta was the largest settlement of this period at Lake Tzibana, acting as a major political, religious, and economic center. It likely served as a "threshold locality," possibly the first stop on the ceremonial route toward the summit of El Mirador. This landscape contains five major cliffs with rock art, such as Risco Mensabak, and numerous rock-shelter shrines containing bizzarely high amounts of ritual ceramics, copal residues, and human remains.

Methodological and Analytical Procedures

The dataset comprises $1,694$ turtle fragments recovered from $62$ archaeological contexts. Analysis centers on the Number of Identified Specimens (NISP) as the primary quantitative measure, as the highly fragmented nature of the material makes Minimum Number of Individuals (MNI) or Minimum Animal Units (MAU) unreliable. The turtle anatomy is classified into the dorsal carapace and ventral plastron, joined by the bridge. Taxonomic identification focuses on locally available species: Kinosternon leucostomum (Chak’ich), Trachemys venusta (K’anak), Staurotypus triporcatus (Led), and Testudinidae continental (Ak). The larger river turtle Dermatemys mawii (T’inak) is also present, suggesting possible transport or trade. A $3\%$ analytical subsample ($51$ valid specimens) was selected for detailed morphological and taphonomic examination. Stratigraphy is organized into five working horizons based on $10\,\text{cm}$ arbitrary excavation spits: Level $1$ (superficial/disturbed), Level $2$ (upper use surfaces), Level $3$ (main concentration horizon), Level $4$ (deeper deposits), and Level $5$ (oldest horizon).

Data Reconciliation and Comparative Baseline

A critical methodological step involved reconciling the turtle NISP with a total faunal NISP baseline compiled by Yael Sánchez. Discrepancies were noted, most notably in context LP/2/A/4/3, where the turtle count ($499$ or $423$) exceeds the recorded total animal bone count of $91$. Such incompatibilities are documented as limitations—possibly resulting from transcription errors, recovery biases, or the exclusion of turtle bone from general faunal counts—rather than being silently corrected. To maintain transparency, the thesis uses aggregated level-scale counts to reduce the impact of single-context inconsistencies. Comparison with Los Olores, a nearby functional node associated with specialized craft and metallurgical production (per Josué Gómez), serves as a comparative check to test if La Punta’s turtle density is unique within the local waterscape.

Quantitative Structure of the Assemblage

Statistical analysis reveals that La Punta dominates the record, accounting for $1,676$ fragments ($98.9\%$), versus only $18$ fragments ($1.1\%$) from Los Olores. Within La Punta, the distribution is highly uneven. Level $3$ is identified as the primary turtle-rich horizon, containing $1,011$ turtle fragments, representing $56.9\%$ of the total faunal NISP for that level. Even when applying conservative tests to account for the LP/2/A/4/3 discrepancy, Level $3$ remains the most significant horizon. Major concentrations occur in contexts such as LP/2/B/1/3 ($237$ fragments) and LP/1/E/3/2 ($198$ fragments). These high-NISP contexts are often associated with Late Postclassic and Protohistoric radiocarbon dates ranging from AD $1027$ to AD $1634$. In contrast, Los Olores lacks the dense accumulations seen at La Punta, suggesting functionally differentiated roles for the two sites.

Taphonomic and Qualitative Results

Taphonomic recording highlighted burning, fragmentation, and intentional modification. In the analytical subsample, $33$ out of $51$ specimens ($64.7\%$) showed thermal alteration. All examined specimens from La Punta in the subsample displayed a high frequency of burning, whereas the five specimens from Los Olores were unburned. Qualitative modifications include polishing, perforation, and working. Context LP/2/B/1/3 is particularly notable for combining a high NISP ($237$) with evidence of polishing and perforation. Context LP/1/C/2/3 demonstrates symbolic association, where $27$ turtle fragments were found alongside animal teeth beads, claws, deer antlers, and fish bone. Butchery marks are rare, with only one specimen (T01 from Los Olores) showing a clear gash or "corte de tajo." The high fragmentation (noted as less than one quarter of the original element for $85\%$ of specimens in the region by Kestle) is attributed to both cultural practices and tropical preservation conditions, including acidic soils and recovery through $5\,\text{mm}$ and $2\,\text{mm}$ sieves.

Interpretation of Faunal Patterning

The interpretation moves away from a false dichotomy between food and ritual, instead arguing for "overlapping practices." While turtles were dietary staples, their presence at La Punta—marked by high density, thermal alteration, and specific associations—points to ritualization. The study considers four models: domestic refuse, communal consumption (feasting), secondary dumping, and ritual deposition. The evidence best fits a model of structured ritualized deposition. Parallels are drawn to Lamanai, where turtles were dismembered, roasted, and potentially used as musical instruments. The concept of turtle shell idiophones (drums) is explored; though no complete instruments were found, the proximity of a ceramic whistle in Structure $1$ and the presence of copper bells suggest a landscape rich in ritual sound. These bells, often associated with liminal rituals or burials in Mayapán and western Mexico, support the idea of La Punta as a site for complex, multi-sensory performances involving food, fire, and sound.

Conclusion and Implications

The thesis concludes that turtle remains at La Punta are not merely refuse but are material evidence of a socio-ecological ritual system. The concentration in Level $3$ and the repeated taphonomic markers of burning and modification distinguish these deposits from ordinary household waste. La Punta served as a threshold within a sacred landscape, where humans engaged with aquatic resources and spiritual forces during a time of Postclassic resilience. While issues like the LP/2/A/4/3 count discrepancy and the exact function of modified shells (ornaments vs. instruments) remain unresolved, the study demonstrates that turtle remains are vital for understanding the material life of ritual. Future research is suggested in the form of use-wear analysis on polished fragments, ZooMS for refined taxonomy, and further excavation at Los Olores to clarify its functional relationship with the primary ritual center at La Punta.