sample plot consisted of: diameter at breast height 

 (d.b.h.) of live and dead trees over 2 inches (5 cm) d.b.h.; 

 stem count of live and dead reproduction in classes of 

 less than 20 inches (0.5 m) tall, 20 inches to 6.6 ft 

 (0.5 m to 2 m) tall, and greater than 6.6 ft (2 m) tall up 

 to 2 inches (5 cm) d.b.h., and number of stumps and 

 diseased trees (primarily conks). Species numbers, 

 composition, and frequency of the understory were deter- 

 mined by 33 nested frequency frames in each sample 

 plot (U.S. Department of Agriculture 1983). Vegetation 

 litter, rock, and bare ground were estimated from six 

 points per frequency frame on the sample plot. Under- 

 story biomass was determined by both clipping and 

 estimating current year's growth, up to 4.9 ft (1.5 m) 

 high, by vegetation categories. Four sets of microplots, 

 were distributed randomly on the 1,076 ft 2 (100 m 2 ) 

 macroplot. Each set consisted of five circular 5.4 ft 2 

 (0.5 m 2 ) microplots clustered so that the biomass of four 

 could be estimated as a percentage of the fifth, which 

 was then clipped. The clipped vegetation was dried at 

 least 48 hours in an oven at 158 °F (70 °C) and weighed 

 for biomass. Percentages of the shrub, grass, and forb 

 components were estimated on all five microplots. The 

 dry weight biomass by vegetation categories is based on 

 20 microplots per stand. Photographs of the overstory 

 and understory vegetation were taken on all plots. 



On the plots sampled less intensively, the tree and 

 environmental data were determined the same as on the 

 more intensively sampled plots. The understory species 

 cover was ocularly estimated on the entire 1,096 ft 2 



(100 m 2 ) plot. The sets of microplots to determine 

 understory biomass were estimated by vegetation 

 categories. 



More sprayed than unsprayed stands were sampled 

 because variation was expected to be greater in the 

 sprayed stands. All plots were used in an analysis using 

 the one-tailed t-test for independent means. 



RESULTS AND DISCUSSION 



We sampled 17 sprayed areas that met our criteria in 

 Idaho, Wyoming, Colorado, and Utah (fig. 2); 34 sprayed 

 plots and 22 unsprayed plots for comparison were 

 sampled (table 1). Sagebrush was the target species for 

 most of the herbicide treatments (table 1). Aspen was 

 sprayed for stand regeneration on four areas and for 

 conifer release on two areas. The herbicide most com- 

 monly used was 2,4-D in high volatile butyl ester formu- 

 lations mixed with water or diesel. The indicated rates of 

 2 lb/acre (2.2 kg/ha) are possibly higher than that actu- 

 ally applied in the aspen stands associated with sage- 

 brush because of drift and volatilization. Quality control 

 and equipment capabilities would also be factors in non- 

 target applications. Most of the sagebrush spray 

 projects in which aspen was hit used fixed-wing aircraft 

 in the late 1950's and early 1960's. The rarity of sprayed 

 aspen stands in the 1970's when sagebrush was the 

 target species coincides with the use of helicopter and 

 low-volatile esters. Volatilization effects on nontarget 

 species were reduced markedly when low volatile esters 

 replaced the high volatile esters. 



Figure 2.— Locations of herbicide treatments in Western United States. 



2 



