When Was the Earthwork Built?

Webster's initial assumption in 1970, influenced by the earlier speculations of Thompson and Pollock, was that defenses like Becán's were products of the breakup of Classic society in the eighth and ninth centuries. This idea turned out to be seriously wrong as excavation progressed and as Ball examined the recovered sherds each day. Our eventual conclusion that the earthwork was instead Late Preclassic was controversial in the early 1970s, and strangely still is for some. Webster is responsible for a certain chronological sloppiness because in his dissertation he allowed a bit of wiggle-room by putting construction somewhere in the second to the fifth centuries a.d. Ball was then still refining his original ceramic sequence, and in particular dividing the 550-year long Late Preclassic Pakluum complex into shorter facets. He concluded that most Preclassic material at Becán was from late and terminal Pakluum times, or roughly 100 b.c. to a.d. 200, and that the abundant (albeit buried) Preclassic structures, plazas, and other features at the site, including at least one large pyramid exceeding 10 m in height, were of similar age. So too were the ceramics from the most revealing stratigraphic contexts associated with the earthwork. Ball accordingly argued strongly for a Late Preclassic date of a.d. 100–200 for the defensive system (Ball and Andrews Reference Ball and Andrews V1978:14) and Webster adopted this reinterpretation, a possibility he had always envisioned (see the preface to the 1976 version of his dissertation and Webster Reference Webster and Adams1977).

When we did our research, the putatively Early Classic Tikal earthwork (Puleston and Callender Reference Puleston and Callender1967) was the only comparable feature, and it stimulated interest in Maya warfare, particularly possible conflicts between Tikal and neighboring Uaxactun.Footnote ¹⁰ Amidst the dearth of evidence for Preclassic warfare, our Becán earthwork date generated skepticism because it seemed unique. Today we have evidence for Preclassic defenses at Tintal, Xulnal, Holmul, El Mirador, Cival, Punta de Chimino, Macabilero, Muralla de Leon, Cerros, La Cuernavilla, and perhaps at Edzna, Los Naranjos, and Nixtun-Ch'ich’, among other places, so Becán is hardly exceptional.Footnote ¹¹ Not only that, but some of these early defenses are bigger than those built by the Maya during Classic or Terminal Classic times. If we had first argued that Becán's earthwork was Late Preclassic in, say, 1995, no one would have blinked an eye. Nevertheless, we shall see that archaeologists who worked at Becán in the 1980s, 1990s, and 2000s continue to assert, echoing Pollock, that the earthwork is just a gigantic Terminal Classic borrow pit or reservoir, but not on the basis of any new evidence.

We have many clues about chronology, but the most important ones derive from architectural stratigraphy (see Webster [Reference Webster1976] and Ball [Reference Ball and Adams1977b] for details). The principal “built” feature of the earthwork is the imposing inner embankment, composed of material thrown up out of the ditch and only along its inner edge (Figures 4 and 5), and in some places directly associated with datable structures.

Figure 4. Section through the embankment near Structure XXVII showing white sascab with basal streaks of old topsoil, but no large limestone fragments. Note the beautifully preserved floor in front of the building, which was obviously very clean when buried by the embankment. Photograph by the authors.

Figure 5. Section through the embankment near Causeway 1. Image by the authors.

Structure XXV

Most of the embankment fill consists of white, or sometimes grey-brown, sascab with scattered dark streaks of old topsoil in the lower levels. Fortunately for us, this fill covered up an earlier building and built surfaces, and elsewhere was leveled off to serve as foundations for later ones. We examined two such buildings along the northwestern segment of the embankment. The first was Structure XXV (Figures 1 and 6), a 50-m long, well-preserved range structure built atop a flattened-off section of the embankment just behind the imposing palace Structure XIII. Structure XXV is firmly of Late Classic date and is itself underlain by floors dating from a.d. 350 to 650 that were also built atop the old embankment deposits. The latest possible date for the earthwork is thus about a.d. 350.Footnote ¹²

Figure 6. Structure XXV during excavation (left) and schematic drawing of its position in relation to the ditch and embankment. Photograph and image by the authors.

Structure XXVII

Even better information came from the small Structure XXVII, whose rear wall Webster noticed eroding from the outer slope of the embankment between Causeways 4 and 5 (Figure 1). Although the structure begged for excavation, Adams, codirector of our project, disagreed. We both fondly remember a very public argument Webster had with Adams, who finally admitted that Webster was right, but only from the rational point of view! The rational view won out, and we jointly excavated Structure XXVII. We exposed several buried plaster floors beneath a little Late or Terminal Preclassic (Pakluum) building and its associated plaza surface (Figures 7 and 8). Structure XXVII had been carefully covered with the sterile sascab of the embankment after the Maya removed its roof and upper walls. Although later damaged by root action, much of the building was beautifully preserved, especially its little front façade that retained extensive patches of a single coat of fine, red-painted plaster, and a smooth, almost pristine plaster floor that extended from it to the south.Footnote ¹³ Both floor and building had been scrupulously, almost reverently, cleaned before they were buried.

Figure 7. Structure XXVII and adjacent Preclassic floor. Photograph by the authors.

Figure 8. Schematic section through the center of Structure XXVII showing its position below the fill of the embankment and two floors. Image by the authors.

All of the identifiable sherds (n = 1,860) from the Operation 13 trenches were of Pakluum or earlier date. Contexts included the embankment deposits covering the building and its associated plaza floor, the fill of that floor and the earlier ones, and the middens beneath them. The rubble core of Structure XXVII produced only 31 sherds, but also an intact Pakluum Usulutan vessel, probably a dedicatory sub-floor cache. Subsequent 1973 excavations below the two floorings that abutted onto and ran beneath Structure XXVII yielded an additional 8,771 sherds, all again of mixed late and early Pakluum or earlier Acachen date. The upper Floor 1 (454 sherds) is emphatically of late Pakluum age, whereas Floor 2 (8,317 sherds) might have been of earlier Pakluum date with a subfloor core laid directly over and grading into a Middle Preclassic Acachen midden and soil surface immediately overlying bedrock.

