Maintaining native plant diversity, detecting exotic species, and monitoring rare species are becoming important objectives in rangeland conservation. Four rangeland vegetation sampling techniques were compared to see how well they captured local plant diversity. The methods tested included the commonly used Parker transects, Daubenmire transects as modified by the USDA Forest Service, a new transect and "large quadrat" design proposed by the USDA Agricultural Research Service, and the Modified-Whittaker multi-scale vegetation plot. The 4 methods were superimposed in shortgrass steppe, mixed grass prairie, northern mixed prairie, and tallgrass prairie in the Central Grasslands of the United States with 4 replicates in each prairie type. Analysis of variance tests showed significant method effects and prairie type effects, but no significant method X type interactions for total species richness, the number of native species, the number of species with less than 1% cover, and the time required for sampling. The methods behaved similarly in each prairie type under a wide variety of grazing regimes. The Parker, large quadrat, and Daubenmire transects significantly underestimated the total species richness and the number of native species in each prairie type, and the number of species with less than 1% cover in all but the tallgrass prairie type. The transect techniques also consistently missed half the exotic species, including noxious weeds, in each prairie type. The Modified-Whittaker method, which included an exhaustive search for plant species in a 20 x 50 m plot, served as the baseline for species richness comparisons. For all prairie types, the Modified-Whittaker plot captured an average of 42.9 (+/- 2.4; 1 S.E.) plant species per site compared to 15.9 (+/- 1.3), 18.9 (+/- 1.2), and 22.8 (+/- 1.6) plant species per site using the Parker, large quadrat, and Daubenmire transect methods, respectively. The 4 methods captured most of the dominant species at each site and thus produced similar results for total foliar cover and soil cover. The detection and measurement of exotic plant species were greatly enhanced by using ten 1 m2 subplots in a multi-scale sampling design and searching a larger area (1,000 m2) at each site. Even with 4 replicate sites, the transect methods usually captured, and thus would monitor, 36 to 66% of the plant species at each site. To evaluate the status and trends of common, rare, and exotic plant species at local, regional, and national scales, innovative, multi-scale methods must replace the commonly used transect methods of the past.