In most respects, the effects of Phos-Chek, and 0.5% or 1.0% Silv-Ex on vegetation in our Great Basin study sites did not vary substantially from each other or from the control. The lack of difference among treatments is best illustrated by the CVA, which tended to separate plots based on phenology and burning, rather than chemical treatment (Figures 3 and 4). With few exceptions, effects we observed did not persist to the end of this 1-year study. Effects that did persist were related to Silv-Ex applications. Unburned and burned riparian plots treated with 1.0% Silv-Ex in June still had a greater deficit in total live stems than any other treatment at the end of the study (Figure 1a and 1b). At the community level, after July applications on unburned upland plots, 1.0% Silv-Ex plots remained distinct from other unburned plots with respect to the second CVA axis (Figure 4b); 0.5% Silv-Ex plots remained distinct from other unburned riparian plots at the 13th week after June applications (Figure 3a).
Burning itself had a much greater influence on community-level changes than did any of the chemical treatments (Figures 3 and 4). Pre-treatment plots varied little with respect to CV1 in either habitat type; burning tended to result in declines in the dominant grasses and sedges represented by this axis, especially in riparian habitat and after July application in upland habitat. Not surprisingly, burned plots remained distinct from unburned plots in uplands with respect to CV2 after June applications (Figure 4a); in this analysis, CV2 was positively correlated with Artemisia tridentata, which is harmed by fire (Young and Evans 1978).
Comparison of Habitat Types
Riparian habitat was marginally more sensitive to chemical treatments than upland habitats. Of the 10 vegetative characteristics we measured, only species richness showed a significant treatment effect in upland habitat (Figure 2c), and this effect was a subtle change in trend between burned and unburned treatments. Change in stems/m2 (Figure 1) and change in species richness (Figure 3) both showed significant treatment effects in riparian habitat. The reason for greater response on riparian plots may be related to moisture availability. Because moisture is limiting in the Great Basin (West 1988), the capacity for response is greater in the more mesic (i.e. riparian) compared to the more xeric (i.e. upland) sites.
Time of Application
In general, chemical effects that appeared after June applications also appeared after July applications in riparian habitat, although statistical significance tended to be greater in June applications. The only effects that persisted at the end of the study were the result of June applications (Figures 1a and 2c). Because the final sampling for both June and July applications were completed within a 2 week period, it is unlikely that this difference can be accounted for by a later final sample for July applications. More likely, some cool-season species that had yet to senesce when June treatments were applied had done so by the July applications, making the July experiment inherently less powerful.
Lack of Response
The majority of vegetative characteristics we measured showed no response to chemical application over the course of the growing season in which the chemicals were applied. We detected no treatment effect on species diversity or evenness, or on any characteristic of the two woody plants we examined (Tables 2, 3 and 4). Flowering progressed normally in Artemisia. Chemicals did not disrupt the well-known (Young and Evans 1974) post-fire sprouting of Chrysothamnus (Table 2). Activity of galling insects was not influenced by either chemical (Table 4), which suggests either that structural components and nutrients of leaves and stems were unaffected by Silv-Ex and Phos-Chek, or that they were affected in ways not detected by galling insects. The chemically treated communities were generally similar to the control communities (Figures 3 and 4).
Unlike a similar study in North Dakota in which P. pratensis increased its growth dramatically, apparently in response to fertilization by nitrogen in Phos-Chek (Larson and Newton 1995), no single species (including P. pratensis, which was common in riparian plots; Table 1) seemed to respond out of proportion to any other in this study (i.e. evenness was unaffected by chemical treatments). In addition, the only effects that persisted at the end of the study were the result of June applications (Figures 1a and 2c). We suspect these differences were related to lack of precipitation at the Nevada study sites. Precipitation was less than half the 30-year mean near our study sites in 1994 (2.03 cm for June–September 1994 compared with a yearly average of 4.90 cm for 1966-1995; NOAA National Climatic Data Center, U.S. Surface Data for Paradise Valley, NV). The lack of precipitation likely limited late-season growth of most species, and may have dampened fertilization effects. James and Jurinak (1979), working in Great Basin desert vegetation, found that response to experimental nitrogen fertilization was smaller when conditions were drier. Working in an annual grassland in California, Larson and Duncan (1982) also found that annual and seasonal weather patterns and soil moisture affected vegetation more than did treatment with air-dropped diammonium phosphate fire retardant.
The lack of significant differences among most chemical treatments applied after burning may reflect the short duration of the study rather than an actual lack of effect. Responses to burning in the sagebrush steppe are more appropriately measured over the course of several years (Young and Evans 1978), or even decades (Harniss and Murray 1973). Early-season annuals, in particular, could not be expected to grow on plots until the following season (West 1988). Many upland species, after early-spring growth that largely occurred before roads were passable to the study site, were dormant through most of the study. It should be kept in mind, however, that most natural fires also occur during this dormant season; if chemicals do not persist in the soils until the next growing season, there may, in fact, be little long-term effect of their use.