The chipped stone tools recovered from the Maya site of San Pedro on Ambergris Caye, Belize, provide one of the only glimpses into Late Postclassic–Early Spanish Colonial (ca. AD 1400–1700) lithic technology along the Caribbean coast. In this report, we focus on the chert and chalcedony assemblage in terms of raw materials, tool types, and tool production. We also include a discussion of reduction strategies, raw material conservation, tool curation/recycling, and use-wear. This analysis of these chipped stone tools provides a more detailed understanding of household lithic economy during a time of significant cultural change.
San Pedro
The San Pedro site lies near the southern end of Ambergris Caye on the grounds of the Parham Plaza Hotel immediately west of the town park in the central part of San Pedro Town on town lots 991 and 992. Despite several seasons of investigation, the full areal extent of the San Pedro site has not yet been determined. Geomorphological studies by Dunn and Mazzullo (Reference Dunn and Mazzullo1993:129–130) found that south of the hotel property—in the area of the Paradise Hotel and the Rosalita site—a layer of anthrosols with abundant Maya artifacts was overlain by approximately 1 m of modern beach ridge sediments. It is possible then that the San Pedro site extends along the beach ridge to the south, but it is also possible that these finds represent part of a separate archaeological site or several discrete sites. Salvage excavations at six different locations spread throughout San Pedro produced Maya cultural materials, and human remains were recovered at each one (Stemp Reference Stemp2001; Weinberg et al. Reference Weinberg, Graham and Pendergast2003). So, it is possible that either the San Pedro site is quite expansive in terms of its areal extent, or that multiple Maya sites lie beneath modern San Pedro Town. Future archaeological studies will be necessary to confirm this.
Archaeological investigations conducted in 1990 and 1993 (Graham and Pendergast Reference Graham and Pendergast1994; Pendergast and Graham Reference Pendergast and Graham1991) and in 2017 and 2023 (Simmons and Mayfield Reference Simmons and Mayfield2019; Simmons et al. Reference Simmons, Mayfield, James J., Courtney and W. James2025) have revealed that San Pedro was a small fishing community for no more than around two to two and a half centuries that was probably first settled by around AD 1350–1400. The earliest settlers may have been migrants from the Maya community at Marco Gonzalez, where elevated sea-level rise and siltation led to mangrove encroachment and eventual abandonment of the community (Dunn and Mazzullo Reference Dunn and Mazzullo1993; Graham et al. Reference Graham, Macphail, Turner, Crowther, Stegemann, Arroyo-Kalin, Duncan, Whittet, Rosique and Austin2017).
The remains of Late Postclassic Maya domestic structures consist of portions of hard packed sascab-like house floors, within which numerous subfloor pit features along with human burials were found. Some of these features held intact ceramic vessels, others were almost certainly house post holes, and some could have functioned as subfloor storage spaces. Superstructures likely consisted of walls made of wooden poles that supported a wood framed, thatch-covered roof. Reflooring episodes were noted in several individual house floors, and midden materials, including well-preserved faunal remains and abundant lithic and ceramic artifacts, were consistently found in association with these house floors (Simmons and Mayfield Reference Simmons and Mayfield2019; Simmons et al. Reference Simmons, Mayfield, James J., Courtney and W. James2025). Ceramic analyses revealed predominantly Late Postclassic types, including red-slipped Red Payil Group sherds, Unslipped Navula Group sherds, and Hubelna Unslipped sherds (Aimers Reference Aimers2019, Reference Aimers2024). The Spanish presence is primarily demonstrated by olive jar fragments (Simmons and Mayfield Reference Simmons and Mayfield2019; Simmons et al. Reference Simmons, Mayfield, James J., Courtney and W. James2025; Stemp Reference Stemp2001; Weinberg et al. Reference Weinberg, Graham and Pendergast2003).
