Although Lynch (1990) argued for a trans-altiplanic entry from north to south for the entry of humans into the Andean highlands, consensus has emerged that the earliest inhabitants of the highlands came from the Pacific littoral. Evidence for this is found in three sites: Quebrada Jaguay in southern Arequipa, Asana in the Rio Osmore drainage of Moquegua, and Acha-2 in northern Chile. At Quebrada Jaguay, Sandweiss et al. (1998) discovered that the inhabitants of this coastal site most probably made logistical (transhumant, in their terminology) forays into the highlands as early as 11,000-9500 BC and obtained obsidian from the Alca source, which is found at 2,850 masl. Aldenderfer (1998) has observed a similar pattern at Asana. The assemblages that pertain to the earliest occupation of the site (Puruma Phase, 9500-8900 BC) contain small quantities of coastal lithic raw materials, and I attribute their presence to logistical forays by lowlanders making initial explorations of the highlands. In the following Khituna Phase (8900-7600 BC), the assemblages contain only raw materials from high elevation sources, suggesting that the foragers at Asana no longer made long-distance forays to the coast and were now living at high elevations permanently. At Acha-2, a coastal site on the Rio Azapa, Munoz et al. (1993) have observed similar regional-scale settlement patterns at around 8100 BC.
Despite claims for a very early occupation (18,000 BC) at Pikimachay in the Ayacu-cho Valley by MacNeish et al. (1980), most scholars working in the region agree that the highlands are not likely to have been occupied before 10,000 BC. Prior to this date, the high Andes were humid but significantly colder than at present, which would have depressed primary productivity. Around 10,000 BC in the northern Andes and slightly later in the southern Andes, deglaciation begins and conditions for human occupation improve significantly. Familiar plant and animal species would have begun to move to high elevations, and new and attractive resource patches would have been established quickly. However, a trend toward aridity develops that is exacerbated by shorter and cooler austral summers until 8000 BC. This 2,000-year span, then, saw both opportunities and constraints develop for the human populations beginning to move into the highlands (Aldenderfer 1999b).
The data from Asana suggest that the move to a permanent habitation of the highlands was rapid, perhaps culminating in no more than 500 years once their initial “discovery” had been made. Artifacts likely used by both males and females are present at Khituna phase Asana, suggesting the entire co-resident group is present, including children and older adults. The discovery and use of the Andean highlands was likely to have been driven by short-term decisions designed to minimize effort while simultaneously ensuring sufficient caloric returns for all members of the co-residential group. The resource pull of the rapidly changing highlands, especially as regards faunal resources, would have permitted foragers to achieve these goals as newly emergent highland animal and plant communities became more productive and stable. In the earliest stages, such as the Puruma phase at Asana or the earliest phase of occupation at Pachamachay (Rick 1980), land use patterns of the highlands would have been made on a logistical basis only, with hunting parties of adult males moving into the highlands and returning to lower elevation base camps with their kills. However, it would not have been possible to increase or even maintain for a long period returns under this system because of limitations in transport technology and the inability of members of the foraging party to overcome the effects of hypoxia that limited their work capacity.
A more permanent occupation of the highlands was probably driven by the eventual stabilization of resource patches within them. As more permanent and productive patches emerged on the landscape, their pull would have been sufficient to draw the entire coresidential group to them. Note that this does not imply resource shortfall in lower elevations but instead stresses that these changes would have permitted the entire group to obtain sufficient caloric return while simultaneously reducing work effort for the group’s most productive members—its hunters. The rapidity of this process, however, would have been contingent upon how quickly these new patches stabilized as well as the mix of floral and faunal resources present in any locality.
The shift to a smaller scale of residential mobility is consistent with the predictions of the model of high elevation foraging. Although the effects of hypoxia would have led to a reduction in work capacity, this negative was balanced by the growing size and density of resource patches in the highlands. However, the hidden costs of early high elevation life— effects on fetal growth and pregnancy—would not have been noticed immediately. Women would have experienced more fetal loss, and successful pregnancies would have been characterized by infants with lower birth weights when compared to lowlanders, which would have led to higher rates of neonatal mortality. And although birth at high elevation confers at least partial biological acclimatization to the rigors of high elevation life, recent data from Ladakh in northwestern India suggest that it may take at least 60 generations for pregnancy-related acclimatization to become widespread in the population (Wiley 2004). The effect of this would have been relatively slow rates of population growth across the highlands, an observation supported in all instances where we have good regional survey and reliable population estimates, such as on the Junin puna (Rick 1980), the Rio Moque-gua drainage (Aldenderfer 1998), and the Rio Ilave drainage in the circum-Titicaca region (Aldenderfer 2002; Klink 2005).