Few scientific experiments have influenced more land than one conducted in the Wallowa Mountains of eastern Oregon by the US Department of Agriculture’s Bureau of Plant Industry and US Forest Service in 1907–1909. Four square miles of land were enclosed with a “coyote-proof fence,” guarded by a hunter, and stocked with an untended band of sheep. Data were collected on vegetation and sheep performance inside and outside the fence, and two years later success was declared. By 1910, the Forest Service had wrested range research from the Bureau of Plant Industry, subordinating the emerging field to timber production and fire suppression for decades to come. The young scientist who conducted the experiment, James Jardine, was promoted to Inspector of Grazing for the fledgling Forest Service, while his Wallowa collaborator, Arthur Sampson, went on to become “the father” of range science. The model of range management that they pioneered was applied across the US West and, later, on many rangelands in the developing world. Fencing and predator control are now generally viewed as unrelated management practices, but in the Forest Service model they were intimately connected. A critical physical geography of the Wallowa experiment reveals that the institutional context in which it occurred was more important than the findings themselves, and that although the results appeared to be scientifically rigorous and ecological, the methods were weak and the real criteria for “success” were economic. The high costs of fencing could be justified only if they were offset by a reduction in labor costs for herders. But without herders to guard the livestock, predators would have to be eliminated. Enormous public subsidies were required to implement the model, which continues to affect rangelands around the world.