The Late Postclassic–Early Spanish Colonial Chipped Stone Tool Assemblage
The lithics derive from excavations that took place in the early 1990s, 2017, and 2023 (Graham and Pendergast Reference Graham and Pendergast1994; Pendergast and Graham Reference Pendergast and Graham1991; Simmons and Mayfield Reference Simmons and Mayfield2019; Simmons et al. Reference Simmons, Mayfield, James J., Courtney and W. James2025). Although the majority of the artifacts were recovered from the property of the Parham family, small quantities came from the Alamilla’s, Elvi’s, and Nuñez’s properties. Given that Maya-produced material culture in Belize did not substantially change with the arrival of the Spaniards (e.g., Oland Reference Oland2014; Pendergast et al. Reference Pendergast, Jones and Graham1993; Simmons Reference Simmons2002), in addition to the significant postdepositional disturbance at the site, the lithics could not be reliably divided into discrete Late Postclassic and Early Spanish Colonial subassemblages and were thus treated as a single assemblage. In all, 1,460 chipped stone artifacts from Late Postclassic–Early Spanish Colonial contexts were recovered from San Pedro town (Stemp Reference Stemp2001, Reference Stemp2016, Reference Stemp, Scott and Mayfield2019, Reference Stemp2024). Unlike the majority of Maya sites, most of the assemblage consisted of obsidian (55.9%; n = 816) with the rest being chert, chalcedony, limestone, quartz, and slate (44.1%; n = 644; Table 1)—a ratio of 1.27:1.
Non-obsidian Chipped Stone Artifacts from Late Postclassic–Early Spanish Colonial San Pedro.
1 Total does not include three artifacts made from European flint (a gunflint and two chunks of ship ballast) and one chert strike-a-light.
Raw Material
Because no known chert and chalcedony sources exist on Ambergris Caye, all cryptocrystalline silicates were acquired from the mainland through trade and possibly direct procurement (Cornec Reference Cornec1985; Hult and Hester Reference Hult, Thomas, Guderjan and Garber1995). Four hundred and forty-six (69.3%) nonobsidian artifacts were visually sourced to the Northern Belize Chert-bearing Zone (NBCZ; Figure 1; Shafer and Hester Reference Shafer and Hester1983:521; Stemp Reference Stemp2001:26). There were 74 (11.5%) artifacts of cherts other than NBCZ material; however, their sources on the mainland of northern Belize are not specifically known (Stemp Reference Stemp2001, Reference Stemp2004). Likewise, chalcedony (4.8%; n = 31) was procured from the mainland, but specific source locations have not been identified (Stemp Reference Stemp2001, Reference Stemp2004). The rest of the chert and chalcedony (11.3%; n = 73) is too badly burnt or patinated to reliably assign the artifacts to a specific source; consequently, these artifacts were classified as “indeterminate.” A small amount of limestone (2.2%; n = 14) was used to make informal tools at San Pedro. Four (0.6%) pieces of quartz and two (0.3%) slate pieces, both from the mainland, were also found at San Pedro.
Map of Belize with sites mentioned in the text and the location of the Northern Belize Chert-bearing Zone (NBCZ).
Tool Types and Production
Seventy-six (11.8%) of the chipped stone artifacts were classified as formal tools (Table 1). Based on tool morphology and production technology, most of these formal tools are identical to those from the workshops at Colha (Shafer and Hester Reference Shafer and Hester1983). Almost half (n = 35) of the formal tools are thin bifaces, specifically lenticular and lozenge forms (Figure 2) like those from the Early to Middle Postclassic (ca. AD 900–1250) at Colha (Hester and Shafer Reference Hester and Shafer1991; Shafer and Hester Reference Shafer and Hester1983). Similar lenticular and lozenge bifaces are noted from other sites on Ambergris Caye, including San Juan, Ek Luum, and Marco Gonzalez (Hult and Hester Reference Hult, Thomas, Guderjan and Garber1995; Stemp Reference Stemp2001; Stemp and Graham Reference Stemp and Graham2006). As at these other sites, some of the thin bifaces from San Pedro were significantly resharpened (Figure 2, right). A smaller number of large bifaces, both oval (n = 8) and general-utility (n = 3) forms, were also found (Figure 3). Several of the bifaces have been heavily resharpened/repaired, as demonstrated by their reduced sizes and steeper edge angles resulting from additional flaking. Moreover, three of the oval bifaces and some of the thick biface fragments were recycled into hammerstones, demonstrated by significant edge crushing, which has been noted at other sites on the caye (Hult and Hester Reference Hult, Thomas, Guderjan and Garber1995; Stemp Reference Stemp2001). Also of note are four small side-notched projectile points made from blades (SSNP; Figure 4) like the Late Postclassic–Colonial period examples reported from other sites in Belize and Guatemala (Marino et al. Reference Marino, Martindale Johnson, Meissner and Debra2016; Meissner Reference Meissner2020; Oland Reference Oland2013; Simmons Reference Simmons1995, Reference Simmons2002).
