VOLCANISM, ECOLqGY AND CULTURE: A REASSESSMENT OF THE VOLCAN ILOPANGO TBJ ERUPTION IN THE SOUTHERN MAYA REALM

Robert A. Dull, John R. Southon, and Payson Sheets

The Tierra Blanca Joven (TB.!) eruption o(lhe 1I0pango caldera in central EI Salvador was one c!(lllC largest Holocene vol­ canic events in Central America, and its ecological and cultural impacts were felt throughout EI Salvador and adjoining areas olGuatemala and Honduras. Early radiocarbon measurements established a ca. AD. 260 ± 114 calendar date for the erup­ I4 tion. However, a reevaluation ofthe original C dates, in addition to new AMS I4C assays, shows that the TBJ eruption occurred at least a eentury and a hall later than originally estimated, The revised composite supports an Early Classic Period cal­ endar date for the eruption: I sigma = AD. 421(429)526; 2 sigma = AD. 408(429)536. A review olarchaeological settlement, ceramic, and radiocarbon evidence from sites throughout the area ofgreatest devastation reveals a large-scale demographic collapscfollowing the event. We believe that the population crash was caused both by the biophysical effects ol the eruption and by the resulting disarticnlation oOhe "Miral/ores" culrural-economic sphere. The affected areas (dEl Salvador and south­ eastern Guatemala did not completely recover umil the seventh cemury A.D.

La erupcion del Tierra Blanca .!oven (TB.!) de la caldera de lIopando en la porcion central de EI Salvador,fue uno de los eventos volcanicos mas grandes del Holoceno en Cemroamerica. EI Salvador y zonas limitrofes de Guatemala y Honduras resintieron el impacto ecologico y cultural de este evemo. Con base en los fechamientos de radiocarbono con que se comaba previameme la erupcion se dato hacia 260 + 114 d.C. Sin embargo, una reevaluacion de esasfechas ademas de otras reciemememe procesadas, muestran que la erupcion del TB.! ocurrio cuando menos siglo y medio despuCs de 10 que se habia calculado originalmeme, De esta manera, se fundamema la adscripei6n temporal de la eruption en el periodo Clasico Temprano: 1 sigma = 421 (429) 526 d.C.; 2 sigma =408 (429) 536 d.C. Una revisi6n de los asentamientos arqueol6gicos, la ceramica y lasfechas de radiocarbono de varios sitios en elarea que sufri61a mayor devastacion, en el occideme de EI Salvador y el sureste de Guatemala, evidencian un colapso demografico a gran escala despuCs del evcnto. La mayoria de las personas que habitaban en un area de 1000 km2 al occideme de la caldera deillopango -incluyendo el valle de San Salvador y partes del valle de Zapotitan- deben haber perecido instamaneamente a causa del impacto de materiales piroclasticos; quienes vivian mas al occideme, en la cuenca del Rio Paz y en la costa suroeste de EI Salvador, probablememe abandonaronlos pueblos a causa del colapso agricola, el hambre y las enfer­ medades. En el sureste de Guatemala, donde los efectos biolisicos de la erupcion nofueron tan severos, al parecer eI decremento poblacional se debio a la desarticulaci6n resultame de la esfera econ6mico-cultural Miraflores. Los desplazados pOl' el desastre que migraron hacia elnorte, posiblemel1te contribuyeron al crecimiento de la poblaci6n rural del valle de Guatemala y del cen­ tro urbano de Copan, Honduras. Las areas que resultaron atei'tadas porelevel1to volcanicoen EI Salvador)' el sureste de Guatemala lograron recuperarse hasta el siglo VII d,C.

he Neovolcanic axis of Central America tra­ has been the greatest ofany mainland country in the verses El Salvador along its entire length, and Americas (Daugherty 1969; Denevan 1992; Lovell T the country has experienced numerous vol­ and Lutz 1995; Wilkie and Ortega 1997). Unfortu­ canic eruptions in historic times and throughout the nately, the corollary to this agricultural success is that late Quaternary (Rose et al. 1999; Sapper 1925; innumerable human inhabitants ofEl Salvador have Simkin and Siebert 1994).Young volcanic soils have been either killed or displaced by the direct impacts supported tremendous agricultural productivity in ofvolcanic eruptions throughout the Holocene (Feld­ El Salvador (Stevens 1964), and correspondingly the man 1993; Larde y Larin 1957; Sheets 1971, 1979, population density during most ofthe late Holocene 1983a),

Robert A. Dull - Department of Geography, University of California, Berkeley, CA 94720 John R. Southon - Center for Accelerator Mass Spectrometry, Lawrence Livermore National Laboratory, Livermore, CA 94550 Payson Sheets - Department ofAnthropology, University of Colorado, Boulder, CO 80309

Latin American Antiquity, 12(1),2001, pp. 25-44 Copyright© 2001 by the Society for American Archaeology

25 26 LATIN AMERICAN ANTIQUITY [Vol. 12, No.1, 2001

Honduras Guatemala

E1 Salvador

Figure 1. TBJ tephra isopachs in central and western EI Salvador (adapted from Hart and Steen-Mcintyre 1983).

The greatest known Holocene eruption in EI Sal­ logical sites buried by this distinctive pumiceous vador, and one of the largest late Quaternary events tephradeposit have been discovered throughout cen­ in all of Central America, was the Tierra Blanca tral and western El Salvador (Amaroli and Dull 1999; Joven (TBJ) event. This Plinianeruption was roughly Sheets 1979). A comprehensive study of the TBJ the size of the 1991 Mt. Pinatubo eruption, ejecting eruption was initiated in the mid 1970s (Sheets a total volume of ~ 18 km"' DRE (dense rock equiv­ 1983a), including analyses of pre-TBJ soils and 2 alent), and blanketing an area ofat least 10,000 km archaeological sites, as well as tephra distribution, under more than 50 em of tephra (Hart and Steen­ petrography, and geochemistry. The Tierra Blanca McIntyre 1983; Rose et al. 1999; Simkin and Siebert .loven eruption was defined in this study as "the 1994). The large volume of tephra emitted by TBJ deposit resulting from the youngest major eruption eruption covered virtually the entire area from the of the Ilopango vent complex" (Hart and Steen­ Valley ofSan Salvador west to the international bor­ McIntyre 1983). der with Guatemala (Figure 1).1 The TBJ eruption had a major impact on the cul­ The importance of the TBJ tephra as a correlative tural development ofcentral and western El Salvador stratigraphic marker in archaeology was recognized in (Sharer 1974; Sheets 1971, 1979, 1984). During the the early twentieth century by the Salvadoran histo­ Early and Middle Formative periods, sedentary vil­ rian Jorge Larde, whoin 1917 discovered the "remains lages supported by maize-based agriculture appeared of an archaic civilization, much older than the Maya throughout the alluvial valleys and coastal plain of civilization" [translation by author] buried below the central and western El Salvador (Amaroli and Dull tierra blanca, or"white earth" (Larde 1926:154). Rec­ 1999; Demarest 1988; Sharer 1974; Sheets 1983a). ognizing that the tierra blanca tephra was distributed By the close of the Formative period and during the in a radial pattern around Lago Ilopango, both Larde "Protoclassic,,,2 western El Salvador was a densely and his American colleague Samuel Lothrop deduced populated region with strong cultural and economic that this tephra must have come from Volcan Ilopango ties to the Southern Maya Highlands and the Pacific (Larde 1926; Lothrop 1927). coast ofGuatemala (Brady et al. 1998; Demarestand Since the initial reporting of the tierra blanca Sharer 1986; Neff et al. 1999; Sharer 1974; Sheets tephra by Larde and Lothrop, numerous archaeo- 1984). Dull et a!.] VOLCANISM, ECOLOGY, AND CULTURE 27

The TBJ eruption abruptly tenninated the Late 2300 Formative/ "Protoclassic" cultural florescence in western El Salvador (Demarest 1988; Sharer 1978b; 2100 Sheets 1984). It is doubtful that anything (or any­ 2 :tl one) living in the -1000 km zone ofdevastation sur­ ~ vived. This area was subjected to pyroclastic flows :e'".... 1900 iU that uprooted, engulfed, and carbonized forest trees, (J t: even at distances of -25 km west of the vent. 2- ll. 1700 Although sites in eastern El Salvador such as Quelepa a:l emerged from the TBJ disaster relatively unscathed e (Andrews 1986:241), survivors in western El Sal­ .:;.'" u 1500 vador (e.g., the Rio paz drainage basin) were left with "'" a sterile, denuded landscape. This ash-choked region - A Previous dates could not support any significant agricultural pro­ 1300 duction for some decades after the eruption (Dull • This study 1998; Sheets 1979, 1981, 1983a). There is unequivocal evidence that the TBJ erup­ 1100 +-+-+-+--+--+--+--+--+-+-+-+-+-+--4 tion had a significant effect on the cultural trajectory Figure 2. Uncalibrated TBJ radiocarbon dates listed in of central and western El Salvador; more challeng­ Table 1 with one sigma errors. The two ellipses highlight the ing is gaining a clear understanding ofits role in the bimodal distribution of the dates. See text for discussion of accepted vs. rejected dates. social, political, and economic evolution of the greater Maya area (Demarest 1988; Henderson 1997; One sampling locality (Figure 1: SiteA) is located Schortman and Urban 1991; Sharer and Gifford 2 km NNW of Santa Tecla on the highway between 1970; Sheets 1981, 1987). Toward this end, more pre­ San Salvador and Santa Ana. At this locale, charred cise dating ofthe eruption is critical (Earnest 1999). plant remains are incorporated into a -6 m high The original suite of nine radiocarbon dates pumiceous pyroclastic flow, erupted from the yielded a composite 14C age of 1708 ± 114 years Ilopango vent during the TBJ eruption. A total of (A.D. 260 ± 114 cal yr RP.) for the TBJ eruption eight intact samples of carbonized wood material (Sheets 1983b). However, closer inspection ofthese (tree trunks, branches, twigs, etc.) were collected nine 14C dates reveals a bimodal distribution, with from four fresh exposures in the field and wrapped dates clustered respectively around 1950 and 1600 in tin foil for transportation. 14C yr. RP. (Figure 2). In the research reported here These carbonized samples retained the overall we reject the earlier cluster of dates in favor of the shapes of the original plant forms and also showed latter, and three new AMS dates obtained from plant clear microscopic wood-grain structures, but had material directly associated with the TBJ eruption cracked into hard, apparently well-preserved, char­ support this analysis. We conclude that the eruption coal blocks. They seemed at first sight to be ideal for occurred between A.D. 410 and 535 (2 sigma cali­ radiocarbon dating, but we found that in all cases the brated date), during the Early Classic period.3 Fur­ material was completely base-soluble (which we thennore, we suggest that the TBJ eruption caused speculate was somehow due to charring in the a large-scale demographic and economic collapse unusual chemical environment ofa pyroclastic flow). that was felt throughout the southern Maya Realm. Thus, in some sense the samples were not true char­ coal; more importantly, they could not be subjected Methods to a standard acid-base-acid pretreatment to remove Samples of organic material from two field sites in soil humics. Therefore, to minimize possible conta­ El Salvador were collected for AMS 14C dating in mination, we dated only the largest ofthese samples April 1998. The material sampled was chosen (acomplete section ofa fully carbonized 35 cmdiam­ because we believe these plants were living at or eter tree trunk). very shortly before the time ofthe TBJ eruption and We took a 2 cm3 block from 1cm below the cam­ that their 14C content reflected that of the contem­ bium of the trunk, scraped off the outer few mm porary atmosphere. from all sides of the block, and treated it with 1N 28 LATIN AMERICAN ANTIQUITY [Vol. 12, No.1, 2001

