Millions of hectares of range and wild lands in the western U.S. have been invaded by non-indigenous forbs, including spotted knapweed (Centaurea stoebe), resulting in negative impacts like decreases in native plant abundance (May and Baldwin 2011) and forage production (Watson and Renney 1974). Because of these negative impacts, invasive forbs are often listed as noxious weeds and routinely controlled through integrated management methods like herbicide applications and release of biological control agents. While these tools can be quite effective, they often do not result in a plant community that meets management objectives because remnant desired vegetation is too scarce to fill voids previously occupied by weeds. Subsequently, the weedy forb may be replaced by an equally invasive secondary invader like the annual grass cheatgrass (Bromus tectorum) (Story et al. 2006, Stephen et al. 2009).

Integrating re-seeding of desired species with weed control methods can help to alleviate secondary invasion and improve longterm restoration outcomes. For example, Rinella et al. (2012) found that the seeded grass intermediate wheatgrass (Thinopyrum intermedium) reduced spotted knapweed biomass by 93% 15 years after seeding. However, re-seeding does not guarantee restoration success. In the case of integrating herbicides and re-seeding, outcomes may be influenced by choice of herbicide and plant materials as well as environmental conditions and site characteristics. Choosing herbicides that effectively control the target invasive plant(s) can be a catalyst for maximizing success of seedlings of desired species.

The goal of our study was to restore desired, perennial grasses to rangeland impacted by spotted knapweed and cheatgrass through the integration of herbicides and reseeding. Two ecological assumptions guided our study: first, seeded species establishment would increase as control of the weedy species spotted knapweed and cheatgrass increased (Sheley et al. 2001, Davies 2010); second, increased establishment of seeded species would result in lower abundance of the weedy species spotted knapweed and cheatgrass over time (Whitson and Koch 1998, Rinella et al. 2012). We tested eight herbicide treatments and six seeding treatments at a site near Missoula, Montana (46°57’04” N, 114°07’07” W). Total yearly precipitation averages 34 cm and average annual high and low temperatures are 14.5°C and 1°C, respectively. Soils are gravelly loams of the Argixerolls-Haploxerolls complex. The site was dominated by spotted knapweed and cheatgrass, and Kentucky bluegrass (Poa pratensis), Canada bluegrass (P. compressa), and bulbous bluegrass (P. bulbosa) also occurred frequently.

Herbicide treatments were designed to provide control of spotted knapweed, cheatgrass, or both; we also included a non-sprayed treatment to help determine whether seeded grasses would establish without direct control of the target weeds. Herbicide treatments were as follows: 1) nonsprayed; 2) aminopyralid (Milestone^®) at 123 g ae/ha; 3) imazapic (Plateau®) at 105 g ai/ha; 4) aminopyralid at 88 g ae/ha plus imazapic at 105 g ai/ha; 5) aminopyralid at 123 g ae/ha plus imazapic at 70 g ai/ha; 6) glyphosate (Accord® XRT) at 471 g ai/ha; 7) aminopyralid at 53 g ae/ha plus clopyralid (Transline^®) at 242 g ae/ha plus imazapic at 105 g ai/ha; and 8) picloram (Tordon^® 22K) at 560 g ae/ha plus imazapic at 105 g ai/ha. Aminopyralid, clopyralid, and picloram are broadleaf herbicides that provide control of spotted knapweed; imazapic provides control of cheatgrass; and glyphosate is non-selective. Six seeding treatments were applied: 1) non-seeded to observe response of the plant community in the absence of seeding; 2) bluebunch wheatgrass (Pseudoroegneria spicata) at 27 kg/ha; 3) big bluegrass (Poa secunda) at 4.5 kg/ha; 4) tall wheatgrass (Thinopyrum ponticum) at 27 kg/ha; 5) bottlebrush squirreltail (Elymus elymoides) at 29 kg/ha; and 6) basin wildrye (Leymus cinereus) at 27 kg/ha. Herbicide treatments were applied on August 20, 2009 using a CO2-powered backpack sprayer delivering 142 L/ha water. [End Page 16] Seeding treatments were applied November 4, 2009 using a Haybuster 77^® all-purpose drill seeder. Treatments were applied in a strip-plot design (herbicide and seeding treatments crossed each other in a perpendicular fashion) and replicated three...