Tanzania Journal of Agricultural Sciences 2005, Vol. 6(2) : 80-91An experiment (5 x 5 Latin Square) was conducted to estimate the physical fill of tropical forages and maximum intake capacity of five mature non-pregnant crossbred dairy heifers kept under zero grazing system. Five (5) forages [Brachiaria brizantha hay (BH), Maize (Zea mays) silage (MS), Lucerne (Medicago sativa) hay (LH), Lucerne and grass hay (LGH) and urea-treated rice straw (UTRS)) were used. The forages were fed as sole diets ad libitum with only minerals and vitamins supplements. Water was provided all the time and measured daily. Voluntary feed intake (VFJ) and faecal output were measured for seven days. Degradability characteristics were obtained in situ using the nylon bag technique. Rumen pool size of NDF was measured by rumen evacuation technique. Passage rates were calculated based on faecal output and rumen pool size of NDF. There was a marked difference (P<0. 000l) between the rate and extent at which NDF for tropical forages was degraded. The rate of passage (% h-1) was different (P<0.03) between forage diets with values ranging from 1.4 to 1.8 for MS and UTRS, respectively. Rumen pool sizes of NDF were different (P<0.01) between forage diets and weighed 4.8, 3.8, 4.7, 5.2 and 4.5 kg for BH, MS, LH, LGH and UTRS, respectively. Fill (days) also varied between forage diets and ranged from 1.4 for UTRS to 1.8 for MS. The intake capacity of animals for NDF were different (P<0.01), highest in animals fed LGH (5.0 kg-1) and lowest in animals fed MS 3.1 kg-1). Using predicted NDF intake (PNDFJ) based on NDF degradability characteristics and passage rates derived from faecal output and rumen pool size of NDF, good prediction of dry matter intake was obtained (R2 = 0.70). It was concluded that a system of describing the physical fill of NDF in tropical forages could be used to predict VFI in cattle.