Structures XXV and XXVII together create a kind of post quem ante quem stratigraphic sandwich consistent with an early date for the embankment, and by inference the earthwork as a whole. Think about this another way. Imagine that the earthwork had been built at either the beginning or the end of the Late Classic Bejuco ceramic phase (i.e., around a.d. 600 or 750), or even as late as the Terminal Classic. If so, Structure XXVII and its adjacent pristine floor somehow stood intact for at least 350 years without seriously deteriorating, without being altered, without being replastered, and without accumulating late sherds and other trash. This is wildly improbable. We reburied Structure XXVII in 1971 because leaving it exposed would have resulted in its rapid destruction.Footnote ¹⁴

Artifacts from the Earthwork Components

Other chronological clues came from causeway and embankment excavations and from debris in the ditch. All the causeways originally consisted of a hard limestone caprock underlain by much deeper and softer sascab (Figure 9). The upper sections of two of them had been purposely cut, possibly in the face of some threat, and later repaired. Fill from the Causeway 7 repair yielded a large (n = 463) collection of entirely Early Classic sherds (Chacsik and Sabucan phases). The upper three meters of Causeway 2 lacked the usual caprock, but instead consisted of fill contained by two retaining walls. Eighty-seven percent of the identifiable sherds from the fill were also Early Classic (late Chacsik) with a small admixture of earlier Pakluum material. Not a single Late Classic or later sherd was recovered from either causeway fill, so they both seem to have been repaired or modified before that time. Either Causeway 2 was originally intended to have this form and was an Early Classic construction, or more plausibly, a Late Preclassic causeway was cut and repaired during the Early Classic. Webster (Reference Webster1976:43) suggested that such causeway alterations were reactions to an Early Classic threat or attack.

Figure 9. The side of a natural limestone causeway sloping down from the inner edge of the ditch (right). (a) The hard caprock is being undermined by (b) erosion of the underlying softer sascab. Photograph by the authors.

We made several long cuts through the embankment, but most deposits were sterile. Only some 171 sherds were found sealed in the undisturbed fill or in the old topsoil beneath it. Of these, 170 were of Pakluum date and the other was Chacsik (Webster Reference Webster1976:86). It was primarily this Chacsik sherd that caused Webster to entertain the possibility of Early Classic construction. Ball (Ball and Andrews Reference Ball and Andrews V1978:14), however, later noted that several Chacsik types emerged during Pakluum times, and so this sherd could well be contemporary with the others.Footnote ¹⁵

In her general survey of Maya fortifications, Cortes (Reference Cortes2007:140) claims that the Becán earthwork must be Early Classic, primarily because she misreads the stratigraphy of Webster's embankment Trench 21 (Webster Reference Webster1976:36–37). She says that because the 19 sherds “from the retaining wall” he exposed along the inner side of the embankment were entirely Chacsik, the embankment must be of similar date.Footnote ¹⁶ These sherds actually came from a rubble layer above a floor that was clearly later than both the wall and the embankment, and the wall itself might long postdate the original earthwork construction. There were no sherds from the retaining wall, which is simply rough stone, or from the intact sascab fill behind it. In any case, as just noted, several Chacsik types have their origins in Late Preclassic times. The unfortunate scarcity of sherds in the embankment fill was disappointing, but also revealing in a negative sense. If a Terminal Classic ditch had been cut through ground already used by the Becán community for centuries, one would expect all sorts of earlier ceramic, architectural, and other debris to wind up in the embankment fill, but this is not the case anywhere we tested.

We recovered large sherd samples from trash thrown into the ditch, especially near the causeways, and also from some intact middens. Most sherds were mixed with sascab washed down from the embankment, and four of the causeways had been widened by purposeful infilling, probably using material from the adjacent embankment. Although most of these contexts are not sealed, the refuse patterns are clear. Ball found that some 86 percent of all the sherd material from the ditch (n = 11,465) dates to Late Classic Bejuco times, with earlier Pakluum, Chacsik, and Sabucan sherds mixed in as minority components. Apparently, the fortifications were kept clean and in good repair until about a.d. 600, and minor local modifications were probably made as well. After maintenance ceased, the sascab fill of the embankment eroded and the trash accumulated, ultimately burying the original ditch floor to a depth of several meters.

Two other points about the relationship between the earthwork and Becán's architecture bear noting. First, with the exception of Structure XXVII, we found no indications that the earthwork buried, destroyed, or overrode earlier structures. Webster (Reference Webster1978) later encountered just such contexts at the northern walled center of Cuca in Yucatan. Such superposition is what we would expect at Becán if the earthwork were a Late or Terminal Classic addition to an already well-established and thriving community, as were the little walls at Dos Pilas where such stratigraphy is common (see illustrations in Houston [Reference Houston1993]).Footnote ¹⁷ Second, except on the northwest, there are no large structures close to the earthwork along its inside perimeter, and only two within 100 m outside it. Apparently, the architecture was adjusted to the earthwork, rather than the earthwork to the architecture, and in any case adjacent buildings outside the ditch seem to long postdate construction of the defenses.

All these (and many other) things considered, we reiterate that the embankment, and by inference the earthwork as a whole, dates to sometime between a.d. 100 and 200 and was actively maintained until around a.d. 600. It was a source of building material after that time and was situationally altered to accommodate buildings or to widen causeways. Nevertheless, considerable defensive potential remained long after maintenance ceased. Anyone clambering around the earthwork today in its current state of disrepair will still encounter one of the most formidable defensive barriers in the central and southern Maya lowlands (Figure 10).