HCl at 80°C, followed by a five-minute base treat­ quoted calendar date (A.D. 260 ± 114) for the erup­ ment with 1N NaOH at 80°C, which dissolved about tion was obtained from the unweighted mean of 9 half the material. Two different fractions-(l) the conventional radiocarbon dates using the 1979 reacidified undissolved residue and (2) a fraction MASCA radiocarbon calibration (Michael and Klein reprecipitated from the base wash with 6N HCl­ 1979; Sheets 1983b). In assessing these 9 previously were independently washed, dried, combusted, con­ reported dates for the TBJ event, we chose to dis­ verted to graphite and assayed for 14C using standard card Hv5001 and 5004 (1450 ± 310 and 1925 ± 215 AMS techniques (CAMS 46574 and 46575). B.P.), which had very large uncertainties due to the The second sampling locale is a small lake -I km small size of the samples dated (M. Geyh, Hanover north of the town of Chalchuapa, some 75 km west 14C laboratory, personal communication 2000), not ofthe Ilopango caldera (Figure I: Site B). This site­ because they were in any sense "wrong" but simply Laguneta El Trapiche-was cored from an anchored due to the lack ofprecision. We also rejected the soil platform with a 3 cm-diametermodifiedLivingstone carbon date (Tx2324: 1940 ± 50 BP), because soil square rod piston coring device (Wright et al. 1984). humics typically have built-in reservoir ages ofsev­ Sediment cores were extruded into plastic lining and eral hundred years (Harrison et al. 1993; Vogel et aI. wrapped in tin foil and PVC pipe for transportation. 1990). In the laboratory, the core containing the TBJ Sharer (1978a) also reports a single date on char­ tephra was split and a .5 em sample of partially car­ coal from the TBJ stratum at Laguna Cuzcachapa, bonized sedge (Cyperaceae) peat was extracted from Chalchuapa (Lawn 1974). This date (P-1803, 1709 immediately below the TBJ tephra for AMS dating ± 62 B.P.) is well within 2 sigma of our new deter­ (CAMS 60527). Monocot leafmaterial from the peat minations (Figures 2, 3), but we consider it unreli­ was subjected to a standard acid-base-acid treatment able, since it is from a feature that is specifically and then assayed for 14c. ol3C was -27.2, indicative described as "deposition and redeposition ofcultural ofterrestrial or emergent aquatic plants that receive debris, outwash and volcanic ash" (Sharer 1978a:51, their carbon directly from the atmosphere, rather emphasis added). than subaquatic vegetation that might produce biased Five ofthe 6 remaining TBJ dates cluster around dates through photosynthesis ofgeologically derived 1600 14C yr B.P., but one is clearly older (Table 1, bicarbonate in "hard" water. Figure 3). Tx31 14 (1940 ± 50 BP) was measured on charcoal taken from a small carbonized tree in a TBJ Results and Discussion pyroclastic flow 3 km NNWofSantaTecla, very near the sampling locality ofCAMS 46574 and 46575. It Radiocarbon Ages appears to be an ideal sample from short-lived mate­ The two fractions ofthe SantaTecla tree-trunk sam­ rial well associated with the TBJtephra, and we have ple yielded uncalibrated 14C dates of 1590 ± 50 and no explanation why the date is so old. Nevertheless, 1610 ± 50B.P., which we combine for a best 14C age since it is so discrepant from the other five results for the sample of 1600 ± 35 B.P. Because the dated (which have a weighted mean of 1610 ± 30 B.P. in "charcoal" could not be subjected to a full chemical excellent agreement with our new data), we reject it 4 2 pretreatment, we cannot completely rule out the pos­ as an outlier (X / u = 6.5, u =5, P < .001). sibility that it contained some residual contaminat­ Calibrated Age ing humics. However, we believe any age bias is insignificant because the tree was located beneath The weighted mean ofthe remaining fi ve concordant three meters ofsorbent ash; the tree itselfclearly pro­ results and the dates for our two new samples yield vided the great majority of the base-soluble mater­ a 14C age of 1605 ± 20 B.P. for the TBJ event (Fig­ ial; and the two fractions dated gave virtually ure 2). Figure 3 shows calibrated age ranges obtained identical results. The Laguneta El Trapiche peat using the INTCAL98 tree ring 14C calibration (Stu­ macrofossils yielded a date of 1610 ± 50 B.P., which iver et al. 1998). We use the two sigma age range of concurs precisely with the Santa Tecla result. A.D. 408(429)536 as our estimate for the TBJ cal­ Previous dates for the TBJ eruption were mea­ endar age (Figure 3; Stuiver et al. 1998: method A), sured on acid-base-acid treated charcoal, or in one while noting that the probability distribution for the case on soil humics (Table I; Figure 2). A widely calibrated date is weighted toward the early part of Dull et al.] VOLCANISM, ECOLOGY, AND CULTURE 29

1750 .,.------......

1700

1650

0- m... III 1600 ~ () ...."it 1550 Sg~g~~~ c;O'O'\C\C'>.CVi OJ) ---- -< 1500 U

1450 -F'-'-...... ~...... i_'_'_'"...... ,~~4...... J 350 400 450 500 550 Calendar Year· A.D. cb ~ u ~~r-­ ('1 N N I I Figure 3. The INTCAL98 (Stuiver et al. 1998) calibration of the accepted composite 14C date for the TBJ eruption, 1605 ± 20 yr. BP. Stippled lines =1 sigma, solid lines =2 sigma. .= Note that the calibration also reflects the one sigma error of the calibration curve itself (~15 years).

.~ "'" ::; "'"::: this range (Stuiver et al. 1998: method B): A.D. ~ "'".= .; "'"..= 0 iZ· 0 415-476 (p .58), A.D. 477-530 (p .42) . U C) C) Q Q t; = = ~ ca c: C'c; ~ C':S ~ P .:2 ~ g 0 0 0 0 0 E; e u -= 2 ~ v @ 2 2 2 2 2 v "E w C) w w w cc; a Il) -g Refining the Calendar Age N ~ ..c: ..c: ..c: ..c: ..c: N c:; :;E ~ " (J " ,::::: 1) 0 ::: 'c u u u u "5 'c: ::: I;/) 0 0 Several lines ofrelated evidence may help to further ~ -€ t E.. ~ ~ w v u u "u "u ~ constrain the date of the TBJ eruption: ceramic evi­ dence, ice core records ofprehistoric volcanism, and tree-ring temperature records. A compelling argument for an early fifth-century TBJ eruption has recently been put forth on the basis of a study comparing central and western Salvado­ ran ceramic sequences to major regional sequences elsewhere in the southeastern Maya area (Earnest 1999).5 Earnest notes that although much of the Copan and Chalchuapa ceramic-assemblage sequences can be roughly matched, the correspon­ dence breaks down during Copan's Brujo facet of the Acbi phase at ca. A.D. 400. Earnest interprets the lack of early Acbi- and Esperanza-phase (Kami­ naljuyu) equivalents at Chalchuapa as evidence for an occupation hiatus. He estimates that the time period characterized by a lack of ceramic produc­ tion at Chalchuapa after the eruption lasted for roughly 100-150 years (1999:288). The archive ofprehistoric volcanism recorded in the GISP2 ice cores from Greenland provides a high­ resolution temporal record of large explosive north- 30 LATIN AMERICAN ANTIQUITY [Vol. 12, No.1, 2001

35 r------....., 30 E25 ""C, :. 20 >< ~ 15 ~ 10 III 5 o '----..11---- 405 415 425 435 445 455 465 475 485 495 505 515 525 535 Year AD

Figure 4. S04-Z volcanic acidity record for the 2 sigma TBJ age range as recorded in the GISPZ Greenland ice core (cour­ tesy of Greg Zielinski).

em hemisphere eruptions (Zielinski 2000). Based and Siebert 1994), especially those at low latitudes, on the original AD. 260 TBJ date, Zielinski and col­ result in planetary cooling for periods of 1-3 years leagues (Zielinski 1995; Zielinski et al. 1994) iden­ (Briffa 2000; Zielinski 2000). tified two prominent third-century A.D. sulfate The 1835 Coseguina eruption in Nicaragua concentration peaks as candidates for the TBJ acid­ (Williams 1952) provides some idea of the climate­ ity signal, but subsequent detailed chemical analy­ forcing potential ofNeotropical eruptions. Although ses oftephra from these strata indicated that they did of a lesser magnitude than the TBJ event, the not originate from Ilopango (G. Zielinski, personal Coseguina eruption (VEl = 5) was followed by sig­ communication 2000). Ourdating evidence suggests nificant summer cooling in the northern hemisphere that the most likely candidate for the TBJ tephra in from 1835-1837 (D'Arrigo and Jacoby 1999; Jones the GISP2 record is one ofthe seven volcanic events et al. 1995). No such volcanically induced cooling recorded betweenAD. 410 andAD. 535 (Figure 4). periods have yet been linked to Ilopango (VEl =6), Four of these events occurred during the 25-year but at least two northern hemisphere dendroclima­ period A.D. 415-440, in the interval during which tological records show pronounced cooling within the eruption is most likely to have occurred. The the early part of the fifth century. Marked tempera­ conclusive attribution ofone ofthese acidity spikes ture depressions-as evidenced by ring width and to Ilopango awaits future microprobe analyses ofthe wood density minima-are recorded in the Sierra composition ofthe tephra glass shards in the GISP2 Nevada, California between ca. A.D. 420 and 428 core for comparison with the TBJ tephra. (Scuderi 1990:76), and in a composite record from A potential global impact of a large explosive three high-latitude northern hemisphere sites6 atAD. volcanic eruption such as the TBJ is climate forcing 421,426, and 436 (Briffa 2000:89). (Bryson and Goodman 1980; Zielinski 2000). Although ice-core volcanic acidity peaks and tree­

