Chrysothamnus viscidiflorus

We found no significant effects of Phos-Chek or Silv-Ex on number of C. viscidiflorus plants per plot for either June or July applications (Table 2). Growth and resprouting were also unaffected by any chemical application (Table 2).

Change in Number of Stems/m²

We found significant treatment by sample period interactions in the change in number of stems/m² between pre- and post-treatment for both June and July applications in the riparian zone (F = 7.05; df = 7, 6; P = 0.01 and F = 4.39; df = 7, 6; P = 0.04 for June and July, respectively; Figure 1). Plots treated with 1.0% Silv-Ex had significantly fewer stems than the control on burned and unburned plots through 13 weeks after June applications. All other treatments were indistinguishable from the control at 13 weeks post-treatment. Number of stems on upland plots did not differ significantly in response to any chemical treatment (F = 1.23; df = 7, 6; P = 0.41 and F = 1.02; df = 7, 6; P = 0.50 for June and July applications, respectively).

Figure 1.  Change in stems/m2 between pre- and post-treatment on unburned and burned riparian vegetation plots. Shown are least square means + one standard error of the mean for each chemical treatment. Means with the same letters do not differ significantly from other means at that sampling period using Fisher's LSD test.

Change in Species Richness, Evenness, and Diversity

We found a significant treatment by sample period effect for species richness for both June and July applications in the riparian zone (F = 7.27; df = 21, 18; P = 0.0001 and F = 2.56, df = 21, 18; P = 0.02 for June and July applications, respectively; Figure 2). Overall, burned plots tended to gain species over the course of the study, while species richness on unburned plots remained relatively stable. Plots treated with Phos-Chek were an exception, however. Species richness declined more in plots treated with Phos-Chek, but not burned, than in plots subjected to any other non-burned treatment. Although species richness in Phos-Chek plots showed a significantly greater decline than either the Silv-Ex or control treatments at intermediate sampling periods, no differences were observed among any treatments by the 13^th week after June applications. The same trend was evident in Phos-Chek plots after July applications (Figure 2b), but chemically treated plots did not differ significantly from control plots in any sampling period.

Figure 2.  Change in species richness between pre- and post-treatment on unburned and burned vegetation plots. Shown are least square means + one standard error of the mean for each chemical treatment. Means with the same letters do not differ significantly from other means at that sampling period using Fisher's LSD test. See legend in Figure 1 for key to treatments.

We found a significant treatment by sample period interaction in change in species richness on upland plots after June (F = 2.29; df = 21, 18; P = 0.04) but not July (F = 1.29; df = 21, 18; P = 0.29) treatments. Unburned plots showed little trend through the season, and chemically treated plots were never statistically different from control plots at any sampling period (Figure 2c). Burned plots showed a slight tendency to gain species over the course of the study, particularly on plots treated with 0.5% Silv-Ex.

Shannon’s index of species diversity (H’) and Pielou’s index of evenness (J’) did not differ among treatments in either upland or riparian habitats after either June or July applications (Table 3).

Community Characteristics

The canonical variate analysis suggested that the chemical treatments had relatively little effect on the community composition (Figures 3 and 4). In each of the four analyses, the first axis tended to separate pre-treatment from post-treatment samples, with pre-treatment conditions associated with higher values of axis one. This axis accounted for 48% of the variation in riparian plots after June application and 47% after July application, 85% in upland plots after June application, and 63% after July application. In riparian habitat, the first axis was most strongly correlated with Poa pratensis (r = 0.86 for June and r = 0.89 for July applications), Carex praegracilis (r = 0.75 for June and r = 0.72 for July), and Juncus balticus (r = 0.75 for both June and July). In upland habitat, the first axis was most strongly correlated with Poa secunda (r = 0.84 for June and r = 0.70 for July) and Agropyron trachycaulum (r = 0.72 for June and r = 0.68 for July).

Figure 3.  Results of canonical variate analysis for burned and unburned riparian vegetation treated with 0.5% Silv-Ex, 1.0% Silv-Ex, Phos-Chek, or water. Symbols on the graph refer to treatment (B = burned, 5 = 0.5% Silv-Ex, 1 = 1.0% Silv-Ex, P = Phos-Chek) and sample period (-1 = pre-treatment sample; -4 = fourth sample period, 8 weeks after June applications and 6 weeks after July applications; and -5 = fifth sample period, 13 weeks after June applications and 11 weeks after July applications). Dashed lines enclose (a) pre-treatment, post-treatment burned and post-treatment unburned plot means; and (b) pre-treatment and final post-treatment plot means.		Figure 4.  Results of canonical variate analysis for burned and unburned upland vegetation treated with 0.5% Silv-Ex, 1.0% Silv-Ex, Phos-Chek, or water. Symbols are as in Figure 3. Dashed lines enclose (a) pre-treatment, post-treatment burned, and post-treatment unburned plot means; and (b) pre-treatment, fourth post-treatment, final post-treatment burned, and final post-treatment unburned plot means.

The second axis, which accounted for 14% and 13% of the variation in riparian plots after June and July applications, respectively, separated post-treatment burned from post-treatment unburned plots, although the unburned 0.5% Silv-Ex plots were more similar to burned plots than to other unburned treatments after June applications (Figure 3a). For June applications, the second axis was most strongly correlated with Taraxacum officinale (r = 0.56) and Thermopsis montana (r = 0.50); for July applications, the second axis was correlated with Agropyron trachycaulum (r = 0.62). By the 11th week after July applications, riparian plots all had similar communities, as defined by the first two axes of the CVA, although unburned plots were all > 1 and burned plots were all < 1 on CV1, implying somewhat greater recovery of the dominant grass and sedge species on unburned plots (Figure 3b). In contrast, burned and unburned plots were still mainly separated on both axes at 13 weeks after June applications (Figure 3a).

In upland habitat, the second axis accounted for 5% and 15% of the variation in plots after June and July applications, respectively. The second axis was most strongly correlated with Achillea millefolium (r = -0.66) and Artemisia tridentata (r = 0.46) for June applications, and with Achillea millefolium (r = 0.53) for July applications. As in the riparian zone, the second axis tended to separate burned from unburned treatments, although more completely for June applications (Figure 4a) than for July applications (Figure 4b); 11 weeks after July applications, 1.0% Silv-Ex plots were more similar to burned plots than to other unburned treatments.

Artemisia Study

The 40 Artemisia tridentata shrubs that we treated did not vary significantly in height (F = 1.45; df = 3, 36; P = 0.24), volume (F = 0.04; df = 3, 36; P = 0.98), or amount of chemical applied (F = 0.44; df = 3, 36; P = 0.72). We found no significant effect of any chemical treatment on growth, flower production, or galling insect activity (Table 4).