This experiment was carried out at the Faculdade de Ciências Agrárias Veterinárias - Unesp - Jaboticabal during dry season, to evaluate the effects of sampling methods on the Marandu grass dry matter and cell wall in situ degradability. The data were analyzed according to a randomized block design with three replications (paddock), in a split plot scheme. Five sampling methods were evaluated on the parcels: metallic square method; Nellore steers extrusa, Red Angus x Nellore crossbreeding steers extrusa, simulated grazing of the Nellore, simulated grazing of Red Angus x Nellore crossbreeding, and two years (2003 and 2005) on the subparcels. The following rumen kinetic fractions were determined: soluble "A", degradable potential "B", degradability rate "Kd", the potential degradability (DP) and no degradable fraction "C" of the FDN, and FDA. According to the results obtained, the metallic square method underestimated the characteristics of the forage degradation. The results of the simulated grazing method resembled the extrusas method; however, it was the simulator practice that assured efficient sampling, as shown by evidence from data obtained between 2003 and 2005.
Brachiaria brizantha (Hochst. ex A. Rich.) stapf; hand-plucked; extrusa; square