Lozenge and lenticular chert bifaces from San Pedro. Note: heavily resharpened/reworked biface (right). (Color online)
Oval chert bifaces and biface fragment from San Pedro. Note: heavily resharpened/reworked biface (left) and crushed edges on biface from use as a hammerstone (right). (Color online)
Small side-notched obsidian (left) and chert (middle, right) points (SSNPs) from San Pedro. (Color online)
The informal tools from San Pedro mainly consist of hard-hammer flakes, including flake-blades and retouched flakes (45.7%; n = 294), although there is a significant quantity of bifacial thinning/retouch flakes (18.2%; n = 117) in the assemblage. Like some of the oval bifaces and biface fragments, six cores and core fragments and three blocky fragments were recycled into hammerstones. The ratio of noncortical to cortical hard-hammer flakes, flake-blades, and retouched flakes (1.94:1) indicates that most core reduction at San Pedro was the result of flaking from partially, but minimally, decorticated cores of NBCZ and other mainland cherts. The flakes, including flake-blades and retouched flakes, to cores ratios for NBCZ chert (13.43:1) and other mainland chert (9.25:1) support the interpretation that partially decorticated cores were brought to San Pedro but not all were flaked to exhaustion. This can be demonstrated by the recovery of a very large NBCZ chert core fragment weighing 1.7 kg and measuring 382.7 cm3 (Stemp Reference Stemp2001:42–43). It may be that the Maya were curating cores for flake production given that no local sources of chert exist on Ambergris Caye (see preceding text). The use of some cores and core fragments as hammerstones reveals recycling efforts as well. The high ratio of noncortical to cortical bifacial thinning/retouch flakes (7.36:1) and the low ratio of bifacial thinning/retouch flakes to bifaces (1.98:1) support the argument that the majority of the bifacial thinning flakes are likely repair/rejuvenation flakes rather than tool production flakes. This interpretation is bolstered by the lack of biface preforms and manufacturing failures, as well as edge damage and use-wear on the dorsal surfaces of some bifacial thinning flakes that indicate removal from used bifaces (Stemp Reference Stemp2001, Reference Stemp2004).
