LETTERS https://doi.org/10.1038/s41562-019-0671-x 1 US Geological Survey, Menlo Park, CA, USA. 2 Department of Geography, University of California, Berkeley, CA, USA. 3 Department of Anthropology, Tulane University, New Orleans, LA, USA. 4 Department of Anthropology, University of Alabama, Tuscaloosa, AL, USA. *e-mail: dwahl@usgs.gov Despite over a century of archaeological research, the nature and broader consequences of Maya warfare remain poorly understood. Classic period (250–950 CE) Maya warfare has largely been viewed as ritualized and limited in scope 16 . Evidence of violent warfare in the Terminal Classic period (800–950 CE) is interpreted as an escalation of military tac- tics that played a role in the socio-economic collapse of the Classic Maya civilization 7 ,8 . The implications of specific textual references to war events (war statements) remain unknown, however, and the paucity of field data precludes our ability to test collapse theories tied to warfare. Here we connect a massive fire event to an attack described with a Classic period war statement. Multiple lines of evidence show that a large fire occurred across the ancient city of Witzna, coincident with an epigraphic account describing an attack and burning of Witzna in 697 CE. Following this event, evidence shows a dramatic decline in human activity, indicating extensive nega- tive impacts on the local population. These findings provide insight into strategies and broader societal impacts of Classic period warfare, clarify the war statement’s meaning and show that the Maya engaged in tactics akin to total warfare earlier and more frequently than previously thought. The paradigm of prehistoric Maya military conflict has long held that total warfare, defined here as targeting both civilian and military resources resulting in widespread destruction of a city, was an aber- ration unique to the Terminal Classic period (TCP; 800–950 ce). Under this model, warfare before the TCP maintained rigid rules of engagement centred on procuring elite captives for tribute and ransom, with minimal involvement of non-combatants and limited socio-economic impact on the population 5,6,911 . The escalation in frequency and intensity of warfare during the TCP has been pre- sented as an important contributing factor in the collapse of Maya state institutions 7,8,12 . This view is supported by abundant evidence of violent conflict in the TCP from the Petexbatun region (Fig. 1). Complete destruction of sites, which occurred in the late eighth and early ninth centuries at Dos Pilas, Aguateca and Cancuen, is interpreted as a departure from warfare norms, symptomatic of TCP socio-political fractionation 1214 . These hypotheses regard- ing specifics of Classic period military tactics, however, remain widely untested. Hieroglyphic inscriptions contain several kinds of warfare refer- ences 15 and have the potential to shed light on the nature of Classic period warfare. Inscriptions with warfare references have been dif- ficult to interpret in terms of connotation, however, due (in part) to the lack of any connection between Maya textual records of war and archaeological data. Here we present palaeoenvironmental evidence of a massive fire event and link the event to a specific attack recorded in the epigraphic record. We show that the common Classic period war statement puluuy (it burned) was used to describe a battle that took place on 21 May 697 ce at the Classic period site of Witzna (Fig. 1; 17.423° N, 89.255° W). The archaeological evidence of this battle includes widespread destruction and burning of the major monuments across Witzna and a distinct charcoal layer deposited in a lake adjacent to Witzna. The charcoal deposit indicates that a fire event, by far the largest of the entire 1,700-yr sediment sequence, occurred during the last decade of the seventh century ce. We interpret palaeoenvironmental evidence for a dramatic reduction in local land use after the attack as reflecting decreased population. These findings provide a means to clarify the connotation of the textual data and suggest that puluuy describes acts of total warfare. Puluuy was a common war statement, and such tactics were not rare or limited to the TCP, but instead were present during the apex of Maya socio-political complexity as early as the seventh century ce. The Classic Maya name for Witzna is Bahlam Jol, as indicated by the two best-preserved monuments at Witzna (Stelae 1 and 2). Both stelae represent a ruler with the customary sceptre of the lightning god K’awiil and a shield standing on a bound captive. The name phrases of the protagonist on the stelae feature a dynas- tic ‘emblem glyph’ title of ‘holy [place-name] lord’ 16 . The toponym in the emblem glyph on Stela 2 is better preserved than the one on Stela 1 and appears to combine logograms JOL and BAHLAM (Fig. 2), indicating that, unless the monument is referencing a non-local ruler, Bahlam Jol was the toponym of Witzna during the Classic period. At Naranjo, a Classic Maya city located 32 km south of Witzna (Fig. 1), the same toponym, Bahlam Jol, appears on Stela 22 (Fig. 2). Naranjo’s Stela 22 reports a series of successful campaigns against neighbouring kingdoms, including Bahlam Jol, that allegedly once belonged to Naranjo’s greater domain 17 . The passage in blocks G8– H9 of the inscription states that “[on the day] 3 Ben, 16 Kasew (‘Sek’), Bahlam Jol ‘burned’ for the second time”. Using the calendric anchors in the text, the date of the attack has been reconstructed as 21 May 697 ce. There are no references to the timing of the first attack implied by the term ‘second time’ in the inscription. The burning of Bahlam Jol happened a year after the same type of attack against Komkom (Buenavista del Cayo) on 27 March 696 ce and a year before the conquests of K’inchil (unknown location) and K’an Witznal (Ucanal), all detailed in the monument’s text 18,19 . The connotation of the expression puluuy has been elusive. In some syntaxes puluuy refers to local fire rituals. In others, when Palaeoenvironmental, epigraphic and archaeological evidence of total warfare among the Classic Maya David Wahl  1,2 *, Lysanna Anderson 1 , Francisco Estrada-Belli  3 and Alexandre Tokovinine  4 NATURE HUMAN BEHAVIOUR | www.nature.com/nathumbehav