Our purpose was to investigate the influence of growth reserves of perennial range grasses on their tolerance to applications of atrazine [2-chloro-4-(ethylamino)-6-(isopropylamino)-s-triazine]. Competing annual grasses were removed with atrazine treatment at 1.1 kg/ha. The carbohydrate reserves of the perennial grasses were reduced by mowing the previous season. Crested wheatgrass [Agropyron desertorum (Fisch. ex Link) Schult.] plants with reduced carbohydrate reserves, as indicated by reduced etiolated growth, were less tolerant to atrazine than their nonmowed counterparts. Without reduced carbohydrate reserves, squirreltail [Sitanion hystrix (Nutt.) J. G. Smith] was much more susceptible to atrazine than crested wheatgrass. Thus, the weed control treatment did not necessarily increase the density of desirable perennial grasses.

Predictions of root architecture have been tested using only a narrow set of environmental conditions and species. There is little information on root architecture of plants growing on semi-arid grasslands where soil nutrient heterogeneity and defoliation might impose particular restrictions on root growth. We conducted a split-root study to evaluate the effect of nonuniform nutrient supply and clipping on root architecture and biomass allocation of three range grasses found in the Intermountain West of the USA: ‘Whitmar’, a cultivar of Pseudoroegneria spicata, Bromus tectorum (cheatgrass), and ‘Hycrest’, a hybrid between Agropyron desertorum and A. cristatum. Root systems of individual seedlings were equally divided between the two divisions of a partitioned 10-l pot, and each of the divisions was independently supplied with ‘high or ‘low’ nutrient solution. Plants were either clipped or not. Characteristics of root architecture (topology and link length), root size (root length and number of root tips), and biomass distribution (specific root length and root mass[ratio ]root volume ratio) were evaluated. Root topology (branching distribution) changed with nutrient treatments and defoliation in all three species. Length of links (a segment of root between two branches, or between a node and a root tip) responded to nutrient treatment only in defoliated plants of ‘Whitmar’, in which relative growth rate was associated with changes in architectural traits (external–external and internal–internal link length), while biomass accumulation and relative growth rate in cheatgrass and ‘Hycrest’ maintained allometric relationships with root length. Our results suggest that large plasticity in root architecture, together with inherently low relative growth rate are adaptations that allow ‘Whitmar’ to grow well in heterogeneous and poor soil-nutrient conditions. Cheatgrass and ‘Hycrest’ exhibited characteristics suggesting capability for rapid exploitation of nutrient patches.