Dozens of palaeoecological sequences from lakes and large swamps are available from the full range of types of tropical forest - evergreen, semi-evergreen, and deciduous - reaching from southwestern Mexico to the Amazon Basin. The data show that the development and spread of agriculture in the American lowland tropics sometimes exerted profound influences on the structure and composition of the vegetation. Regions far removed from ancient centers of civilizations such as Central Panama and parts of the Amazon Basin experienced systematic interference with tropical forest thousands of years ago. Fire was an important instrument of vegetational modification for people practicing agriculture in tropical America. Between 7000 BP and 5000 BP, depending on the region, palaeoecological records show that slash and burn cultivation had emerged in Caribbean Mexico, central Panama, the Colombian Amazon, the Ecuadorean Amazon, and probably eastern Amazonia. In some regions of Central and South America, forest clearance by agricultural populations was so severe millennia ago that defining climatic events from vegetational patterns in lake records is nearly impossible for Middle and Late Holocene periods. Often, especially in highly seasonal forests in Central America, human impacts were intense and prehistoric agriculture removed or greatly reduced in frequency a significant number of primary forest trees and shrubs. A review of some of these records follows.
Pollen and charcoal data generated in my laboratory from a large swamp called El Venancio, located in Guerrero, Mexico very close to the presumed cradle of maize domestication, provide the first indications of ancient agriculture and associated deforestation in the deciduous forests of the Central Balsas River Valley (Figure 5). Here, maize, high amounts of charcoal, and severe forest clearing are evidenced from the base of the sequence dated to 4000 BP. An older Mexican sequence generated by Kevin Pope, Mary Pohl, and John Jones from San Andres, located on the Caribbean coast of Tabasco in what would become the Olmec heartland 3000 years later, records slash and burn cultivation with maize starting at 6200 BP. Two pollen and charcoal records studied by the same group of investigators from swamps called Kob and Cobweb, located about 55 km apart in northeastern Belize, show an early phase of intensive deforestation resulting from slash and burn agriculture starting by 4000 BP (Figure 6).
A number of lake sequences generated recently from the Pacific and Caribbean watersheds of Costa Rica demonstrate slash and burn cultivation dating from 4800 BP to 2950 B. P. Maize is present in the earliest deposits of all of them. Moving further south to Lake La Yeguada, central Panama, the initiation of slash and burn agriculture is indicated by 7000 BP, when primary forest trees decrease greatly in the phytolith record and pollen from early secondary woody growth (Cecropia, Ficus) increases significantly (Figures 3 and 4). Shortly after this time maize pollen is first recorded and it continues to be present throughout the sequence. Cucurbita (squash) phytoliths make an appearance at about 6000 BP. At 4200 BP even the secondary growth trees and charcoal decline greatly, indicating that a larger portion of the woody growth was being cut and could no longer contribute charcoal to the record because fallow periods were probably being shortened (Figures 3 And 4). This region of Panama has also seen longterm archaeological research by Anthony Ranere and Richard Cooke and its pre-Columbian cultural records are among the best in the lowland Neotropics. Excavations and systematic archaeological foot surveys indicate that sites identified as hamlets and hamlet clusters near La Yeguada during the 7000 BP to c. 5000 BP period are very similar in size and other aspects to those of modern slash and burn agriculturists, and that site numbers increased 15-fold at that time compared with the 10 000 to 7000 BP period. Several of these occupations have yielded phytolith and starch grain records of agriculture from 7000 to 3000 BP, making it clear that agriculture supported ever increasing human populations, who in turn placed increasing pressure on their environment. Here is a case where combined archaeological and palaeoecological investigations provided a particularly detailed picture of tropical forest history.
Moving further south into the Amazon Basin and other areas of South America, a varied scale and intensity of human forest disturbance becomes more apparent. For example, disturbance resulting from human agricultural pressure started in the western and eastern Amazon Basin (Lakes Ayauchi and Geral, respectively) during the sixth millennium BP and intensified during the next few milleniums. However, periods of particularly intensive agriculture occurred between c. 3350 BP and the Conquest period, later than in sequences just discussed from Central America, and during the maximum periods of disturbance there are never the declines in tree pollen and phytoliths nor increases in frequencies of herbaceous plants that indicate large-scale destruction of the forest, as occurs in many Central American sequences.
Records recently generated by various investigators for the Chocio, western Colombia and the Rio Napo region of the Ecuadorian Amazon are examples of even more minimal human disturbance on very wet, aseasonal, and still densely forested northern South American landscapes. In these cases, maize is not recorded until c. 1000 BP in a 4200 year-long sequence, but this signal of agriculture is still not accompanied by signs of significant forest clearing (the Chocoi record). Charcoal is common during certain periods, but neither maize nor signs of forest clearing in the way of declining tree frequencies are recorded at all over more than 10 000 years of time (Rio Napo area). Moreover, three sequences now available from high rainfall areas in northern Brazil dating to the past 10 000 years and more record no burning, cultivar presence, or tree felling at all.
The Central American sequences contain evidence for significant to severe forest clearing more often than do existing South American records. When the ecological settings of the sites are compared it can be seen that lake records from Central America that demonstrate the most intensive disturbance (El Venancio, Mexico; Belize, Costa Rica, La Yeguada, Panama) are located in areas where the vegetation is or was highly seasonal tropical forest. In other ecological settings in Central and South America, especially those that supported wet, nonseasonal, evergreen formations, human disturbance is often less severe and even undetectable.
World History