Figure 10. The relict earthwork as an obstacle in 1970. Photographs by the authors.

Water Management or Reservoir?

In 2005 the archaeologist Benevides (Reference Benevides2005:16) recycled the old idea that “Becán is particularly important for being surrounded by a ditch that gathered rain water and facilitated its use,” although he grudgingly admitted that the reservoir might also have come in handy during “war times.” Bueno Cano had earlier advanced this opinion in his 1999 overview and synthesis of Becán's archaeology, asserting that the ditch and parapet constituted no more than “un almacén para agua,” or reservoir (Bueno Cano Reference Bueno Cano1999:36). Campaña Valenzuela (Campaña Reference Campaña2005:49–50), who oversaw major consolidation and restoration at the site between 1999 and 2004, similarly proffered a number of unsupported assertions that the great ditch and parapet amounted to nothing more than an elaborate drainage system and giant borrow pit bordering commoner residential platforms, all possibly dating no earlier than the beginning of the Terminal Classic period (Bueno Cano Reference Bueno Cano1999).

These opinions were both puzzling and groundless, and recently Campaña (Reference Campaña and Castillo2014:141) doubled down, insisting again that the ditch was no more than a giant water-control drainage system and with no new supporting evidence other than expedient logic:

Let us be very clear about this water business. Despite its unfortunate name, the ditch was not formed by water, it is not a barranca in any sense of the word, it was not a water reservoir, nor was it any other kind of intentional water feature.Footnote ¹⁸ A short anecdote shows how the water management hypothesis has nevertheless long exercised a tenacious grip on some archaeological imaginations. In the mid-1990s Arqueologia Mexicana solicited an article on Becán from Webster (Webster Reference Webster1996). The editors requested reconstruction drawings of the earthwork, which he duly provided. When the article appeared in 1996, the composite illustration made from Webster's drawings of a dry ditch was curiously transmuted into a water-filled moat (Figure 11), despite the fact that he has always denied such an explanation and did not make this claim in the text of the article, in his submitted drawings, or anywhere else.Footnote ¹⁹

Figure 11. Webster's submitted images (left) and the mysteriously transformed published version.

Becán is located in relatively well-watered country, both in terms of rainfall and surface sources. Annual precipitation at the time of our work averaged around 1,500 mm (Vogeler Reference Vogeler and Adams1974) and in this part of Campeche generally ranges between 1,000–2,000 mm, with comparatively low yearly variation. Thomas (Reference Thomas1981:96) noted that the attractions of the locale for early settlers included the “well–drained terrain and abundant water of the nearby natural aguadas.” Large seasonal bajos border the site on the northeast and south and were natural deterrents to wet season attacks. These bajos are now much silted in, and they must have had more standing dry season water during earlier times. There is a small aguada within the fortifications, but the main water source when we worked there was Aguada Carmelita, about 500 meters south of the ruins and connected to them by a causeway (Figure 12). The aguada, at least partly artificial, is a rough square about 125 meters on a side and was clearly the main water source for Becan's inhabitants (Nicholas Dunning, personal communication 2021). Nevertheless, like most of the central and southern lowlands, the Becán region is prone to serious and unpredictable short droughts. Thomas (Reference Thomas1981:2) described the effects of droughts in 1972 and 1975 on the communities of newly-established migrant farmers around Becán. These people would probably have been forced to abandon their homes and relocate had trucks not delivered water from distant sources (we got most of the water for our camp the same way). That there were similar shortages in earlier times seems likely given new reconstructions of climatic history (see section Climate Change).

Figure 12. Flooded bajo to the south of Becán. Photograph by the authors.

The two main arguments against the water management explanation focus on engineering design and sedimentation. Thompson carefully read Ruppert's and Denison's description of the ditch, which said that its bottom in one section was some 16 feet higher than the adjacent swamp—the water source for a true moat. He also saw what appeared to be an inlet channel on the outer side of the ditch between Causeways 5 and 6. Because Thompson believed the moat was engineered to be flooded, he concluded that it was never completed because water could not have flowed around the whole ditch. Webster's later contour-mapping showed that Rupert's and Denison's calculations were correct. Although the ditch maintains its general cross-section around the perimeter of the site, its bottom on the northwest is seven to 10 m higher than on the southeast. Had water seeped into the ditch from the bajos to the northeast it would simply have puddled-up in the sections there. Another telling point is that during wet intervals the water in Aguada Carmelita is perched at a higher level than the original ditch bottom. Nevertheless, there are no sediments indicating any overflow or seepage onto the original floor of the adjacent southern ditch.

Two other obvious engineering objections were anticipated by Armillas. The causeways are simply solid rock left in place when the ditch segments were dug. They were never pierced to allow the flow of water because this was not the intention of the Maya, not because the moat system was unfinished. Armillas also correctly stressed the porousness of the sascab. Had the causeways been pierced to channel water, their sascab foundations would have dissolved, causing them to collapse.

