Robusta coffee (Coffea canephora Pierre ex Froehner) is well suited to the humid tropical climate of the Amazon basin. It is often produced under contrasting conditions of low-input agroforestry systems and intensively managed monocultures that differ in N constraints on productivity. We evaluated indicators of soil N cycling and coffee plant N sufficiency using a full factorial of two input managements (organic vs. conventional) and the absence or presence of the interplanted leguminous tree (LT) Erythrina spp. in a replicated robusta coffee field trial in the Ecuadorian Amazon. Activities of soil protease, three aminopeptidases, N-acetyl-β-D-glucosaminidase, cellobiohydrolase, and β-glucosidase were evaluated in tandem with soil NH 4 -N and NO 3 -N, potentially mineralizable N (PMN), and permanganate oxidizable C (POXC), as well as coffee leaf N, leaf N/P ratio (N/P), and yield. The LT decreased soil enzyme activities, PMN, and extractable NH 4 -N, as well as leaf N and leaf N/P. Enzyme activities and NH 4 -N were greater under organic input, but conventional input resulted in greater PMN, leaf N, leaf N/P, and yield. Permanganate oxidizable C and NO 3 -N were similar across input and interplanting but were more variable in the presence of LT relative to its absence. Lower soil enzymatic activities, labile N pools, and leaf N in coffee systems with Erythrina spp. suggest that, although a common smallholder practice in this region, interplanting this leguminous perennial has marginal impacts on soil N cycling and may not necessarily improve N supply for robusta coffee.
INTRODUCTIONCoffee is an important cash crop in many developing tropical countries as the economic basis of an estimated 20 mil-