Much has been written about the chinampas of central Mexico. One of the commonly repeated themes is that these wetland fields were self-irrigated from below in a process known as sub-irrigation. According to this model, water infiltrates the planting platforms from adjacent canals and then rises to the root zone by capillary action. Thus, chinampas are thought to have needed little supplemental irrigation, and produced dependable and high yields. Here I report the results of field and lab studies of soils (...) of 10 chinampas in the region east of Xochimilco, in the Mexican Federal District. These interpretations are combined with discussion of maize rooting tendencies, regional rainfall patterns, and the presence of willow tree roots in chinampa soils. I also consider whether farmers strive to create and maintain optimum field morphologies for subirrigation when faced with both flood and drought risk, and when manual irrigation is aready option. I conclude that while sub-irrigation is possible under certain combinations of field morphology, crop type, and soil properties, it was a relatively minor factor in the overall decision-making process of past chinampa farmers, and should not be a major determinant of future planning for chinampa preservation or reconstruction. Furthermore, sub-irrigation must be modeled carefully in relation to soil properties, croptypes, and seasonality if it is to beconsidered a significant function of wetland fields in any context. (shrink)