The legumes used for green manure have been introduced in agricultural systems for recovering degraded soils by intensive use. The objective of this work was to evaluate the response of herbaceous legumes to liming and P application on dry matter and its nutritional efficiency. The experiment was developed in greenhouse at the Soil Science Department of the Federal University Agricultural of Amazon (Brazil). Soil used was a dystrophic Yellow Latosol, sampled in the 0 - 20 cm depth. A randomized block experimental design was used with treatments arranged in a factorial 3³, comparing three legume species: velvet bean (Stizolobium atterrimum), "mucuna cochinchinensis" (Stilozobium cochinchinensis) and jack bean (Canavalia ensiformes); three doses of liming (0, 4 and 8 t ha-1 of limestone) and of triple superphosphate (0, 25 and 45 mg P dm-3 of soil), with four replications each one. Results showed that "mucuna cochinchinensis" and jack bean produced higher amounts of dry matter than velvet bean. The first species accumulated higher amount of P and the second species higher amount of Ca in the tissue. The "mucuna cochinchinensis" species showed higher use efficiency of Ca, Mg and P in the absence of liming. The jack bean and mucuna cochinchinensis can have better performance than velvet bean, in weathered soils with low concentration of P, Ca and Mg, since they presented higher translocation efficiency and use of these nutrients, respectively.
Liming; phosphate fertilization; nutrient uptake; nutrient translocation; nutrient use efficiency