Discussion
Despite living on an offshore island with no naturally occurring cryptocrystalline silicate outcrops, it appears the San Pedro Maya had reasonably good access to chipped chert and chalcedony tools in the Late Postclassic–Early Spanish Colonial period and developed strategies to extend the overall use-life of the lithic materials they acquired from the mainland. The San Pedranos developed a few different procurement and use strategies. They relied on finished formal tools produced in the Colha workshops in the NBCZ (Hester and Shafer Reference Hester and Shafer1991; Shafer and Hester Reference Shafer and Hester1983). Once acquired, bifaces were heavily used, curated, and recycled as evidenced by significant biface resharpening (which also produced usable flakes), the conversion of some biface fragments into ad hoc flake cores, and the use of some bifaces and biface fragments as pounding tools or hammerstones. This pattern of tool curation has been noted at other sites on Ambergris Caye (Hult and Hester Reference Hult, Thomas, Guderjan and Garber1995; Stemp Reference Stemp2001; Stemp and Graham Reference Stemp and Graham2006). Some of the formal tools may have been scavenged from other sites on the caye. This is suspected because some of the tools, notably thin bifaces (lenticular and lozenge), are like those produced in the late facet Early/Middle Postclassic (ca. AD 850–1250) workshops at Colha (Shafer and Hester Reference Shafer and Hester1983). Moreover, there are no tool preforms/manufacturing failures and there is not enough NBCZ debitage to indicate formal tool production at San Pedro (Stemp Reference Stemp2001, Reference Stemp2004, Reference Stemp, Scott and Mayfield2019, Reference Stemp2024). Given the import of finished tools from Northern Belize and the heavy curation of these tools, it seems doubtful that the San Pedro Maya traded them with people off the caye. In addition to the acquisition of finished tools from Colha, the San Pedranos also acquired a relatively small quantity of nodules/cores of NBCZ and other mainland cherts, which they used for ad hoc or expedient flake production. It is also likely that the San Pedro Maya scavenged some debitage (flakes and blocky fragments) from other locations on the caye that were occupied prior to the Late Postclassic, but this is more difficult to determine based solely on the debitage (see Oland Reference Oland2013).
Use-wear analysis of bifacial thinning flakes from thin bifaces, including lenticular and lozenge forms, recovered from the site reveals that these tools served multiple functions. The San Pedro Maya used some of these finely made bifaces for both acquiring and processing fish and medium-to-large terrestrial and marine mammals, as well as working plants, shell, and wood (Stemp Reference Stemp2001, Reference Stemp2004). Faunal remains from middens associated with Maya house floors at the site yielded bones from marine mammals, such as manatee (Trichechus manatus), terrestrial mammals, such as deer (Odocoileus sp.), and marine reptiles, including sea turtles (Eretmochelys, Chelonia, Caretta, or Dermochelys sp.), as well as a variety of fish species. Based on tool fragments and bifacial thinning flakes, some oval and general-utility bifaces were used to chop/adze wood, dig in soil, and as hammerstones. Use-wear on chert and chalcedony hard-hammer flakes and obsidian blade segments reflect a variety of everyday tasks, such as fish and animal processing, working ceramics, plants, shell, and wood (Stemp Reference Stemp2001, Reference Stemp2004, Reference Stemp2016). Although use-wear analysis was not performed on the small side-notched projectile points, they may have been used for hunting birds, as well as reptiles and small mammals.
Conclusion
Overall, the San Pedranos developed a series of successful strategies to both acquire and maximize the use of their chipped chert and chalcedony tools in the Late Postclassic–Early Spanish Colonial period. When many communities on the mainland fell under the yoke of the Spaniards, the San Pedro Maya managed to survive and continue to contribute to a coastal trade network, in part, because they had stone tools (Stemp Reference Stemp2016). Lithic studies like this one are important because they provide information about the coastal Maya at a critical time of cultural transformation that is typically missing from Spanish accounts.
Acknowledgments
We thank Drs. Elizabeth Graham and David Pendergast for their work on Ambergris Caye and for sharing their insights into the Maya at San Pedro. Many thanks are owed to the Belize Institute of Archaeology for their support and permission to conduct our research on Ambergris Caye. We would also like to acknowledge the continued support of Ms. Maria Parham, Mr. Tim Hagerty, Mr. Mito Paz, Ms. Jan Brown, Reef Radio and Television, the San Pedro Sun, Ambergris Today, and the people of San Pedro Town.
Funding Statement
The lithic analysis was funded by the Philip F. Vineberg Graduate McGill Major Fellowship, the David Stewart Memorial Graduate McGill Major Fellowship, and two Keene State College Faculty Research Grants (2017, 2023) awarded to W. James Stemp.
Data Availability Statement
Raw data are available in Table 1. Additional data available upon request from W. James Stemp (jstemp@keene.edu).
Competing Interests
The authors declare none.