As far as we know, Webster and his workmen are the only people who have recently stood on the original bottom of the ditch and examined its deposits (Figure 13). Any hole in the ground, dug for whatever purpose, in a tropical region that gets 1,000–2,000 mm of rain annually, admittedly accumulates water. That said, we encountered sediments indicative of significant water input and retention (e.g., dark clays and organic materials) only in the two northeastern segments of the ditch that are closely flanked by bajos. This part of the ditch, however, is still perfectly clear to anyone walking along it, as is the embankment. Although some water-deposited sediments accumulated in it, it was never silted up enough to obscure the outer or inner lips of the ditch—the presumed maximal level of the bajos. Had there been danger from flooding in this zone, the hydrological solution would have been to allow the water to drain off to the east and south, but solid causeway construction prevented such flow. The sediments we observed might, moreover, have been deposited long after the earthwork fell into disuse and the bajos had silted in. Even then, however, this northeastern segment of Becán's perimeter was never vulnerable to significant flooding, and we know of no flood deposits found anywhere inside the perimeter of the embankment. Neither, as far as we know, has anyone found drainage features like those at Tikal that funnel runoff from the interior architecture into the ditch segments to supply reservoirs. In short, there was never a water-filled barrier like those common in medieval Europe, although the ditch is often represented in that way despite our repeated protests. Whatever its incidental hydrological or other functions, the earthwork was intended as a dry ditch and embankment fortification pure and simple, despite the data-free assertions of Benevides, Bueno Cano, Campaña, and others.

Figure 13. Webster, Ignacio, and Gilberto on the floors of two deep ditch soundings. Photographs by the authors.

Finally, imagine for the moment that the ditch segments were purposely constructed reservoirs or flood control barriers. One could make the argument that drought and flood vulnerability support our Late Preclassic date. If flooding were a serious problem, why wait until Late or Terminal Classic times to control it? If, on the other hand, the purported ditch/reservoir was built that recently, how did earlier populations cope with water shortages or floods? We shall see shortly that there was a major regional dry spell during the Late Preclassic that would probably have lowered water levels in the bajos.

Borrow Pit?

None of the foregoing tells us much about Becán's Late Preclassic settlement or construction efforts, which are central to the borrow-pit hypothesis. There are some impressive Late Preclassic buildings, most notably Structure IX-sub and Structure E-IV-sub. More generally, “Pakluum structures, floors, and rubbish are distributed densely and consistently within the fortified zone and across the surrounding countryside” (Ball and Andrews Reference Ball and Andrews V1978:15). Later intensive reworking of the site destroyed or buried many insubstantial early buildings. An example is Structure XXII, a low (2.5 m) but extensive platform on a slight rise in the bedrock (Figure 10). Webster tested it with several large trenches and uncovered many fragmentary floors and walls. The locale was used well into Late Classic times, but his probes produced a high frequency (about 40 percent) of Late Preclassic sherds, probably scraped up locally and used as fill. Preclassic occupation in this vicinity helps explain the odd southern bulge in the earthwork between Causeways 7 and 1—it accommodated a prime early habitation spot. Most early constructions inside the ditch were generally modest but widely distributed, and Becán's Terminal Preclassic population might have compared favorably with its Late Classic one.

Our insistence on the primacy of fortification does not preclude other uses, and in this Campaña is half-right. The ditch was certainly a borrow pit, as Webster said in his dissertation. He calculated the volume of fill dug from the ditch at roughly 117,607 m³, equivalent to more than nine Tikal Temple 1s (Webster and Kirker Reference Webster and Kirker1995). This represents a mass probably greater than all the other structures ever built at Becán combined.Footnote ²⁰ Of this amount, about one-third (the volumetric equivalent of three Temple 1s) was in the form of hard limestone caprock. Because the Maya removed this material first, we expected to find it redeposited, along with old topsoil, to form the basal layers of the embankment. Instead the embankment fill, as shown above, consisted almost entirely of loose sascab with dark soil lenses at its base, so the Maya clearly hauled off the caprock to use elsewhere. We also saw that there are low sections of the parapet where the sascab fill was removed or leveled off, as beneath Structure XXV. Where did all this material go, and when did such reconfigurations take place?

Our conventional Becán map shown in Figure 1 is a kind of reverse palimpsest of early and late features, and it takes real mental effort to disentangle the visible surface architecture, which is Late Classic or later, from the much earlier earthworks.Footnote ²¹ In Figure 14 we eliminate it and show where there were some Late Preclassic buildings. Not depicted here are the “essentially unbroken expanses of sascab-cement plaza floors, terraces, platforms, and pyramidal structures—all of late through terminal facet Pakluum age” that were scattered widely over the intramural zone (Ball and Andrews Reference Ball and Andrews V1978:15; see also Bueno Cano Reference Bueno Cano1999). Becán's vigorous Late Preclassic settlement expansion required masses of hard fill to level off the outcrop, to build the first big formal plazas and structures, and to produce plaster to surface them. Caprock from the ditch was ideal for these purposes.

Figure 14. Locations of Pakluum construction; Pakluum debris encountered widely in the shaded area. Image by the authors.

Many Times or Intervals of Trouble

Consider the many potential times of trouble for Becán, both general and specific, that have been identified by research since 1970:

1. (1) A major Late Preclassic dry spell from about a.d. 100–300 (Dunning et al. Reference Dunning, Wahl, Beach, Jones, Luzzadder-Beach, McCane and Iannone2014; Ebert et. al Reference Ebert, May, Culleton, Awe and Kennett2017).

2. (2) Disruption of the great El Mirador polities at about a.d. 150–250, characterized by political collapse, new institutions of kingship, and also by population declines or diasporas.

3. (3) Pulses of general Teotihuacan-related influence in the Maya lowlands between a.d. 200–400. that roughly correlate with Ball's detection of an Early Classic decline at Becán at about a.d. 250–450.Footnote ²⁸

4. (4) The “arrival of strangers” at Tikal in a.d. 378 and the expansion of a Teotihuacan/Tikal political order.

5. (5) Archaeological evidence strongly suggestive of an attack on Becán during late Chacsik or initial Sabucan times, between ca. a.d. 400 and 460/500.

6. (6) The struggles between the Kaanu'l and Tikal hegemonies that began about a.d. 540 and lasted until about a.d. 750.

