Long-term herbaceous response data following herbicidal treatment of honey mesquite (Prosopis glandulosa Torr.) are needed to develop more accurate projections regarding economic feasibility of these treatments and to model ecological interactions between woody and herbaceous plants in rangeland systems. Our objective was to measure herbaceous yield and mesquite regrowth 10 or 20 years after mesquite was aerially sprayed with either mesquite top-killing or root-killing herbicides. Treatments evaluated included mesquite top-killing herbicides at 10-12 years (T10) and 19-21 years (T20) post-treatment, mesquite root-killing herbicides at 10-12 years (R10) and 19-21 years (R20) post-treatment, and an untreated control where mesquite were >30 years old (C30). Treatments were applied in the late 1970's or late 1980's. Grass yields, measured annually from 1998 through 2000, were quantified within patches of 3 perennial grass functional groups: cool-season mid-grasses, warm-season mid-grasses, or warm-season short-grasses. Coolseason annual grass yields were also quantified within these perennial grass patches. By 1998, mesquite canopy cover was 55, 47, 36, 24, and 12% in C30, T20, T10, R20, and R10 treatments, respectively. Warm-season mid-grass yields were most sensitive to differences in mesquite cover in all 3 years and declined sharply when mesquite cover exceeded 30 %. Cool-season midgrass yields declined slightly with increasing mesquite cover. Warm-season short-grass and cool-season annual grass yields were not related to mesquite cover, except in 2000 when warmseason short-grass yield beneath mesquite canopies increased with increasing mesquite cover. Results suggest that herbicide treatment life (defined by increased perennial grass yield in response to mesquite treatments) was at least 20 years for the root-killing herbicide, but no longer than 10 years for the topkilling herbicide.

Assessment of herbaceous standing crop in heterogeneous range plant communities requires large numbers of samples to account for inherent variability. The dry-weight-rank method (DWR) was developed to eliminate the need for clipping and sorting of herbage to determine relative proportions on a dry weight basis. The technique was assessed for applicability and accuracy in the mixed prairie of the Texas Rolling Plains. Much of the herbage within the communities investigated occurred in monospecific patches that resulted in only 15% of quadrats having 3 species ranked for which DWR was designed. Non-harvest methods of determining grass proportion by species were compared to harvested proportions in mesquite (Prosopis glandulosa Torr.) and redberry juniper (Juniperus pinchotii Sudw.) communities. Estimation methods evaluated were 1) harvest by species, 2) weight estimation by species, 3) DWR with quadrat weighting, 4) unweighted estimated proportion by species, and 5) unweighted DWR. Correlations of non-harvest to harvest proportions were improved with quadrat weighting. Weighting improved values more in the juniper than in the mesquite communities. Although cumulative ranking of DWR multipliers was necessary in 85% of sample quadrats, there was a high correlation (r 2 >0.995) between weight estimation and weighted DWR and between estimated proportion and unweighted DWR. This indicates that cumulative ranking with the original DWR multipliers was virtually the same as evaluator estimation. Analysis of variance indicated significant differences in nonharvest methods compared to harvesting. Quadrat weighting with DWR was necessary to draw the same statistical conclusions between means that harvest data provided. Ranks are easier to apply and more likely to be applied similarly by individual evaluators than estimated proportions. For sites with high standing crop variation and patchiness of species that require considerable use of cumulative ranking, DWR with quadrat weighting provides adequate determination of species proportions of biomass.

This paper presents a comparative simulation analysis of the economics of prescribed fire and aerially applied root-killing herbicide treatment as methods for maintaining livestock productivity on rangeland in the Texas Rolling Plains. A "no-treatment" scenario is used as the base for comparison. In almost all the simulated scenarios both herbicide application and prescribed burning were economically feasible since net present values were > 0 and benefit/cost ratios were >1. However, the net present values for prescribed fire were much higher that those for the herbicide treatment even with a lower increase in carrying capacity with burning. The cost of herbicide would have to be less than half the current cost of $57 ha-1 before it would be economically competitive with fire in controlling mesquite. If cattle numbers were not increased after treating brush, burning had an even greater net present value and benefit/cost ratio advantage over herbicide treatment than if cow numbers were increased after treatment. Even if fences have to be constructed to implement adequate deferment for burning, the net present value and benefit/cost ratios of the fire option were higher than those for herbicide scenarios. This analysis indicates that there is an economic advantage to using fire wherever possible, and use of herbicides is restricted to those instances when fine fuel amount is < 1,700 kg ha-1 yr-1 when fire is not a viable option. The analyses indicate the economic response is most sensitive to the treatment effect on wildlife income.