Aerosols (most notably H2S04) formed by S02 ring temperature records are consistent with a cli­ injected into the stratosphere by volcanic eruptions matically effective ca. A.D. 420-435 volcanic can effectively increase stratospheric absorption and eruption, there is no "smoking gun" linking these reflection of incoming solar radiation. The result is events to Volcan Ilopango. Given the new TBJ 14C increasing temperatures in the stratosphere and results reported here, we propose that the TBJ erup­ decreasing temperatures in the troposphere. Global tion is a likely candidate, but these assessments are temperature depressions resulting from large explo­ clearly speculative pending further research. sive eruptions have been recorded in both instru­ Disaster and Recovery: mental (Angell and Korshover 1985) and tree-ring The TBI in Ecological Context records (Briffa 2000). Dendroclimatological prox­ ies ofpaleo-temperature show that most large erup­ The devastated area probably remained biologically tions, Volcanic Explosivity Index (VEl) >5 (Simkin impoverished and geologically unstable for decades 32 LATIN AMERICAN ANTIQUITY [Vol. 12, No.1, 2001 posed by thick tephra deposits. These adjustments land surrounding Paricutin that was covered in 50 include deep tap roots deployed to reach the under­ cm. oftephra was still uncultivable 35 years after the lying pre-eruption soil and wide networks ofadven­ eruption (Rees 1979:283). titious roots developed to overcome nutrient and Previous studies of the Ilopango TBJ eruption water deficiencies in porous tephra (Thornton allowed for some two centuries of weathering of its 2000: 1060). Cultivated annuals, such as maize, highly silicic tephra, vegetation succession, animal squash, and beans do not exhibit the same adapt­ recolonization, and formation of an A-horizon suf­ ability (Scarth 1999). The pollen record from ficient to support maize-based agriculture. Here we Chalchuapa (Dull 1998) shows that forest recovery revise the estimate slightly downward, but maintain preceded agricultural recovery in the Rio Paz basin, that agricultural recovery within the arearepresented and this is probably what happened throughout cen­ by the 50 cm TBJ tephra isopach (Figure 1) proba­ tral and western El Salvador. bly took a century or more. The most densely populated areas of Pre­ Reevaluating the Archaeological Record: The TBl columbian El Salvador were the interior valleys Tephra in Cultural Context (Sheets 1983a), and it is these areas that would have experienced the slowest recovery (Eggler 1963: The human populations most directly affected by 50--51; Rees 1979; Thorarinsson 1979: 139). The val­ the TBJ eruption were those living in the territory leys ofSan Salvador, Zapotitan, Rio Lempa (includ­ extending 100 km directly west of the Ilopango ing the Paraiso Valley) and Rio Paz (Figure 6) were caldera. However, the indirect effects on social, eco­ affected most severely. Tephra eroded from upland nomic, and political systems affected a much wider areas would have added to total tephra depths in area of Mesoamerica. The Ilopango TBJ eruption these valleys. Primary streams and their tributaries permanently severed the previously close cultural would have been overwhelmed by transient bedload and economic relationships that had existed from material, aggrading and depositing reworked tephra the Middle Preclassic to the Early Classic periods well outside the former stream channels. Areas that among peoples living in the southern Maya area were once relatively flat, valley agricultural lands extending from Kaminaljuyu to Chalchuapa and would have become hummocky and gullied, as the beyond. unconsolidated tephra was continually shifted and The decimation of human populations in central reworked by fluvial and aeolian processes (Rees El Salvador has been clearly documented (Sheets 1979). 1971, 1979, 1981, 1983a). It is not likely that any­ Incipient soil development in the valley bottom­ one survived in the San Salvador or ZapotitMexico thwarted repeated attempts to and McBirney 1979: 156-157). The carbonizedplant 8 reintroduce cultivation in that region : "Where the material found in this study at sampling site A (Fig­ ash lay 15 cm or more deep, the land was virtually ure 1) attests to the heat ofthe TBJpyroclastic flows, useless for over a decade for crops such as maize or even at a distance of 25 krn from the source vent. beans" (Scarth 1999:208). Much ofthe former farm- Other volcanic hazards that could have affected Dull et al.] VOLCANISM, ECOLOGY, AND CULTURE 31

Figure s. Photograph of a Guatemalan landscape near Voldn Santa Maria following the 1902 eruption. The photograph was taken by F.M.Cardenas and originally published by Sapper (1927). following the eruption. Volcanic debris would have of historic-period analogs. For example, the VEl 6 choked major waterways and the summer rains eruption ofVolcan Santa Maria, Guatemala can give would have reactivated deep tephra deposits, caus­ us some idea ofwhat the post-TBJ landscape looked ing ma<;sive post-eruption lahars (mudflows), such like (Figure 5).7 Perhaps the best-studied post-erup­ as those following the 1991 eruption ofMt. Pinatubo tion tropical ecosystem is that of Krakatau, Indone­ (Fisheretal. 1997:123-132; Scarth 1999:268-271). sia (Thornton 1996), a VEl 6 eruption (Simkin and Although it is possible that some trees might have Siebert 1994). Although a debate continues as to survived burial by >2 m of tephra (Antos and Zobel whether all life on the island was eradicated by the 1987:250; Thornton 2000), primary succession eruption, it is certain that deep tephra deposits pre­ would have ensued in most areas covered to this vented forest recovery until roughly 50 years after depth (Blong 1984:319; Eggler 1948); no plants the 1883 eruption (Thornton 1996:75). Similarly, 25 would have survived in the areas overrun by pyro­ years after the eruption of Coseguina, Nicaragua clastic flows (Thornton 1996). Slowly the region blanketed that volcano with deep tephra deposits, would have been colonized, first by algae and Squier (1860:536) reports, "The appearance of this mosses, then by pteridophytes (ferns and fern allies), mountain is now desolate beyond description. Not a monocots (grasses and sedges), and rudera1 dicot trace oflife appears upon its parched sides." Another herbs (Egg1er 1963; Thornton 1996). Finally, peren­ reconnaissance of Coseguina made by Karl Sapper nial shrubs and arboreal species would have in 1897,62 years after the eruption, found that for­ appeared, although initially only on steep slopes and est vegetation had recovered up to an elevation of along riparian zones where the thick ash deposits 750 m (Sapper 1925:74--75). were first removed by gravity and fluvial processes Forest recovery, however, does not equate to agri­ (Thornton 1996). cultural recovery (Ugolini and Zasoski 1979: 110). The ecological recovery ofareas with deep (100+ Many tree and shrub species are well equipped to em) tephra deposits following the TBJ eruption can deal with periodic tephra deposition, and can adjust probably be best understood when considered in light to overcome water deficiency and nutrient barriers Dull etal.] VOLCANISM, ECOLOGY, AND CULTURE 33

Tikal f:::,.

Mexico

Guatemala

Honduras

o, 50

t1lJ Volcfm llopango Miraflores-Providencia D ceramic sphere 1- -I fc:tfo~~~r:o~~1~;~~~~pse 6 Archaeological sites/regions

Figure 6. Map of archaeological sites mentioned in the text in relationship to the Mirallores-Providencia sphere (Demarest 1986) and the area depopulated as a result of the TBJ eruption. people in the immediate vicinity of the vent during denizens ofChalchuapa must have remained in west­ the eruption include: lahars, earthquakes, acid burns, ern El Salvador, moving "upoutofash-buried basins," carbon dioxide, carbon monoxide and Rourine poi­ orthat they retreated west to the nearby highlands and soning, and coastal tsunamis along the Pacific coast coastal plain of southeastern Guatemala. As of yet, (Blong 1984; Scarth 1999; Thorarinsson 1979).9 however, there is no archaeological verification of Although there is little doubt that central El Sal­ either suggested movement. Indeed it appears that vador was completely devastated by the TBJ erup­ most ofsoutheastern Guatemala was also abandoned tion and void of human life immediately following in the 5th century A.D. immediately following the it, the post-TBJ demography of western El Salvador TBJ eruption (Estrada Belli personal communication has been the subject of some debate; one side argu­ 2000; Shook 1965; Wauchope and Bond 1989). It ing for the virtually complete and prolonged depop­ might be that the western Salvadoran and southea,>t­ ulation of the upper Rio Paz drainage basin (Sharer ern Guatemalan sites receiving the immigrants have 1974, 1978b; Sheets 1984), and the other for demo­ not yet been found, or it could be that most of the graphic and cultural continuity-suggestingonIy tem­ Early Classic residents of Chalchuapa emigrated to porary local displacement of human populations areas further afield; we believe the latter interpreta­ following the TBJ event (Demarest 1986, 1988). tion is more likely. Demarest (1986: 179-180) speculates that the The crippling of Chalchuapa had a profound 34 LATIN AMERICAN ANTtOUITY [Vol. 12, No.1, 2001

