For decades, the dominant paradigm has been that the first Americans were descendants of populations that migrated from northeast Asia to North America by crossing the now-submerged Bering Land Bridge around 13,000 years ago. Over many generations, these populations would have travelled through an ice-free corridor between the Canadian ice sheets to the south and west to populate North, Central, and South America. Growing archaeological evidence, however, suggests that humans may have arrived many thousands of years earlier than originally thought, perhaps travelling by boat along the Pacific Coast. The often-spirited debate over the timing, manner, and location of arrival of the first humans in this continent has relevance for related areas of research, including the factors involved in the extinction of the American megafauna.

Our recent study at Coxcatlan Cave in the Tehuacan Valley of Puebla, Mexico provides new data relevant to this debate. The archaeologist Richard S. MacNeish was the first to lead excavations of Coxcatlan Cave in the early 1960s. Its location within the arid Tehuacan Valley, today a UNESCO World Heritage site, resulted in exceptional preservation of organic materials, including some of the oldest recorded specimens of domesticated corn, chili peppers, and avocados. Radiocarbon carbon dating registered approximately 10,000 years of human occupation of the cave, over which time the local communities transition from hunting-gathering lifestyles to those based on agricultural food production. Research at Coxcatlan revolutionized our understanding of the origins of agriculture in the Americas and defined the cultural adaptations of the Mesoamerican Archaic period. The lowest levels of the cave, however, were not subjected to radiocarbon dating, leaving open the question of when humans first arrived.

As part of a larger study of the ancient environment of the Tehuacan Valley, our team used accelerator mass spectrometry (AMS) dating on a series of animal bones recovered by MacNeish’s excavations, including six rabbit bones from the lowest levels of the cave. Although most of the 14 AMS ages were similar to those produced by previous studies, the results from the lowest levels were surprisingly old. The earliest occupation was estimated to have occurred between 33,448 and 28,279 years before present (cal BP), a time nearly 20,000 years older than the traditional model of the peopling of the Americas. Nevertheless, the dated specimens were documented in association with 14 stone tools and hundreds of apparently butchered animal bones. Notably, these new radiocarbon ages overlap with those from the recently excavated Chiquihuite Cave in Zacatecas, Mexico, providing a second possible example of humans in Mexico prior to the Last Glacial Maximum.

The early radiocarbon ages from Coxcatlan Cave require us to carefully scrutinize the associated artifacts and bones to determine whether the lowest cave levels do in fact represent human occupational episodes, or whether they were naturally accumulated layers. Future studies will reassess the archived faunal bone collection for evidence of human manipulation and the stone tool collection for signs of human manufacture. Although MacNeish’s excavations occurred 60 years ago, Coxcatlan Cave continues to yield exciting new data relevant to some of the largest questions in archaeological research today.

The full paper New AMS Radiocarbon Ages from the Preceramic Levels of Coxcatlan Cave, Puebla, Mexico: A Pleistocene Occupation of the Tehuacan Valley? by Andrew D. Somerville, Isabel Casar and Joaquín Arroyo-Cabrales has been published open access in Latin American Antiquity.