7. (7) Disruption related to volcanic events of the mid-sixth century, including the eruption of the Ilopango volcano in El Salvador (or maybe in Iceland).Footnote ²⁹

8. (8) Ball's (Reference Ball2014) recent detection of an interval of serious decline or outright abandonment at Becán between about a.d. 730/750 and 830, followed by repopulation by people using distinctive, northern-related pottery.

9. (9) The general turmoil of the southern Maya collapse between about a.d. 750 and 900.Footnote ³⁰

10. (10) The proposed great regional “megadroughts” between about a.d. 850 and 1100.

We can regard these episodes separately, but they resolve themselves into complex concatenations of vulnerability as graphed out in subsequent figures. The question is, how do we apportion potential construction, alteration, or extended use of the earthwork among all this potential instability? We surmised in 1970 that Becán might have suffered some sort of Teotihuacan-related threat or actual attack around a.d. 440–460, more or less in line with presumed conflict between Tikal and Uaxactun, and just before what then seemed to be a major Tikal hiatus. All this was too late for our proposed construction date.

The first two options (Figure 17) roughly correlate with what Estrada-Belli (Reference Estrada-Belli2011:61) identifies as a transformational “conjunction” episode in southern Maya culture history. They fit our construction date for the earthworks almost perfectly and lend credence to it. The “arrival of strangers” option, on the other hand, is too late (but see section A Testable Conjecture). We already saw that the earthwork was maintained and cleaned until about a.d. 600. During this extended interval of active use (Figure 18) there were several other local and regional threats (options 3–7).

Figure 17. Times or intervals of plausible instability or crisis at Becán juxtaposed with period of construction. Image by the authors.

Figure 18. Times or intervals of plausible instability or crisis at Becán juxtaposed with interval of active maintenance of the earthwork. Image by the authors.

Case Study: An Early Classic Attack on Becán During the Great Hegemonic Struggles

The big threat (options 4–6) that might be linked to violence at Becán was the inception of the Tikal/Snake dynasty wars around a.d. 540 (Martin Reference Martin2020; Martin and Grube Reference Martin and Grube1995, Reference Martin and Grube2008), which suspiciously coincides with estimates for the Ilopango eruption. Archaeological data supporting this attack on Becán has never been published, apart from short comments and a brief summary by Ball (Reference Ball1971:25–26, Reference Ball1977a:170–171, Reference Ball1979). Here is what we know.

In 1969 Simmons, Thomas, and Wright did preliminary excavations at Becán in anticipation of our 1970 season (Simmons et al. Reference Simmons, Thomas and Wright1969). Simmons cleared the impressive Structure XIV-sub (see Figure 1), a likely multiroom, Chacsik-phase palace that had been systematically in-filled and buried. Its replacement, the overlying Sabucan phase Structure XIV, was a large platform with probable talud-tablero elements. The superstructure, of unknown function, had been violently and comprehensively demolished. The well-known Teotihuacan-style terracotta host statuette with its interior complement of 10 Teotihuacan “cookie cutter” figurines was recovered from a cache in the rubble core of the Sabucan platform (Ball Reference Ball1974, Reference Ball1979), although its exact stratigraphic context is unclear. The inarguably Teotihuacan-manufactured ensemble was seated within an exceptionally fine, plano-relief carved, brown glossware tripod cylinder vase of southern lowlands origin. Glossware tripod cylinder vases and apron lids fragments are not found at Becán during Chacsik times, but are abundant in the lowermost Sabucan deposits and levels. Then they entirely disappear. Such wares are, of course, hallmarks of Teotihuacan influence during the southern lowland Tzakol 3 horizon of the Early Classic after a.d. 378, and well-known from literally all southern lowland sites. Several authors, including Braswell (in Volta et al. Reference Volta, Gunn, Folan, Braswell, Masson, Freidel and Demarest2020), have remarked in passing that these form(s) do not occur at Calakmul or Dzibanche, at least as far as is known to date from the published literature.

Two separate test pits, probably dug by Thomas and Wright, were situated to the immediate north and northeast of the north end of the Late Classic palace complex, collectively called “Structure XIII.” These two pits, whose contents were later analyzed by Ball and Rovner, tapped into an apparent early Sabucan destruction level strewn with human skeletal remains. Five bifacial broken dart or spear points were recovered from deposits of ambiguous Chacsik/Sabucan date from Structure XIV and from the two destruction-level pits. Four of the points were made of green obsidian, and one from a grey variety from an unidentified central Mexican source according to PIXE analysis. The precise dating and significance of these remains are open questions, but clues come from Becan's obsidian and ceramic assemblages.

Political Implications of Becán's Obsidian

Becán is deep in Snake country, about 70 kilometers northeast of Calakmul (Figure 19). We find it difficult to imagine that it was unaffected by the repeated Kaanu'l wars with Tikal, including those that occurred between a.d. 695 and 744 as Tikal regained dominance. A look at some venerable obsidian data raises some new ideas about this struggle.

Figure 19. Becán in Snakeland. Map by the authors.

A major but generally forgotten contribution to Maya studies made by the Tulane-Wisconsin Becán project was the elaboration of the first true lithic sequence anywhere in the Maya lowlands. It presented a contextually comprehensive study of the obsidian artifacts from the site, and a careful analysis of their chronological as well as their spatial and depositional distributions. The heavy lifting in this endeavor was done by Rovner (Reference Rovner and Adams1974, Reference Rovner1975), who joined the project in 1971 as part of his doctoral dissertation (Rovner Reference Rovner1975) research on the lithics of Dzibilchaltun, Becán, and other northern sites. Rovner and Ball worked together to attempt to tease out some tenable cultural-historical information from the Becán lithic collection. This proved more elusive than anticipated, but a number of intriguing findings did emerge.