opan

~ Kaminaljuyu

Figure 7. Zone of greatest impact from the TBJ eruption, including posited evacuation routes. effect on the interregional economy of the south­ ofthe southeast periphery interaction system focused eastern Maya area. Chalchuapa was one of the two on it" (Schortman and Urban 1991:137). largest regional centers in the southeastern highlands West of Chalchuapa, in the Jutiapa region of at the time of the TBJ eruption, controlling the pro­ Guatemala (Figure 6), a population decline is also duction and exportation of many important trade noted at this time (Wauchope and Bond 1989: 102, items during the Fonnative and Early Classic Peri­ Ill). The eight sites surveyed in this study record a ods (Demarest 1986; Schortman and Urban 1991; cultural hiatus from the end of the Preclassic until Sharer 1974, 1978a). This area produced prodigious ca. A.D. 600. Wauchope and Bond indicate that the amounts of Usulutan ceramics that were circulated TBJ eruption might be responsible for the hiatus, but throughout the highlands and Pacific lowlands of no finn conclusions are reached. Guatemala, and even exported to sites in the Maya The deleterious effects of the TBJ eruption also Lowlands such as EI Mirador and Cuello (Neff et al. extended into the Pacific coastal plain and piedmont 1999; Sharer 1974). Other items that may have been zone ofsouthern Guatemala and EI Salvador (Sheets traded through Chalchuapainclude Ixtepeque obsid­ 1987). Tephra deposits exceed 30 cm. at Laguna ian (Demarest and Sharer 1986), Motagua jade Gamboa, 105 km from Ilopango (Figure I-Site C). (Sheets 1979), and Pacific slope commodities such Located two kilometers east ofLaguna Gamboa, the as cacao (Bergmann 1969; Bove 1993a; Young archaeological site ofCara Sucia (Figures 6, 7) also 1994), cotton (Arroyo 1993; Sheets 1979), and salt contains as much as 50 cm of TBJ tephra (Arnaroli (Andrews 1991). Because of Chalchuapa's central 1987:IX-1). These tephra units are the eroded, weath­ position in this interregional trading network, the ered, and reworked remnants of the original airfall eruption ofIlopango not only caused the collapse and ash deposits, which presumably were much greater abandonment of the city but also "signaled the end at the time ofdeposition. Immediately following the Dull et al.] VOLCANISM, ECOLOGY, AND CULTURE 35

eruption Cara Sucia was abandoned for more than even a light «5 cm) dusting of tephra because peo­ 200 yearslO (Amaroli 1987:IV-1O). ple there relied more heavily on aquatic protein Cara Sucia was not the only Pacific slope site resources (Coe and Flannery 1967) that would have deserted following the TBJ eruption. Indeed, much been greatly damaged by the fine volcanic ash ofthe southeastern Pacific coast and piedmont zone (Sheets 1987). of Guatemala seems to have been abandoned from The abandonment ofCara Sucia can probably be ca. A.D. 400/450 to 600 (Bove 1993b; Estrada Belli explained by the direct effects of the deep airfall 1998, personal communication 2000; Shook 1965). tephra deposition. However, environmental difficul­ The area of southern Guatemala hardest hit ties in this area would likely have been overcome directly by the TBJ eruption was the Department of within two or three decades, not centuries. Pacific Santa Rosa in southeastern corner, just west of the coastal and piedmont sites to the west of Cara Sucia Rio Paz (Figure 6). Although not initially attributed would not have been hindered by long-lasting envi­ to the TBJ eruption ofIlopango, 11 tephra deposits of ronmental problems; to the contrary, tephra deposits 5-10 cm. within the Classic Period stratum have of < 20 cm may even have been beneficial to agri­ been measured in test units at several sites in this zone cultural production in southeastern Guatemala within (Estrada Belli, personal communication 2000). a decadefollowing the eruption (Blong 1984; Ugolini Based on ceramic cross-dating, Estrada Belli origi­ and Zasoski 1979). Why then, was this region largely nally assigned the cultural hiatus to ca. A.D. 200-450 abandoned for nearly two centuries immediately fol­ (Estrada Belli 1997),12 and raised the possibility that lowing the TBJ eruption? We agree with Schortman it could have been caused by the TBJ eruption and Urban (1991) that the demise of Chalchuapa­ (1998:190). However, given the new dating of the and numerous smallersites in central and western El TBJ, and considering the ceramic sequences found Salvador such as Cara Sucia--also signaled the in the region, Estrada Belli now believes that the cul­ downfall of a powerful interdependent southeastern tural hiatus may well have occurred 200-250 years periphery economic sphere (Providencia-Miraflo­ laterthan he had originally estimated (personal com­ res), one with a long history spanning much of the munication 2000). Formative period (Demarest and Sharer 1986). The Slightly to the west ofSanta Rosa, in the Escuintla TBJ event, therefore, should not be viewed exclu­ region (Figure 6), the Balberta project also found sively in terms ofits environmental impacts; the more widespread abandonment between A.D. 400/450and potent blow to the entire southern Maya realm prob­ 600, but apparently no tephra (Bove 1993b). Bove ably came in the form ofa collapsing regional econ­ (l993b:191) gives no material explanation for the omy and, as a result, the political destabilization and Early Classic population collapse, but remarks that, decentralization ofa number offledgling mercantile "Itis difficult to believe in a population abandonment polities. for several centuries of such a rich environmental Tephra fallout from the TBJ eruption was most zone." More recently Bove has argued that the aban­ severe to the west ofthe Ilopango caldera (Hart and donment of the Balberta nuclear zone should be Steen-McIntyre 1983; Figure 1). Accordingly, the attributed to an "explosive and militaristic aggres­ debilitating effects of the TBJ eruption, felt so sive intrusion of Teotihuacan colonists" (personal strongly in western El Salvador and parts of south­ communication 2000). eastern Guatemala, did not extend too far north into The depopulation ofsouthern Guatemala west of western Honduras. Situated40kmnorth ofIlopango, Escuintla (Figure 6) may have been significant fol­ the Paraiso Basin-a sub-basin of the upper Rio lowing the TBJ eruption, if not as dramatic as the Lempa Basin (Figures 6, 7}--did receive ~60 crn of southeastern coastal areas. Bilbao is inland and far­ tephra (Earnest 1999:41; Fowler and Earnest 1985), ther from Ilopango (~215 km., Figure 6), and it suf­ and although there is some evidence for a small rem­ fered a marked decline in the Early Classic, but not nant population following the eruption, the agricul­ an abandonment (Parsons 1967). Even more distant tural lowlands of the valley were "virtually sites in the Ocos area (~340 km) apparently were at uninhabitable for several centuries" following the least largely abandoned in the Early Classic, and TBJ eruption (Fowler and Eamest 1985:24). reoccupied in the Late Classic (Cae and Flannery There is no evidence oftephra deposition 130 km 1967). Ocos might have been more vulnerable to north of Ilopango at Copan. In fact, Copan thrived 36 LATIN AMERICAN ANTIQUITY [Vol. 12, No.1, 2001 immediately following the TBl eruption (Fash 1986; ulation (Michels 1979: 194-197). Sharer et ai. 1999), perhaps filling the void in the The trends in the urban data at Kamina1juyu interregional economy left by the collapse of (declines in population, household wealth) are con­ Chalchuapa. We are not suggesting that the impres­ sistent with the socioeconomic disruption presum­ sive rise ofCopan during the Classic Period was pre­ ably caused by the TBl eruption. The increase in rural cipitated solely by the collapse of Chalchuapa. The populations might be explained by some combina­ TBl eruption was just one of many factors that tion of out-migration from the economically together allowed the remarkable ascent of the bur­ depressed Kaminaljuyu and immigration of subsis­ geoning polity beginning in the fifth century A.D. tence fanners from the ash-mantled Pacific slope Apparently another reason for Copan's emergence region; indeed, post-TBl ceramic linkages affinn a at this time was its increasing association with a Pacific coast-Guatemalan highland association (Neff much larger cultural and economic sphere of influ­ et ai. 1994, 1999). There is evidence to suggest that ence, one controlled by Teotihuacan. other highland sites in Guatemala may also have The expansionist mercantile Teotihuacan civi­ been affected by the TBl eruption and collapse of lization evidently affiliated with one of the power the southern Maya economic sphere. Semetabaj, for blocks in the Early Classic period emerging in the example-located on the north rim of Lago Atitlan Peten: Tikal (Moholy-Nagy 1999; Schele and Frei­ (Figure 6)--was completely abandoned aroundA.D. del 1990; Sharer 1994). By the Middle Classic 400 (Shook et ai. 1979: 16). period, the Tikal-Teotihuacan association expanded SimilarTeotihuacanpolitical-economic objectives south to Kaminaljuyu and subsequently into Copan, may have, at least in part, underlain the founding of possibly to facilitate access to the two major Maya the Copan royal dynasty in A.D. 426 (Sharer et al. highland obsidian sources (EI Chayal and 1xtepeque), 1999:20), The Teotihuacan elements are clear in the and probably the jade source along the middle early talud-tablero architecture, the depictions ofthe Motagua River. earliest kingYax K'uk Mo in later art, and in artifacts. The Teotihuacan presence at Kaminaljuyu began The Maya colonization of the area included move­ slowly, perhaps with indirect exchange relationships ment into the Ixtepeque obsidian source and found­ (Henderson 1997: 135) dating to the early 300sA.D. ing ofPapalhuapa (Graham and Heizer 1968), a Maya or perhaps earlier (Sanders and Michaels 1977; "factory town" for processing that obsidian into Sharer 1994:676). That presence increased dramat­ macrocores for export back into the Maya lowlands. ically in the second halfofthe fifth century, and con­ The dramatic population increase in the Copan tinued strong until the second half of the sixth Valley during the Middle Classic (Fash 1986:79-82) century, when it began a decline and cessation. may not only be explained by the immigration of We suggest here the possibility that highland colonists from the Maya lowlands; the arrival of Maya society in the Kaminaljuyu area was weakened refugees from Chalchuapa and neighboring sites in from the Ilopango disaster and by the abrupt cessa­ western EI Salvador left uninhabitable by the TBl tion of economic ties with the larger Miraflores eruption may also have played a part. 13 Residents of sphere (especially Chalchuapa), thereby facilitating Chalchuapa fleeing the Rio Paz basin would have the Teotihuacan strengthening ofdemographic, polit­ found a similar physical environment (elevation, pre­ ica
' and economic influence there in the mid to late cipitation, temperature) in Copan to the one they left 400s. The weakening of Kaminaljuyu is confirmed behind in Chalchuapa (Portig 1965; Sharer 1994). by Michels (1979:30), who reports that "the stabil­ This notion is bolstered by evidence linking Copan ity that largely characterized the Early Classic is ter­ to western EI Salvador ceramic traditions after the minated dramatically with the onset of the Middle TBl eruption (Neff et ai. 1999). Classic phase." Several notable demographic Beyond the southeastern region of the Maya changes occurduring this time at Kaminaljuyu: a sig­ realm, there is no clear evidence for economic or nificant drop in Rank I (Paramount Elite) households political disruption precipitated by the physical from 39 percent in the Early Classic (A.D. 200--400) effects ofthe TBl eruption. However, the event may to 14 percent in the Middle Classic (A.D. 400-600); have had a great psychological effect on human pop­ a 33 percent reduction in total urban population; and ulations over an area much largerthan the devastated a 12.8 percent increase in the "sustaining" rural pop- zone. Comparisons to other large Plinian eruptions Dull et al.] VOLCANISM, ECOLOGY, AND CULTURE 37 suggest that earthquakes, heavy rain, lightening eruption starvation and disease have caused only 4 stonns, and daytime darkness probably accompa­ percent of the volcano-related deaths worldwide nied the TBJ eruption (Sheets 1987:45; Thorarins­ since 1900, this percentage swells dramatically to49 son 1979: 139; Williams 1952). percent for the pre-industrial period from 1600 to The ]835 Coseguina eruption, for example, 1899 (Tilling 1989:5). More people died from dis­ caused numerous earthquakes felt strongly in Leon, ease (-3000) following the 1902 eruption ofVolcan Nicaragua (88 km away) and San Miguel, El Sal­ Santa Maria in Guatemala (Figure 5) than from the vador (90 km), dry season torrential rains in Tegu­ direct effects ofthe eruption itself(Blong 1984:126). cigalpa, Honduras (130 km), and several days ofnear One account of the Santa Maria disaster is particu­ complete darkness stretching at least 550 km from larly revealing in this respect: "One of our greatest Guatemala City southeast to Rivas, Nicaragua troubles was that of sickness, owing to the balance (Williams 1952:32-33). Residents of Leon, ofNature having been upset by the eruption, which, Nicaragua attributed the earthquake, darkness, and having killed all the birds for some hundreds ofmiles, violent electrical stonns to divine anger, and flocked enabled the flies, mosquitoes, and rats to multiply to to the temples imploring the mercy ofheaven (Scarth suchanextent that life to man became nearly unbear­ 1999: 130). The Coseguinaeruption was heard as far able" (Anderson 1908, cited in Blong 1984:126). away as Bogota, Colombia (1,750 km) in the south Thus, malnutrition, starvation, and pestilence fol­ and Oaxaca, Mexico (1,100 km) in the north, and lowing the TBJ eruption might have been partly tephra deposition was reported as far away as south­ responsible for progressive demographic collapse ern Mexico (700 km) and Jamaica (1,280 km) (Sap­ throughout the abandonment zone. per 1925; Scarth 1999; Squier 1860; Williams 1952). Those who did evacuate the zone outlined in Fig­ Effects similar to those caused by Coseguina ure 6 probably fled north and northwestward toward would have also been generated by the much larger Copan and Kaminaljuyu respectively (Figure 7). TBJ eruption. To the prehistoric Maya, the accom­ Ceramic evidence implies an economic realignment panying wind, lightening, darkness, and earthquakes following demographic recovery in central and west­ would have held great religious significance, and ern El Salvadorand southeastern Guatemala. Where would likely have been attributed to supernatural formerly the Miraflores economic sphere united vir­ forces (Marcus 1978; Thompson 1970). As a result, tually the entire area (Demarest and Sharer 1986), we can speculate that the psychological trauma the post-TBJ ceramic associations show two distinct induced by the TBJ eruption might have contributed north-south trending economic spheres, one uniting to temporary social disruptions over an area of western El Salvador and Copan, and another linking Mesoamerica much wider than the zones of demo­ Kaminaljuyu to the Pacific slope region ofGuatemala graphic and economic collapse (Figures 1, 6, 7). (Neff et al. 1999). Although we now have a much more precise Depopulation and Demographic Recovery understanding of when the TBJ eruption happened, The areaofdemographic collapse following theTEJ the geography and chronology ofreoccupation in the eruption is still not clearly demarcated, but it depopulated zones is limited by a dearthofpublished stretched west at least to the Escuintla region and radiocarbon data, and by inaccurate ceramic cross­ north to Asuncion Mitla/Jutiapa, overlapping much dating. Furthermore, the post-TBJradiocarbon dates ofthe Providencia-Miraflores economic sphere (Fig­ that are available are generally not taken from the ure 6). Itmay be that the western and northern fringes earliest reoccupation archaeological contexts. Ifwe ofthis area were not completely deserted, but it seems adopt the working hypothesis that the eruption at a minimumthat centralized political systems were occurred at ca. A.D. 425,14 we can begin to stan­ abandoned and monumental construction was halted. dardize reports of an Early Classic population con­ Not everyone necessarily fled the area outlined in traction throughout central and western El Salvador Figure 6 immediately following the eruption. How­ and southeastern Guatemala. Table 2 shows what we ever, any survivors in the area of deep (> 50 cm) consider to be the "best,,15 radiocarbon dates for the tephra airfall would have encountered many tephra­ post-TBJreoccupation ofseveral sites in central and related hardships, most notably in the areas of food western El Salvador, including: Ceren, Chalchuapa procurement and disease outbreaks. Although post- and Cara Sucia. Table 3 contains sites from south- c.> (Xl

