Relationship between Fecal Components and Forage Consumption and Digestibility

Abstract

Thirty-nine in vivo total collection digestion determinations were performed with steers with the purpose of relating a selected array of fecal components to forage intake and digestibility. These determinations were made during the spring, summer and fall of 1978 with eight steers averaging 218 kilograms. The forages were freshly cut fescue and frozen green legumes (crimson clover, three cuttings of red clover and lespedeza). Composite fecal samples were analyzed for dry matter (DM), N, ether extract (EE), crude fiber (CF), ash, cell wall constituents (CWC), acid detergent fiber (ADF), acid detergent lignin (ADL), acid insoluble ash (AIA), urobilinogen (UROB), in vitro DM digestiblity (IVDMD), Na and Zn. Hemicellulose, cellulose, acid soluble ash (ASA) and N-free extract (NFE) were calculated. Fecal components related (P<.10) to DM intake (DMI) were: DM, N, CF, EE, NFE, ASA, IVDMD and Zn. Those related (P<.10) to DM digestibility (DMD) were: DM, N, CF, ash, NFE, ASA, IVDMD, CWC and Zn. The best multiple regression equations (highest R2, with all independent variables significant) for estimating DMI, fecal DM output (FDMO), DMD and digestible DMI (DDMI) had R2 values of .69, .35, .65 and .74, respectively. The respective residual standard deviations (RSD) for these models were: .46 kg/day, .20 kg/day, 7.07% and .41 kg/day. The equation for estimating DDMI was: DDMI = −6.66 + 1.65(N) + .37 (EE) + .08(CWC) − .08(DM). The equation, DDMI = −1.13 + 1.32(N), had an R2 of .44 and a RSD of .58 kg/day. The addition of between-animal variation to these models improved (P<.01) R2 only for DMI and DDMI, accounting for about 11% of the variation in each variable. Most of this between-animal variation was due to variation in weight. Fecal N indices do not have broad application, because the relationship between fecal N and forage intake or digestibility changes as season of year, N in forage and species consumed change, but the addition of EE, CWC and DM to models containing N largely overcomes these problems. Therefore, the development of useful fecal indices with fairly broad application appears possible.