RESUMEN
We examined the nutritive value of common grass species in the semi-arid rangelands of Borana in southern Ethiopia using local experience based herbage preference (LEBHP) perception and laboratory techniques. Local pastoralists in the study area were asked to identify common grass species and rank them according to the species' preferences and palatability to cattle. The pastoralists listed a total of 15 common grass species which were then sampled during the main rain and cold dry seasons and analyzed for crude protein (CP), Neutral Detergent Fiber (NDF), Acid Detergent Fiber (ADF) and ash content to verify pastoralists' claim regarding the quality of individual species. The relative feed value (RFV) and dry matter digestibility (DMD) were also calculated using NDF and ADF contents. Spearman's rank correlation was used to examine possible relationships between laboratory results and pastoralists' experience on grass quality. Cenchrus ciliaris, Chrysopogon aucheri, Digitaria milanjiana, Eragrostis papposa and Panicum maximum were the top five species based on LEBHP perception. There were indications of inconsistency in terms of LEBHP perception among the different pastoral communities. The chemical composition of all grass species showed significant (p<0.05) variation between sites, seasons and species. The results showed that the CP values for the Borana rangelands were in the range of 8.7% in the main rain season to 5.1% for the cold dry season. The fiber constituents were relatively low in the main rain season compared to the cold dry season. Overall, Digitaria milanjiana had the highest CP (16.5%) content, while the least was recorded with Heteropogon contortus (10.8) and Aristida adoensis (9.8%) during the main rain season. It seems that the spatial variability of landscapes within the wider geographical regions, soil properties and texture, and land-use patterns probably contributed to site differences in species quality. Generally, the RFV of individual grass species was significantly (p<0.05) varied between and within sites. The ranking of species by pastoralists according to their preferences by cattle was highly correlated with the chemical composition of laboratory results of individual grass species with 'r' values for CP (0.94), ash (0.95), NDF (-0.98), ADF (-0.93) and ADL (-0.93). We suggest the complimentary use of LEBHP and laboratory techniques in evaluating the nutritive quality of rangeland forage species for sustainable animal production.