Table 2. Radiocarbon Dates for Latest Post-TBl Reoccupation in Central and Western EI Salvador.

Measured Age Calibrated Age--IO' Calibrated Age -20' Lab# Material Site Sampling Context (14C yr B.P.) (calendar yr A.D.) (calendar yr A.D.) Source Tx3113a T plumosus" charcoal Cercn Thatch roofing, structure I, south room 1330 ± 70 Sheets 1983b:7 Tx3119a wood charcoal Cercn Roof beam, strueture I, center room 1420 ± 50 Sheets 1983b:7 Tx6600 1'. plumosusa charcoal Cercn Thatch roofing, structure 3 1520 ± 70 authors Tx6601 7~ plumoslls" charcoal Cercn Thateh roofing, structure 2 1350 ± 90 authors ~ EIS40 1'. pillmosus" charcoal Cercn Thatch roofing, structure I 1440 ± 135 Sheets 1983b:7 z A-10743 7~ nlumosus" charcoal Ceren Thatch roofing 1360 ± McKee personal communication 2(XX) ~ :nm Cercn Composite (weighted mean) 1390 ± 30 642(656)662 605(656)684 ('; Y-682 charcoal Tazumall Chalchuapa Pot I, Grave 16, structure I 148()± 120 Stuiver and z» Deevey 1961: 131 » Y-685 charcoal Tazumall Chalchuapa Pot 8, Grave 13, structure I 1520 ± 120 Stuiver and Deevey 1961:131 ~ TazumalComposite (weighted mean) 1500±85 434(562,592,596)644 395(562,592,596)678 ~ A-1949b carbonized maize cob Cara Sllcia 'Thmasha Phase basurero 1380± 180 533(658)859 260(658) 10 18 Amaroli 1987: IV-4 Note: All dates are calibrated using the INTCAL 98 calibration (Stuiver et aI., 1998). Tx the University of Texas, Austin Radiocarbon Laboratory: EIS Universidad Nacional, Physics Department Radiocarbon Laboratory, EI Salvador; A-10743 University of Arizona Radiocarbon Laboratory; Y Yale University Radiocarbon Laboratory; A-1949b University of Miami Radiocarbon Laboratory. "Trachypogon plumosus b o1JC =-11.9. This is the only Ceren 14C date that has been o1JC corrected.

~ ~ z p :" ~'" Dull et al.] VOLCANISM, ECOLOGY, AND CULTURE 39

Table 3. Post-TBI Reoccupation in Southeastern Guatemala.