Most pertinent to our rehabilitation of Becán is the chronological distribution of central Mexican green obsidian within the confines of the earthwork. Telltale clear green to specular gold-sheen obsidians were quantitatively abundant. They appeared in a range of forms from simple prismatic blades and blade segments to small leaf-shape bifaces and well-made tanged dart points (Ramirez Reference Ramirez2001; Rovner Reference Rovner1975; Rovner and Lewenstein Reference Rovner and Lewenstein1997). All this central Mexican green obsidian came from Chacsik and terminal facet Pakluum-Chacsik contexts. Until well into the Terminal Classic tenth century none (apart from the four points found in 1969) came from any contexts dating after about a.d.460/500 when the full Middle Classic Sabucan ceramic complex abruptly replaced Chacsik at Becán. To reiterate, other than in three redeposited instances of mixed Sabucan and earlier materials, central Mexican green obsidian was relatively abundant at Becán throughout the Early Classic Chacsik phase. It vanished completely and abruptly with or during Sabucan, the cultural phase that Ball originally associated s with a successful assault on and capture of Becán by Teotihuacan-affiliated Tikal/Petén forces sometime between a.d. 440 and 460 (Ball Reference Ball1977a, Reference Ball1979; Ramirez Reference Ramirez2001; Irwin Rovner, personal communications 1971–1974; Rovner Reference Rovner and Adams1974, Reference Rovner1975; Rovner and Lewenstein 1997).

The overall pattern is:

1. (1) Central Mexican green obsidian is well-represented at Becán from terminal facet Pakluum through late facet Chacsik times (ca. a.d. 150–400/460).

2. (2) Central Mexican green obsidian is also well-represented at Tikal during its Early Classic Manik phase (ca. a.d. 250–550), after first appearing in Terminal Preclassic contexts (Moholy-Nagy Reference Moholy-Nagy1999, Reference Moholy-Nagy2003; Moholy-Nagy et al. Reference Moholy-Nagy, Asaro and Stross1984, Reference Moholy-Nagy, James, Mark and Caleb2013).Footnote ³¹

3. (3) Green obsidian from any source is conspicuously absent from Calakmul throughout the full Classic period (Braswell Reference Braswell2013; Volta et al. Reference Volta, Gunn, Folan, Braswell, Masson, Freidel and Demarest2020), and clearly this commodity was unavailable to the rulers of the Kaanu'l polity, particularly during the Early Classic.

All this suggests direct interactions and commercial or other ties linking fortified Chacsik Becán to Tikal, rather than to the capitals of the Kaanu'l dynasty, Dzibanche and Calakmul, that flank it to the northeast and southwest. Braswell (in Volta et al. Reference Volta, Gunn, Folan, Braswell, Masson, Freidel and Demarest2020:359–360) believes that central Mexican “green” was not commodified at this early date, and argues that its presence reflects the existence of important political and/or social ties to Tikal and thence, indirectly, to Teotihuacan, rather than any commercial interactions. Obsidian from the presumably commodified El Chayal source in Guatemala is well-represented, even abundant, from the Late Preclassic through the entire Early Classic and the following Late and Terminal Classic at Becán, Tikal, and Calakmul. It shows no signs of having been in any way restricted or otherwise difficult to obtain at any of these centers throughout the entire Classic period (Braswell Reference Braswell2013; Ramirez Reference Ramirez2001; Rovner Reference Rovner1975; Volta et al. Reference Volta, Gunn, Folan, Braswell, Masson, Freidel and Demarest2020).

Two other southern Campeche-Petén centers, Uxul and Naachtun, lie well south of Becán between Calakmul and Tikal. They have also produced large quantities of central Mexican green obsidian, seemingly restricted to high status elite contexts in both cases. Braswell omits Becán in his modeling, which is a shame because it apparently fits the pattern of the third- through fifth-century, Tikal-linked, Calakmul-opposing centers that he identifies. From this perspective, the fortifications seem to reflect Becán's very precarious existence within the heart of Snake territory.

Implications of Sabucan-Phase Ceramics

Ball's (Reference Ball1973, Reference Ball1977a, Reference Ball1979) original idea of a probable north-central Petén military site-intrusion at Becán at the start of the Sabucan ceramic phase was based on the few usable ceramics studies then available. During the last three decades he has become increasingly uncomfortable with this assessment on several grounds. Chief among them was the above-mentioned sudden and complete disappearance of Mexican green obsidians coinciding with the appearance of the Sabucan ceramic assemblage (Tzakol 3-equivalent).

Sabucan deposits everywhere at Becán typically lay directly over and on Pakluum floors, sometimes with intervening soil horizons reflecting temporary abandonments. More often than not, the basal layers of these deposits constituted something of a “crush zone,” with late facet Chacsik sherds inextricably mixed with Sabucan materials. Teotihuacan-style tripod cylinder vases and associated apron lids in a variety of glosswares—brown, orange, black, and bichrome—appear abruptly with Sabucan, and are restricted to it (the cache vessel in Structure XIV is a famous example). These two companion forms are generally accepted to have been connected to elite storage, movement, and presentation of raw kakaw (cacao) beans or powdered cacao.

Fragments of such vases and lids were well-represented in virtually all our recorded Sabucan deposits, but they are not uniformly strewn throughout these vertically. They are instead restricted to the lowermost 10 cm of the deposits, effectively the basal layer or “crush zone” of Chacsik/Sabucan intermingling. This distribution might reflect a temporal faceting of Sabucan based on the content of the “crush zone,” with a Peténized early or “initial” Sabucan facet represented by the Chacsik-Sabucan “crush zone” and its indisputably Teotihuacanoid content, and a later facet or “full” Sabucan cleansed of such exotics and influences extending above it.

