MAINTENANCE OF YIELDS AND SOIL FERTILITY IN NONMECHANIZED CROPPING SYSTEMS, BOLIVIA

Slash and burn farmers in the tropical rain forests of eastern Bolivia are abandoning land after one rice (Oryza sativa L.) crop because of declining yields. A trial was conducted on a Typic Paleudult for 41 mo to investigate whether alternative low-input nonmechanized cropping s ystems could prolong soil fertility and yield maintenance, and whether soil fertility or weeds were responsible for declining yields. Twelve cropping systems were investigated in a factorial design, with three summer-winter crop sequences: rice-peanut (Arachis hypogaea L.), com ( Zea mays L.)-bean (Phaseolus vulgaris L.) later substituted by cowpea [Vigna unguiculata (L.) Walp.], and rice-fallow (control); two weed co ntrol treatments: minimal and optimal; and two fertilizer treatments: with and without 60 kg N ha-1 and 17.5 or 35 kg P ha-1. Crop sequences significantly increased exchangeable acidity; the rice sequences sign ificantly reduced exchangeable Ca, and cornbean/cowpea and rice-peanut significantly reduced exchangeable Mg. Fertilization significantly in creased soil P but decreased Ca. Foliar analysis revealed N, Mg, and Z n deficiencies in all cropping systems. Rice yields, unlike corn, were significantly increased by optimal weeding. Corn yields were dominate d by fertilization, whereas rice yields were mainly influenced by fert ilization in the first and fourth years, and by weeds in the interveni ng years. Without fertilizers, rice-fallow was not sustainable, and on ly corn-bean/cowpea was sustainable for 3 yr. With fertilization, rice -fallow plus optimal weed control and corn-bean/cowpea with minimal or optimal weeding were sustainable for 3 yr. Additional fertilization a nd future liming would be necessary for more prolonged sustainability.