Region Evidence Approximate Date for Reoccupation Citation I Uliapa Stratigraphic hiatus, Ceramic seriation ca. A.D. 600 Wauchope and Bond 1989: 102 Santa Rosa Stratigraphic hiatus, Ceramic seliation ca. A.D. 625" Estrada Belli 1998:189,261 Escuintla Stratigraphic hiatus, Ceramic seriation ca. A.D. 600 Bove 1993a: 191 Norc: These dates should be considered tentative at present. as the cultural sequences of southeastern Guatemala have yet to be well-dated by radiocarbon. Future radiocarbon work focused on the reoccupation period should help refine our understanding of the geography and chronology of the post-TBI hiatus in this region. a The Santa Rosa region reoccupation date is based on an A.D. 425 TBI eruption (this study) and Estrada Belli's estimate of a 200-year hiatus during the Early Oscuro phase, immediately following the TBI eruption (Estrada Belli 1998: 261). eastern Guatemala discussed in the text and their vious radiocarbon dates reported for the event. The reported reoccupation times. We conclude from these composite calibrated radiocarbon date for the event data that the depopulated area outlined in Figure 6 is now A.D. 408(429)536 (two sigma range), but the represents a cultural hiatus ofroughly 100-150years. bulk ofthe probability distribution for the calibrated It is clear that much of the area abandoned after the radiocarbon date is concentrated near the older end eruption was resettled by the end of the sixth cen­ ofthis range, before A.D. 476. Ceramic, ice-core, and tury A.D. dendroclimatological data are all consistent with an early-to-middle fifth-century A.D. eruption, but so The TBI Tephra as a Time-Stratigraphic Marker far none of this evidence can definitely be linked to Understanding the impacts of the TBJ eruption on Ilopango. the cultural history of Mesoamerica has provided The trajectory of deposition for the majority of the impetus for much ofthe work done on this spec­ volcanic ejecta emitted by the TBJ event trends to tacular geologic event to date. Its utility as an inter­ the west of the Ilopango vent. Pyroclastic flows and pretive scientific tool, however, goes far beyond deep (2 m +) airfall tephra deposits had the greatest studies of Maya demography and socioeconomic impact on people living near to or directly west of change. Ilopango in the San Salvador basin and on the east­ The TBJ tephra as a tephrochronological marker ern fringes of the Zapotitan basin. Most inhabitants is important to archaeologists and paleoecologists of this region probably perished as a result of their alike. In archaeological contexts, the very presence proximity to deadly volcanic hazards: pyroclastic of the TBJ tephra layer provides a low cost, highly flows, lahars, emissions of noxious gasses, or deep reliable minimum date for cultural remains that lie tephra deposits. below it (Amaroli and Dull 1999; Eamest 1999; Outside ofthe zone ofdevastation, there is a much Fowler and Earnest 1985). Similarly, the presence of larger area of prolonged depopulation (-100-150 the TBJ tephra in marine, estuarine, and lacustrine years) following the TBJ eruption. The demographic sediments allows for regional correlation of strati­ collapse in this area-encompassing most of west­ graphic records of environmental change (Dull ern El Salvador and southeastern Guatemala-was 1998). Moreover, a well-dated, widely distributed probably progressive and can be attributed to a com­ tephra such as the TBJ can help circumvent prob­ bination of three factors: immediate evacuation due lems associated with the lack ofsuitable samples for to volcanic hazards, post-eruption starvation and dis­ radiocarbon dating-for example, the "hardwater ease, and regional economic collapse. effect" that often confounds efforts to use radiocar­ Residents ofChalchuapa, located 77 km west of bon to date stratigraphic lacustrine sequences in the vent, were probably far enough away that most karstic areas like the Maya lowlands (Leyden 1987). ofthem survived, but they would not have been able to remain in the ash-laden Rio Paz basin following Conclusion the eruption. The initial airfall deposits would have New dating evidence indicates that the TBJ eruption gradually been augmented by tephra eroded from the ofVolcan Ilopango, El Salvador occurred not at the upland areas during the first few rainy seasons fol­ close ofthe Fonnative Period, but at least 150 years lowing the eruption. Comparative studies show that later, during the Early Classic Period. Three new it would have taken decades before subsistence AMS detenninations support a reanalysis ofthe pre- agriculture could have been resumed in the valley 40 LATIN AMERICAN ANTIQUITY [Vol. 12, No.1, 2001 floor. The radiocarbon data from Tazumal support graphically situated along a mid-elevation river val­ this finding, showing that Chalchuapa was not sig­ ley, a very similar environment to the one they left nificantly reoccupied until the sixth century AD. behind in the Rio Paz basin. During this period of An Early Classic period depopulation is also rapid demographic, cultural, and economic change posited for the Pacific slope of western El Salvador at Copan-partlyspurred by the TBJ event-thedom­ and southeastern Guatemala. However, the popula­ inant preexisting political infrastructure was appar­ tion collapse in this region cannot be explained ently weakened, facilitating the ascendance of the merely in terms of physical and biological lowland Maya-Teotihuacan alliance. effects/constraints. We believe that the TBJ eruption Finally, it is our hope that this paper will high­ had a profound effect on the economic viability of light the general need in Mesoamerican archaeology this region via its dependence on trading partners for a more critical reading ofradiocarbon results and who were more heavily impacted directly by the more intense scrutiny of the radiocarbon dates that eruption (e.g., Chalchuapa). So although much of support ceramic seriations. Our understanding of southwestern El Salvador and southeastern incipient food production in Mesoamerica has Guatemala was covered with a measurable layer of changed radically in recent years due to the reanaly­ ash (~5-50+ cm) following the eruption, the envi­ sis (using AMS techniques) of long-established ronmental damage only partially explains the Early archaeological chronologies (Fritz 1994; Long et al. Classic hiatus in the region. The disruption of the 1989), and this is certainly not the end of the road. "Miraflores" southern highland-Pacific coast cul­ The re-dating of the TBJ eruption is just one step in tural/economic sphere was likely more damaging to a long process ofcritically reevaluating the chronol­ the inhabitants ofthe Pacific slope overthe long term ogy of cultural evolution and Holocene environ­ than the environmental effects of the TBJ. mental change throughout Mesoamerica. The southern highlands ofGuatemala may also have been strongly affected by the TBJ eruption. Acknowledgemems. This project would not have been possible At Kaminaljuyu, the 200-year period following the without the help and advice of a number of people in EI TBJ event is characterized by marked declines in Salvador and the United States. We would like to thank Brian McKee, Robert Sharer, and Francisco Estrada Belli for pro­ the wealthiest households and overall urban popu­ viding archaeological and/or radiocarbon data from the study lations, and a modest increase in rural populations. area; Greg Zielinski for providing data on volcanic acidity We attribute the declines in urban dwellers and deposition in the GISP2 ice cores; and Keith Brina and Ian wealthy households to the economic disruption of Harris for supplying northern hemisphere tree-ring tempera­ the southern Maya area caused by the TBJ erup­ ture records. The final manuscript was improved considerably by helpful comments from Edward Schortman, William tion. Furthermore, we believe that the instability Fowler, Frederick Bove, Patricia Fournier, Katharina created by the physical and economic effects ofthe Schreiber, and three anonymous reviewers. We thank Claudia eruption facilitated the penetration ofTeotihuacan Leal for significantly reworking the Spanish abstract. cultural, political, and economic influence. The Fieldwork in EI Salvador in 1997 and 1998 was enabled by the increase in rural populations surrounding Kami­ help of Jose "Gamaliel" Garza and Paul Amaroli. Brian Frantz helped Southon and Dull with the preparation of samples in naljuyu might have resulted from out-migration the CAMS graphite lab. Tom Bowling and Maria Escobar de from Kaminaljuyu orfrom rural farmers fleeing the Bowling graciously provided room and board for Dull in ash-covered Pacific slope of southern Guatemalan Chalchuapa, EI Salvador in 1997 and 1998. Many of the ideas and El Salvador. put forth in this paper by the lead author (Dull) have been Another possible recipient ofthe TBJ refugees is developed and crystallized through discussions of the TBJ eruption with Roger Byrne, Bill Denevan, and Karen Bruhns. Copan, which experienced a pronounced increase in Fieldwork by Dull in EI Salvador was carried out with the per­ population beginning in the fifth centuryAD. Copan mission of PANAVIS, and supported by the following grants was apparently spared the severe environmental for his Ph.D. research: Stahl Grant for Archaeological impacts of the TBJ that were so crushing in areas Research, Tinker Foundation Travel Grant. UC Berkeley Vice directly west and northwest ofthe Ilopango caldera, Chancellor for Research Grant, UC Berkeley Social Science Research Grant, and the UC Berkeley Department of and this is precisely why Copan was probably so Geography Carl O. Sauer Research Fellowship, Portions of attractive to the fleeing urbanites of Chalchuapa. In this work were performed under the auspices of the United Copan, the former inhabitants ofChalchuapa would States Department of Energy contract W-7405-Eng-48. have found a growing, vibrant urban center geo- Dull et al.] VOLCANISM, ECOLOGY, AND CULTURE 41

References Cited Coe, Michael D., and Kent V. Flannery 1967 Earlv Cultures and Human Ecology in South Coastal Amaroli, Paul Guatemdla, Smithsonian Contributions to Anthropology 3, 1987 Informe preliminar de las excavaciones arqueol6gicas Smithsonian Institution Press, Washington, DC. en Cara Sucia, Departamento de Ahuachapan, EI Salvador, D'Anigo, Rosanne D., and Gordon C. Jacoby Ministerio de Cuhura y Communicaciones, Direccion de 1999 Northern North American Tree-Ring Evidence for Patrimonio Cultural. San Salvador, EI Salvador. Regional Temperature Changes after MajorVolcanic Events. Amaroli, Paul, and Robert Dull Climatic Change 41 :1-15. 1999 Milpas prehispanicas en EI Salvador. In XII Simposio Dam,hertv, Howard E. de Investigaciones /irqueol6gicas en Guatemala, 1998, 1969 ~Man-Induced Ecological Change in EI Salvado!'. Ph.D, edited by Juan Pedro Laporte, Hector L Escobedo and Ana dissertation, University ofCalifornia, Los Angeles. Univer­ Claudia Monzon de Suasnaver, pp, 639-650, Museo sitv Microfilms, Ann Arbor. Nacional de Arqueologia y Etnologia. Demar;sr, Arthur A. Anderson, T. 1986 The Arc/weoloay o(Sama Leticia and the Rise ofMaya 1908 The Volcanoes ofGuatemala. Royal Geographical Soci­ Civilizarion, TUlan~' University Middle America Research ety 473-489 Institute, New Orleans. Andrews, Anthony P. 1988 Political Evolution in the Maya Borderlands: The Sal­ 1991 Las salinas de El Salvador: bosquejo historico, etnogra­ vadoran Frontier. In The Southeast Classic Maya Zone, fico y arqueologico. Mesoamerica 12(21 ):71-93. edited bv Elizabeth H. Boone and Gordon R. Willey, pp, Andrews, E. Wvllvs 335-394. Dumbarton Oaks Research Library, Washington, 1986 La arqu;ologia de Quelepa. Direccion de Publicaciones DC. y Impresos, San Salvador. Demarest, Allhur A., and Robert J. Sharer Angell, J. K., and J. Korshover 1986 Late Predassic Ceramic Spheres, Culture Areas, and 1985 SurfaceTemperature Changes Following Six MajorVol­ Cultural Evolution in the Southeastern Highlands of canic Episodes between 1780 and 1980. Journal ofClimate Mesoamerica. In The Southeast Maya Periphery, edited by andApplied Meteorology 24:937-951. PatriciaA, Urban and Edward M. Schortman, pp. 194-223. Antos, Joseph A., and Donald B. Zobel University ofTexas Press, Austin, 1987 How Plants Survive Burial: A Review and Initial Denevan, William M. Responses to Tephra from Mount St. Helens. In IHount St. 1992 Introduction. In The Native Population ofthe Americas Helens 1980: Botanical Consequences of the Explosive in 1492. 2nd ed, Edited by William M, Denevan, pp, xv-xli, Eruptions, edited byDavid E. Bilderback, pp, 246-259. Uni­ University of Wisconsin Press, Madison. versity of California Press, Berkeley. Dull, Robert Arroyo, Barbara 1998 Late Holoeene Human Ecology of the Chalchuapa 1993 Spindle Whorls from Balberta. In The Balherta Project, Archaeological Zone: A 3700Year Stratigraphic Record of edited by Frederick J. Bove, Sonia Medrano, Brenda Lou, Agriculture, Erosion, Fire and Vegetation Change from and Barbara Arroyo, pp, 137-144, University of Pittsburgh Laguna Cuzcachapa, Western EI Salvador. Paper presented Department of Anthropology and Asociacion Tikal, Pitts­ at the 15th Biennial Meeting of the American Quaternary burgh. Association, Puerto Vallarta, Mexico, Bergmann, John E Earnest. Howard 1969 The Distribution ofCacaoCultivation in Pre-Columbian 1999 A Reappraisal 01' the Jlopango Volcanic Eruption in America. Annals of the Association ofAmerican Geogra­ Central EI SalvadOl: Doctoral Disseration, Harvard Univer­ phers 59:85-96, sity, University Microfilms, Ann Arbor, Blong, R. 1. Eggler, Willis A. ~ 1984 Volcanic Hazards: A Sourcehook on the Effects ofErup­ 1948 Plant Communities in the Vicinity ofthe Volcano EI Pan­ tions. Academic Press Australia, North Ryde, New South cutin, Mexico, afterTwo and a HalfYears ofEruption, Ecol- Wales, ogy 29(4):415-436. ._ Bove, Frederick J. I %3 Plant Life of Par/cutin Volcano, Mexico, Eight Years 1993a Balberta in Perspective. In The Balherta Project, edited after Activity Ceased. American Midland Naturalist by Frederick J, Bove, Sonia Medrano, BrendaLou, and Bar­ 69(1 ):38-68. bara Arroyo, pp, 1-14, University of Pittsburgh Department Estrada Belli, Francisco of Anthropology and Asociacion Tikal, Pittsburgh, 1997 Survey in Jutiapa, Southeastern Guatemala. Mexicon 1993b The Terminal Formative-Early Classic Transition. In 20(3):55-59, The Balberta Pr~ieet, edited by Frederick 1. Bove, Sonia 1998 The Evolution of Complex Societies in Southeastern Medrano, Brenda Lou, and Barbara Arroyo, pp, 177-194. Pacific Coastal Guatemala: A Regional GIS Archaeological University of Pittsburgh Department of Anthropology and Approach. Ph,D, dissertation, Boston University. University Asociacion Tikal, Pittsburgh. Microfilms. Ann Arbor. Brady, James E., Joseph W. Ball, Ronald L Bishop, Duncan C. Fash, Jr., William L. Prin" Norman Hammond, and Rupert A, Housley 1986 Historv and Characteristics ofSettlement in the Copan 1998 The Lowland Maya"Protoclassic,"AncientMesoamer­ Valley, and'SomeComparisons with Quirigua. In The South­ ica 9: 17-38. east Mava Periphery, edited by Patricia A. Urban and Edward Briffa, Keith R. M, Schortman, pp. 72-93. University ofTexas Press,Austin, 2000 Annual Climate Variability in the Holocene: Interpret­ Feldman. Lawrence H. ing the Message of Ancient Trees, Quaternary Science 1993 Mountains ofFire, LandsThat Shake: Earthquakes and Reviews 19:87-105, Volcanic Eruptions in the Historic Past ofCentral America Bryson, Reid A., and Brian M. Goodman 11505-1899), Labvrinthos, Culver City, California, 1980 Volcanic Activity and Climatic Changes, Science Fisher, Richard V" Gra;t Heiken, and Jeffrey B. Hulen 207:1041-1044. 42 LATIN AMERICAN ANTIQUITY [Vol. 12, No.1, 2001