Within the thin basal, “initial” Sabucan crush layer, we in places find late facet Chacsik ceramic materials, Sabucan ceramics, green obsidian artifacts, tripod cylinders, apron lids, pieces of slate-backed pyrite mosaic mirrors, and also scattered fragments of human bone. Thicker “sanitized” Sabucan deposits continue above. We now believe that the “crush zone” represents both a relatively short-lived early or initial facet of Sabucan marked by intensified Tikal/Teotihuacan influence or presence, and the infusion of new ceramic and other artifactual elements symbolizing the latter's post-a.d. 378 association with Teotihuacan. Just when the Tikal/Teotihuacan political order extended to Becán is unknown, except that it occurred sometime between a.d. 400 and 562. The “sanitized,” regional late facet of Sabucan persisted, ultimately blending into the Tepeu 1 horizon Bejuco assemblage.

Other insights involve surface finish and form-mode characteristics of the Sabucan complex that seem inconsistent with what we know of north-central Petén pottery. Individual, decontextualized tomb vessels from Dzibanche and Calakmul appeared sporadically and randomly in museum exhibits and coffee table art books, and Ball was struck by their often close similarities with finer Sabucan complex ceramics. This raised the possibility that Becán's Early Classic attackers were Kaanu'l dynasty forces, not Petén-related enemies. We now think it likely that the obsidian and ceramic evidence reflect strong Tikal-Teotihuacan influence, but not any sort of attack and take-over. This view contravenes Ball's earlier suggestions of a fifth-century attack on Becán by Teotihuacan-affiliated Tikal/Petén forces, but older views must and should give way to newer data.

The difficulty of verifying our alternative scenario is the absence of full, or even partial, ceramic sequences for either Dzibanche or Calakmul, despite the prominence of both centers in the archaeological literature. For Calakmul there is a preliminary ceramic study by Domínguez Carrasco (Reference Domínguez Carrasco1994) based on excavations carried out by Folan for the Universidad Autónoma de Campeche in 1984–1985 and 1988–1989, but these were confined to Structures I, II, III, and VII. Her sequence is preliminary and not stratigraphically-based, but rather typologically so, as she makes clear. Careful study of Domínguez's sequence discloses additional problems. Her chronologically floating Early Classic Kaynikte complex clearly extends no later than ca. a.d. 500/550, while her similarly floating Late Classic Ku complex plainly begins no earlier than around a.d. 690/700, these dates based on typology and modal diagnostics in each case. This leaves the critical period between a.d. 500 and 690/700, the era of both Calakmul's ascendancy in the Maya world and the abrupt appearance of the Sabucan complex at Becán, an unrepresented and unknown blank. Trying to fill this gap by hit-and-miss hunting through exhibit catalogues and art books for the occasional, isolated funerary vessel is unconscionable as a research strategy, but the only one available given the continuing absence of robust, full-fledged, and widely-published ceramic sequences for either of the two key Kaanu'l centers.

Pending such new ceramic insights, we suggest that Becán's early Late Classic florescence between a.d. 590/610 and 730/750 reflects likely subjection to, or alliance with, the Calakmul-Dzibanche Kaanu'l polity and a subsequently more amicable or quiescent relationship with it.

The center's rapid failure and abandonment in the first half of the eighth century might readily reflect the collapsing fortunes of the Kaanu'l polity. Throughout most of its archaeologically documented history, Becán was a pivotal crossroad on the overland transpeninsular route between the Caribbean and the Gulf of Mexico (Ball Reference Ball1977a; Ball and Taschek Reference Ball, Taschek, Richard and Berlo1989; Rovner Reference Rovner and Adams1974, Reference Rovner1975; Rovner and Lewenstein 1997), and it is inconceivable that the weakening political power and economic influence of the giant center to the southwest had no serious reverberations on Becán's economic and political fortunes and viability. We submit that this was exactly the case, and that the defeats of Calakmul in the very late seventh and early eighth centuries had devastating consequences for Bejuco-era Becán.

A Testable Conjecture

We think the preceding information boils down to this new, plausible, and testable scenario:

1. (1) Becan's earthwork was originally built before a.d. 200 in reaction to some unknown Late Preclassic threat or crisis.

2. (2) Sometime well before the “arrival of strangers” at Tikal in a.d. 378, the latter had established a peaceful commercial or political presence at Becán, whose earthwork might already have been seen as a strategic asset.

3. (3) In the early to middle fifth century, sometime after the “arrival of strangers,” the influence and perhaps actual presence of the expanding Tikal/Teotihuacan political order at Becán intensified significantly, ushering in the distinctive Sabucan ceramic phase. It began before we have textual evidence of the Tikal/Snake wars, suggesting earlier, unrecorded antagonisms between the fledgling “superpowers.”

4. (4) The Becán-Petén alliance constituted both a direct tactical and strategic threat to the rising Snake rulers. Sometime late in the sixth century they successfully and violently attacked Becán and incorporated it into their local dependencies. Maintenance of the earthwork ceased, and the Snake lords could shift their base south to Calakmul without having a Petén outpost to their rear.

5. (5) The major decline or abandonment of Becán that Ball detected between about a.d. 730/750 and 830 coincided with the final defeat and unraveling of the Kaanu'l political order.

