Although glyphosate is typically used as a nonselective herbicide, low rates have the potential to provide selective control of seedling annuals in the understory of established perennial plants. In a repeated experiment on two adjacent sites at a single location near Alturas, CA (2009 and 2010), we evaluated the efficacy of glyphosate at several different rates on medusahead and nontarget species in northern California sagebrush scrub. We applied glyphosate at 10 rates ranging from 0 to 709 g ae ha⁻¹ (0 to 18 oz product acre⁻¹) at three separate timings in each trial: mid-March (medusahead in early seedling stage), late April to early May (tillering), and late May to early June (boot to early head). Plots measured 3 m by 9 m (10 ft by 30 ft) and were arranged in randomized complete blocks with four replications for each rate and timing. We visually estimated vegetative cover for all dominant species in July before medusahead seed drop using three 1-m² quadrats per plot. Medusahead cover declined with increasing rates of glyphosate, and the middle application timing (at tillering) was the most effective. In rate series regression models, we achieved 95% control of medusahead with 160 g ae ha⁻¹ glyphosate in midseason 2009, compared with 463 g ae ha⁻¹ in early season and 203 g ae ha⁻¹ in late season. In 2010, we achieved 95% control with 348 g ae ha⁻¹ in midseason, compared with > 709 g ae ha⁻¹ in early season. Medusahead seed production reflected changes in cover, though individual plants tended to produce more seed at low densities. We attribute reduced control early in the season and poorer overall control in 2010 to greater tolerance of medusahead to glyphosate at lower temperatures. Treatment effects on big sagebrush, as indicated by shoot tip vigor, were minor, although the midseason timing caused a slight reduction in vigor. These results show that low rates of glyphosate (158 to 315 g ae ha⁻¹) at a treatment timing corresponding to medusahead tillering can give economical and effective control of medusahead without long-term damage to big sagebrush.

Nomenclature: Glyphosate; medusahead, Taeniatherum caput-medusae (L.) Nevski; big sagebrush, Artemisia tridentata Nutt.

Management Implications: In a Great Basin sagebrush community, low rates of glyphosate applied at the medusahead tillering stage in late April to early May provided the best control of medusahead. At this timing, we achieved at least 95% control of medusahead cover and a corresponding reduction in seed production with 160 g ae ha⁻¹ (4.1 oz product acre⁻¹) in 2009 and 348 g ae ha⁻¹ (8.8 oz product acre⁻¹) in 2010. Generic glyphosate products have a lower concentration of glyphosate (3 lb ae gal⁻¹) compared to Roundup ProMax used in this study (4.5 lb ae gal⁻¹) and, thus, would require 6 to 13 oz product acre⁻¹ to provide effective control of medusahead at the optimum timing. These rates are far lower than those required to control perennial species and provide a more cost effective option to ranchers and land managers compared to other herbicide treatments. Earlier applications required significantly higher rates to achieve equivalent control. Later applications (close to heading) also gave excellent control at slightly higher rates than applications at tillering. However, late application timing not only provided less competitive release and increased injury of desirable understory species, but also left a cover of standing dead medusahead, increasing the risk of wildfire. A potential negative aspect to the middle application timing is the overlapping timing with big sagebrush bud break. This level of injury did not cause plant death, but did suppress