1997 Volcanoes: Crucibles ofChange. Princeton University 1999 Mexican Obsidian at Tikal, Guatemala. Latin American Press. Princeton, New Jersey. Antiquity 10(3):300-3II Fowler, William R., and Howard H. Earnest Neff, Hector, Frederick 1. Bove, Eugenia J. Robinson, and Bar­ 1985 Settlement Patterns and Prehistory ofthe Paraiso Basin baraArroyo ofEl Salvador. Journal ofFiled Archaeology 12:19-32. 1994 A Ceramic Compositional Perspective on the Forma­ Fritz, Gayle J. tive to Classic Transition in Southern Mesoamerica. Latin 1994 Are the First American Farmers Getting Younger? Cur­ /lmerican Antiquity 5:333-358. relll Alllhropology 35(3):305-309. Neff, Hector, James W. Cogswell, Laura J. Kosakowsky, Fran­ Graham, John, and Robert Heizer cisco Estrada Belli. and Frederick J. Bove 1968 Notes on the Papalhuapa Site. Colllribwions oflhe Uni­ 1999 A New Perspective on the Relationships among Cream versily of CalitiJrllia Archaeological Research Facilily Paste Ceramic Traditions of Southeastern Mesoamerica. 5:101-125. Latin American Antiquity 10:281-299. Harrison, Kevin, Wallace S. Broecker, and Georges Bonani Parsons, Lee A. 1993 A Strategy for Estimating the Impact of CO) Fertiliza­ 1967 Bilbao, Guatemala: An ArcheologicalStudy orthe Pacific tion on Soil Carbon Storage. Glohal Biogeochemical Cycles Coast Cotzumalguapa Region, Publications in Anthropol­ 7:69-78. ogy, No. I I. Milwaukee Public Museum,Milwaukeee. Hart, William, and Virginia Steen-McIntyre Portig, W. H. 1983 Tien'a Blanca Joven Tephra from the AD 260 Eruption 1965 Central American Rainfall. Geographical Review ofllopango Caldera. In Archaeology and Volcanism in Cen­ 55:68-90. lral America., edited by Payson D. Sheets, pp. 14-34. Texas Rees, John D. Pan American Series. University of Texas Press, Austin, 1979 Effects of the Eruption of Paricutin Volcano on Land­ Texas. fOlms, Vegetation and Human Occupancy. In Volcanic Activ­ Henderson. John S. ity and Human Ecology, edited by Payson D. Sheets and 1997 The World oflhe Ancient Maya. 2nd ed. Cornell Uni­ Donald K. Grayson, pp. 249-292. Academic Press, New versity Press, Ithaca, New York. York. Jenny, Hans Rose, William I., F. Michael Conway, Carlos Pullinger, Alan 1941 Factors uf Soil Formalion. Dover Publications, New Deino. and William C. McIntosh York. 1999 An ImprovedAge Framework for Late Quaternary Sili, Jones, P. D., Keith R. Briffa, and F. H. Schweingruber cic Eruptions in Northern Central America. Bulletin ofVol­ 1995 Tree-Ring Evidence ofthe Widespread Effects ofExplo­ canology 61:106-120 sive Volcanic Eruptions. Geophysical Research Letlers Sanders, William, and Joseph Michaels 22(11): 1333-1336. 1977 Teotihuacan and Kaminaljuyu: A Study in Prehistoric Larde, Jorge Culture Contact. Pennsylvania State University Press, Uni­ 1926 Cronologfa arqueol6gica de El Salvador. Revisla de versity Park. Ernologia. Arqueulogia y Linguistica 1(3-4):153-162. Sapper, Carlos K. (Karl) Larde y Larin, Jorge 1925 Los volcanes de la America Cemml. Estudios sobre 1957 EI Salvador.' historia de sus pueblos. villas y ciudades. America Central. Verlag von Max Niemeyer, Halle. Departamento Editorial del Ministerio de Cultura, San Sal­ Sapper, Karl vador. 1927 Volkankunde.l Engelhorns Nachf., Stuttgart. Lawn, Barbara Scarth, Alwyn 1974 University of Pennsylvania Radiocarbon Dates XVII. 1999 Vulcan's Fill)': Man against the Volcano. Yale Univer­ Radiocarbon 16(2):219-237. sity Press. New Haven. Leyden, Barbara W. Schortrnan. Edward M.. and Patricia A Urban 1987 Man and Climate in the Maya Lowlands. Quaternary 1991 Patterns of Late Preclassic Interaction and the Forma­ Research 28:407-414. tion of Complex Society in the Southeast Maya Periphery. Long, Austin, Bruce Benz, Douglas J. Donohue, AJ. Timothy Jull, In The Formatiun of Complex Society in Southeastern and Larry Toolin Mesuamerica, edited by William R. Fowler, k,pp. 121-142. 1989 First DirectAMS Dates on Early Maize from Tehuacan, CRC Press. Inc., Boca Raton, Florida. Mexico. Radiocarbun 31: I035-1040. Schele, Linda, and David Freidel Lothrop, Samuel K. 1990 The Untold Story ofthe Ancient Maya. Morrow, New 1927 PotteryTypes and TheirSequence in EI Salvador. Indian York. Notes and Monographs 1(4). Scuderi, Louis A. Lovell, W. George, and Christopher H. Lutz 1990 Tree-Ring Evidence forClimatically Effective Volcanic 1995 Demography and Empire. Dellplain Latin American Eruptions. QuaternOlY Research 34:67-85. Studies, Number 33. Westview Press, Boulder. Sharer, Robert J. Marcus, Joyce 1974 Prehistory ofthe Southeastern MayaPeriphery. Current 1978 Archaeology and Religion: A Comparison of the Amhropulogy 15(2):165-176. Zapotec and Maya. World Archaeology 10(2): 172-191. 1978a The PrehistUl)' ofChachuapa, EI Salvador, vol. 1. The Michael, Henry N., and Jeffrey Klein University of Pennsylvania, Philadelphia. 1979 An International Calibration for Radiocarbon Dates. 1978b The Prehistol)' ofChachuapa, EI Salvador, vol. 3. The MASCA Journal 1(2):56-57. University ofPennsylvania, Philadelphia Michels, Joseph W. 1994 The AnciemMaya, Stanford University Press, Palo Alto. 1979 The Kaminaljuyu Chiefdom. The Pennsylvania State Sharer, Robert J., and James C. Gifford University Press Monograph Series on Kaminaljuyu. Penn­ 1970 Preclassic Ceramics from Chalchuapa, EI Salvador, and sylvania State University. Their Relationships With the Maya Lowlands. American Moholy-Nagy, Hattula Antiquity 35(4):441-462. Dull et al.] VOLCANISM, ECOLOGY, AND CULTURE 43