Post-hegemonic Becán

If the above scenario is true, the concurrent emergence of a large new center, the recently discovered Chactun, some 24 kilometers north-northwest of Becán (Šprajc Reference Šprajc2014, Reference Šprajc2021) seems more than coincidental. Chactun has conspicuously strong epigraphic and architectural ties to the north-central Petén, including sculpted stelae and altars replete with glyphic texts (Esparza Olguín Reference Esparza Olguín and Šprajc2015; Šprajc Reference Šprajc2015; Šprajc et al. Reference Šprajc, Olguín, Campiani, Esquivel, Marsetič and Ball2015). Among the 20 stelae and 15 altars recorded thus far, 11 carry remnants of sculpted figures and texts, including six readable dates corresponding to a.d. 731 and 751. At least two of these monuments, Stelae 1 and 12, were dedicated on 9.16.0.0.0 2 Ajaw 13 Sek (May 5, 751) by one Aj K'ihnich Balam. Chactun quite conceivably figured into Becán's demise somehow, but whether in a consequent, causal, exploitive, or purely incidental way is unknown, pending more substantive information about the site and its history. Surface sherds and the ceramic output of the small handful of test pits excavated at the site thus far shed little to no light on this dynamic (Ball Reference Ball and Šprajc2015, Reference Ball and Šprajc2017; Dzul Gongora Reference Dzul Gongora and Šprajc2020). To date, we have little evidence of the occupational history of Chactun or its smaller sister sites Tamchen and Lagunita prior to the Late Classic.

The Río Bec/Chenes Regional Florescence

To work out what happened at Becán between ca. a.d. 400 and 600 and in the mid-eighth century requires a fresh program of carefully conceived and focused investigations. The broad outline of events very probably followed the scenario laid out above. Ball believes that Becán's decline and the final defeat of Calakmul in the late seventh and eighth centuries prepared the way for a more decentralized tradition of dispersed elite groups and a thriving rural regional population. The impressive French and Slovenian surveys to the south, southeast, and north of Becán document this new settlement form (Nondédéo et al. Reference Nondédéo, Charlotte Arnauld and Michelet2013; Šprajc Reference Šprajc1996, Reference Šprajc1998, Reference Šprajc2021). Interestingly, to our knowledge they have identified no obtrusive local fortifications indicating serious warfare. Neither are there signs of defensive systems such as documented around Piedras Negras/Yaxchilan or the La Buenavista regions.Footnote ³² The earliest Becán-region inscriptions and iconography at a.d. 475 do depict themes of war and dynastic founding, so more fortifications might in the future be identified on recent lidar images.

In their own examination of the rise and florescence of the Río Bec regional centers, archaeologists of the 2002–2010 French Río Bec Project also considered the presence and likely impact of Calakmul and its fortunes (Nondédéo et al. Reference Nondédéo, Charlotte Arnauld and Michelet2013:50), but they arrived at a conclusion markedly different from our own (Taladoir et al. Reference Taladoire, Dzul, Nondédéo and Fornéc2013). The French believe that the Río Bec florescence and population boom beginning around a.d. 550/590 and lasting until 700 were largely independent of the florescence and decline of Calakmul and Dzibanche. We find this idea difficult to accept, as did Coe (Reference Coe2012:140), who thought it inconceivable that Becán and other Río Bec centers were not under Calakmul's sway. The French, in contrast, perceive little, if any social or economic impact arising from the florescence or failure of Calakmul on the emergence of Río Bec society and its distinctive settlement and organizational patterns, and instead emphasize a largely autochthonous origin for the zonal tradition.

However they were affected by the Tikal/Kaanu'l struggles, the Río Bec and Chenes regions seem to have thrived after the final decline of the Calakmul order, and the landscape south of Santa Rosa Xtampak is characterized by numerous small but impressive elite groups with no local dominant central place. What these sites share is one consistent and breathtaking architectonic/iconographic theme—building façades displaying enormous, imposing, sculpted serpentine “Monster Mouth” entry portals. In 2018, Stuart remarked to Webster that the obtrusive snake-mouth or Witz monster imagery so common on the façades of Late Classic Río Bec and Chenes buildings might hark back to the time when the whole region, Becán included, was under the sway of the Snake Lords of Dzibanche and Calakmul. At Becán such portals are preserved on the front façades of the mammoth early seventh-century, Bejuco-age Structure X and early Terminal Classic Structure IV (Potter Reference Potter1977). Iconographically distinct Itzam Yeh and/or Chac corner masks (e.g., on Chicanna Structure XX) are not known to occur in the area until the advent of Terminal Classic Xcocom ceramics and architecture. Just how this might relate to the late influx of northern people or influences detected by Ball is unclear. In the early to mid-ninth century the still well-populated southeastern Campeche region was infiltrated by peoples with strong northern ceramic affiliations (Ball Reference Ball1973, Reference Ball and Adams1977b; Ball and Taschek Reference Ball, Taschek and James J.2013, Reference Ball and Taschek2015; Thomas and Campbell Reference Thomas and Janice Campbell2008), and many abandoned standing buildings were selectively reoccupied by them (Thomas Reference Thomas1981). It was at this same time that the grand florescence of the distinctive twin-towers Río Bec style began, best represented by Becán Structure I (Bueno Cano Reference Bueno Cano1999; Michelet et al. Reference Michelet, Nondédéo, Patrois, Gillot and Gómez2013).

Along with the southeastern Campeche zone generally, Becán revived after its several decades of decline or abandonment and thrived mightily during the Terminal and Early Postclassic, experiencing an architectural “boom” beginning about a.d. 830/850. This is presumably one of the reasons why our Mexican colleagues prefer a late date for the earthwork—it provided the raw material for all this construction. Unexplained by their borrow-pit hypothesis are that two-thirds of the excavated material was clearly not used for this or other purposes, but simply left in place, and why almost 90 percent of all the ceramic material recovered from the ditch is of earlier Late Classic date (ca. a.d. 600–730/750). That the embankment was not ravaged for construction material suggests the defenses remained socially or militarily significant until very late. In any case, there was impressive ninth-century construction at Becán and much of its hinterland retained significant population right through a.d. 1050–1100.