Sharer, Robert J., Loa P. Traxler, David W. Sedat, Ellen E. Bell, 1979 On the Damage Caused by Volcanic Eruptions with Marcello A. Canuto, and Christopher Powell Special Reference to Tephra and Gasses. In Volcanic Activ­ 1999 Early Classic Architecture beneath the Copan Acropo­ ity and Human Ecology, edited by Payson D. Sheets and lis. Ancient /'lesoamerica 10(1999):3-23. Donald K. Grayson, pp. 125-160. Academic Press, New Sheets, Payson D. York. 1971 An Ancient Natural Disaster. E\pedition 13(1 ):24-31. Thornton, Ian W.B. 1979 Environmental and Cultural Effects of the llopango 1996 Krakatau: The Destruction andReassembly qfan Island Eruption in Central America. In Volcanic Actil'ityandHuman Ecosystem. Harvard University Press, Cambridge, Massa­ Ecology, edited by Payson D. Sheets and Donald K. Grayson, chusetts. pp. 525-564. Academic Press, New York. 2000 The Ecology ofVolcanoes: RecovelY and Reassembly 1981 Volcanoes and the NaturaIHistorv90(8):32-41. ofLiving Communities. In Encyclopedia oil'il/canoes, edited 1983a Archaeology in Central America. Uni- by Haraldur Sigurdsson, Bruce Houghton, Stephen R. versity ofTexas Press, Austin. McNutt, Hazel Rymer, and John Stix, pp. 1057-1081. Aca­ 1983b Introduction. In Archaeology and Volcanism in Cen­ demic Press, San Diego. tral America., edited by Payson D. Sheets, pp. 1-13. Texas Tilling, Robert L (editor) Pan American Series. University of Texas Press. Austin. 1989 Volcanic Hazards. American Geophysical Union, Wa,h­ Texas. ington, D.C. 1984 The Prebistory of EI Salvador: An Interpretive Sum­ Ugolini, F. c., and J. Zasoski mary.ln The Archaeology ofLowerCentral America. edited 1979 Soils Derived from Tephra. In Volcanic Activitv and by Frederick W. Lange and Doris Z. Stone, pp. 85-1 12. New Human Ecology, edited by Payson D. Sheets and Donald K. Mexico Press. Albuquerque. Grayson, pp. 83-124. Academic Press, New York. 1987 Possible Repercussions in Western Honduras of the Vogel, John S., Winton C. Cornell, D. Earle Nelson, and John R. Third-Century Eruption ofllopangoVolcano. In The Periph­ Southon ei), ofthe Southeastern Maya Realm, edited by Gary Pahl, 1990 Vesuvius!Avellino, One Possible Source ofSeventeenth pp. 43-52. Vol. 61, UCLA Latin American Center Publica­ Century BC Climatic Disturbances. Nature 344:534-537. tions, Los Angeles. Wauchope, Robert, and Margaret Bond Shook, Edwin M. 1989 ArchaeologicalInvestigations in the Department of'./uti­ 1965 Archaeological Survey of the Pacific Coast of apa, GuateJlwla. Middle American Research Institute, Pub­ Guatemala. In Handbook ofMiddle American Indiam. vol. lication 55. Tulane University, New Orleans. 2, edited by Gordon R. Willey, pp. 180-194. University of Wilkie, James w., and Jose Guadalupe Ortega (editors) Texas Press, Austin, Texas. 1997 Statistical Abstracts ofLatin America. 33. The Regents Shook, Edwin M" Marion P. Hatch, and Jamie K. Donaldson of the University of California, Los Angeles. 1979 Ruins ofSemetabaj, Dept. Solola, Guatemala. In Con­ Williams, Howel tributions of the University of California Archaeological 1952 The Great Eruption of Coseguina, Nicaragua, in 1835. Research Facility, edited by John A. Graham, pp. 7-142. University ofCalifornia Publications in GeologicalSciences Studies in Ancient Mesoamerica, IV. vol. 41. Department of 29(2):21-46. Anthropology, University of California, Berkeley. Williams, Howell, and Alexander R. McBirney Simkin, Tom. and Lee Siebert 1979 Volcanology. Freeman. Cooper & Co.• San Francisco. 1994 Volcanoes ofthe World. 2nd ed. Geoscience Press, Tuc­ Wright, H. E., D. H. Mann, and P. H. Glasser son. 1984 Piston Corers for Peat and Lake Sediments. Ecology Smith, Keith 65:657-659. 1992 Enl'ironmental Hazanls: Assessing Risk and Reducing Young, Allen M. Disaster. Routledge Physical Environment Series. Rout­ 1994 The Chocolate Tree: A Naturaillistory ofCacao. Smith­ ledge, LDndon. sonian Institution Press, Washington, DC. Squier, E. G. Zielinski, Gregory A. 1860 Nicaragua; Its People. Scenel)'. Monuments. Resources. 1995 Stratospheric Loading and Optical Depth Estimates of Condition. and Proposed Canal; with One Hundred Origi­ Explosive Volcanism over the Last 21 OOYears Derived from nal Maps andl/Iustrations. Harper & Brothers, New York. the Greenland Ice Sheet Project 2 Ice Core. Journal o/Geo­ Stevens, Rayfred L. physical Research IOO(D I0):20,937-20,955. 1964 The Soils of Middle America and Their Relation to 2000 Use ofPaleo-Records in DeterminingVariability within Indian Peoples and Cultures. In Handbook ofMiddle Amer­ the Volcanism-Climate System. Quaternary Science Reviews ican Indians. Vol. 1: Natural Environment and Early Cul­ 19:417-438. tures, edited by Robert C. West, pp. 265-315. University of Zielinski, Gregory A., Paul A. Mayewski, L.D. Meeker, S. Whit­ Texas Press, Austin. low, M.S. Twickler, M. Morrison, Deborah A. Meese, A. J. Stuiver, Minze, and Edward S. Deevey Gow, and Richard B. Allev 1961 Yale Natural Radiocarbon Measurements VI. Radio­ 1994 Record ofVolcanism'since 7000 B.c. From the Gisp2 carbon 3:126-140. Greenland Ice Core and Implications for the Volcano-Cli­ Stuiver, Minze, Paula J.Reimer, Eduard Bard, J.Warren Beck, mate System. Science 264:948-957. George S. Burr, Konrad A. Hughen, Bernd Kromer. Gerry McCormac, Johannes van del' Plicht, and Marco Spurk Notes 1998 INjCAL98 Radiocarbon Age Calibration 24,000-0 Cal BP. Radiocarbon 40: I041-1083. I. The two innermost isopachs (100+ cm and 50-100 cm) Thompson, J. Eric S. are based on the field measurements by Hart and Steen­ 1970 Maya HistOl)' and Religion. University of Oklahoma Mcintyre (1983:22, Figure 2-10). The outermost partial isopach Press, Norman. (20-50 cm) is based on field measurements by Robert Dull in Thorarinsson, Sigurdur lake sediments from Laguna Llano (20 cm+) and Laguna 44 LATIN AMERICAN ANTIQUITY [Vol. 12, No.1, 2001

Gamboa (30 cm+), both in the department of Ahuachapan, and deposits of onJy 15 em posed an insurmountable hurdle to agri­ measurements by Paul Amaroli in archaeological test units at cultural success for more than a decade after the eruption Cara Sucia (Amaroli 1987). Amaroli found TBJ deposits as (Scarth 1999). thick as 50 cm. in his Cara Sucia excavations (1987: IX-I) 9. Historic examples of these hazards include: lahars such as 2. In using the term "Protoclassic," we follow the interpreta­ those that in 1985 killed more than 23,000 people when the tion of Brady et al. (1998), who suggest that the Protoclassic Nevado del Ruiz volcano erupted in Colombia; carbon dioxide ceramic stage spans the period from ca. A.D. 75 to 420, over­ poisoning such as that which in 1986 killed more than 1,700 res­ lapping both the Late Formative and Early Classic periods. idents of Cameroon; and coastal tsunamis such as those that in 3. We define the Early Classic as the period from A.D. 250 1883 killed more than 32,000 people following the eruption of to 600 (Henderson 1997). Krakatau in Indonesia (Scarth 1999; Tilling 1989). 4. This date (1940 ± 50:4C yr B.P) is also apparently too old 10. The reported hiatus period of 300 years should be con­ in compmison to 2 pre-TBJ dates from the Chalchuapa area. sidered tentative considering the apparent lack of internal con­ One of these dates. 1840 ± 40 14C yr B.P. (P-1547). is fi'om char­ sistency in the reported radiocarbon data series (Amaroli coal within a Caynac complex (Sharer 1978a) vessel from EI 1987:chapter IV). Using the best of Amaroli's reoccupation Trapiche (Lawn 1974). The other date. 1830 40 14C yr B.P dates (A-1949b, Tamasha Phase carbonized maize cob: (CAMS 6(524), comes from a dicot leaf macrofossil sample 533(658)859 cal AD) , we interpret the Cara Sucia hiatus to be over a meter below the TBJ tephra in a Laguna Cuzcachapa lake roughly 200 years (Table 2). sediment core. II. Estrada Belli has shown samples of this material to 5. We only discovered that Eamest was working on the TBJ Sheets, who suggested that a TBJ identification could not be chronology problem in March 2000 as we were completing the made without trace element verification. first draft of this manuscript. The ceramic cross-dating issue is 12. The Pacific coast disruption was also assigned this one that we had not confronted. and Eamest does an excellent FOlmative/Classic date by Neff et al. (1994:333), who conclude job of piecing together that complicated puzzle. Note that he that, "some kind ofdramatic historical discontinuity affected the also reevaluates the radiocarbon evidence for the timing of the lives of the Pacific-slope people around AD. 200." TBJ eruption, and though he offers no new data, his conclusions 13. Given the proximity to the Motagua Valley, it is quite based on a detailed reassessment of the original suite of TBJ likely that Salvadoran TBJ refugees also made found their way dates are very similar to our own. into that region and perhaps to the Maya Lowlands beyond 6. These records include: Tometrask. Sweden (68QN, 20QE). (Sharer and Gifford 1970; Sheets 1979). Yamal, Russia (67.5QN, 70QE) and Taimyr, Russia (noN. 14. A best guess considering the probability distribution of 102QE). the composite radiocarbon data and supplementary ice core, tree 7. Although the post-Santa Maria eruption landscape ring, and ceramic data discussed earlier. depicted here is probably an accurate analog for the TBJ erup­ 15. We decided to consider only those dates from identifi­ tion in terms of tephra volume-NEI 6 (Simkin and Siebert able short-lived samples (plant macrofossils) and charcoal from 1994). the landscape does not convey the violence associated controlled proveniences. It should be noted that many samples with the TBJ pyroclastic flows. The fact that dead trees remain unsuitable for radiocarbon data are reported in the literature, and standing in this Santa Maria photograph indicates that lateral caution must be used in interpreting these results. The dates that base surge pyroclastic flows, such as those that were so devas­ are reported should be considered a, latest possible dates for tating in El Salvador during the TBJ eruption (Hart and Steen­ reoccupation. The Ceren dates actually date houses buried by McIntyre 1983), were not a component of the 1902 Santa Maria the Loma Caldera eruption; clearly some years must have eruption. The photograph is reprinted from Sapper (1927:88, passed after the time people reoccupied the zapotitan basin, Tafel IV). before the Loma Caldera eruption. Similarly the dates from 8. It should be stressed that the middle twentieth-century Tazumal record a period of resumed monumental architecture farmers of Paricutin possessed more advanced farming imple­ construction in the Chalchuapa area at some time following the ments and technological knowledge than the ca. fifth-century initial fe-occupation. farmers ofthe Rio Paz basin. Foremost among the advantages of the Paricutin farmers were metal tools (plows, hoes, shovels, Received May 22. 2000; accepted September 21. 2000; revised etc.) and domesticated animals (e.g., oxen). Nonetheless, tephra November 17, 2000.