Oecologia (2009) 159:549–558 DOI 10.1007/s00442-008-1241-5 PLANT-ANIMAL INTERACTIONS - ORIGINAL PAPER Invasive plant architecture alters trophic interactions by changing predator abundance and behavior Dean E. Pearson Received: 26 June 2008 / Accepted: 14 November 2008 / Published online: 10 December 2008 Ó Springer-Verlag 2008 Abstract As primary producers, plants are known to primary native substrate. Dictyna webs on C. maculosa also influence higher trophic interactions by initiating food captured 4.2 times more large prey items, which are crucial chains. However, as architects, plants may bypass consum- for reproduction. As a result, Dictyna were nearly twice as ers to directly affect predators with important but likely to reproduce on C. maculosa substrates compared to underappreciated trophic ramifications. Invasion of western native substrates. The overall outcome of C. maculosa North American grasslands by the perennial forb, spot- invasion and its transformative effects on vegetation archi- ted knapweed (Centaurea maculosa), has fundamentally tecture on Dictyna density and web building behavior were altered the architecture of native grassland vegetation. Here, to increase Dictyna predation on invertebrate prey C89 fold. I use long-term monitoring, observational studies, and field These results indicate that invasive plants that change the experiments to document how changes in vegetation archi- architecture of native vegetation can substantially impact tecture have affected native web spider populations and native food webs via nontraditional plant ? preda- predation rates. Native spiders that use vegetation as web tor ? consumer linkages. substrates were collectively 38 times more abundant in C. maculosa-invaded grasslands than in uninvaded grass- Keywords Biological invasions Á Community assembly Á lands. This increase in spider abundance was accompanied Preadaptation Á Predator–prey interactions Á by a large shift in web spider community structure, driven Trait-mediated indirect interactions primarily by the strong response of Dictyna spiders to C. maculosa invasion. Dictyna densities were 46–74 times higher in C. maculosa-invaded than native grasslands, a Introduction pattern that persisted over 6 years of monitoring. C. mac- ulosa also altered Dictyna web building behavior and Food webs have been conceptualized as linear, density- foraging success. Dictyna webs on C. maculosa were 2.9– driven, bottom-up versus top-down chains of interactions. 4.0 times larger and generated 2.0–2.3 times higher total Either plants supply resources that fuel higher trophic prey captures than webs on Achillea millefolium, their structure from the bottom-up, or strong predator–consumer interactions cascade downward to structure communities from above (Hairston et al. 1960; Oksanen et al. 1981; Hunter and Price 1992). Although it has long been appre- Communicated by Sven Bacher. ciated that these simple interaction chains are caricatures of more complex trophic structure (Polis and Strong 1996), D. E. Pearson (&) Rocky Mountain Research Station, USDA Forest Service, the varied ways that indirect effects travel through food 800 E. Beckwith Ave., Missoula, MT 59801, USA webs has only recently come to light. For example, in the e-mail: [email protected] Aleutian Islands, exotic foxes indirectly transform sub- arctic grasslands to tundra, not by suppressing herbivores, D. E. Pearson Division of Biological Sciences, University of Montana, but by disrupting nutrient subsidies (Croll et al. 2005), and Missoula, MT 59812, USA in old field systems, spiders affect plant composition and 123 550 Oecologia (2009) 159:549–558 ecosystem properties more by altering the behavior than relatively uncommon native forbs. Thus, the invasion of the density of their prey (Schmitz 2006). these structurally depauperate grasslands by exotic forbs Just as predators can influence food webs via novel that are larger, more structurally complex and more per- pathways, so too can plants. Typically, plants are viewed as sistent than most native forbs may have important producing the energy that travels upwards through food ramifications for these native spiders and their prey. In this webs, thereby affecting higher trophic levels only as food paper, I use data from long-term monitoring, field obser- resources. However, plants are not only primary produc- vations, and field experiments to examine how a ers—they are architects. Plants create three-dimensional functionally unique exotic plant architecture alters native structures that define the physical landscape for predators web spider abundance and foraging behavior, and to and prey. Although it is well established that plants provide determine how these plant-induced changes in predator habitats that affect animal abundance and diversity (Crooks abundance and behavior may affect predation rates and 2002; Langellotto and Denno 2004), few studies have indirect interactions. examined how plant architecture alters predator–prey interactions (e.g., Denno et al. 2002). Furthermore, despite increasing recognition that predator-driven indirect effects Materials and methods can be mediated by changes in both prey behavior and density (Werner and Peacor 2003; Schmitz et al. 2004; Study system Pressier et al. 2005), the potential for plant architecture to influence indirect interactions through their direct effects This research was conducted in semi-arid low elevation on predator behavior and density remains unstudied. Yet grasslands in western Montana. These grasslands are such interactions may have broad ramifications for trophic dominated by the bunchgrasses Psuedoroegneria spicata dynamics, particularly following anthropogenic habitat and Festuca scabrella, and by forbs such as silky lupine changes and exotic plant invasions (e.g., Tylianakis et al. (Lupinus sericeus), arrowleaf balsamroot (Balsamorh- 2007). iza sagittata), and yarrow (Achillea millefolium). Native Biological invasions represent massive natural experi- forbs generally flower in May and June, senesce in July, ments offering potentially significant insights for and their flowering stalks fall and rapidly decompose understanding the roles of trophic interactions in commu- around September. Additionally, their flowering stems are nity assembly. Although extensive research shows that generally not very large or complex. Due to their mini- exotic plant invasions can radically alter native plant malist structure and low persistence, they offer poor communities (Mack et al. 2000; Levine et al. 2003) and substrates for web building spiders. An exception is exotic predators can severely impact native prey (Black- A. millefolium, which produces more rigid and persistent burn et al. 2004), little is known about how exotic plants flowering stems than most natives. Although A. millefolium affect predator–prey interactions and trophic outcomes stems are small and sparse, they often support spider webs (Levine et al. 2003). One underappreciated means by the following spring. Within these grasslands, Centau- which exotic plants may alter trophic interactions is by rea maculosa is a particularly potent invader that can reach changing vegetation architecture. In many western North high densities and usurp native plant communities (Ortega American grasslands, forbs offer the primary source of and Pearson 2005). These plants occur as rosettes early in three-dimensional structural diversity. However, these the growing season, but bolt and produce upright flowering plants are predominantly herbaceous and ephemeral, often stalks in May and June. C. maculosa flowering stalks are disappearing by the end of the growing season. Western larger, more complex, and more rigid than most native grasslands are currently being overrun by Eurasian forbs forbs and tend to persist to the next growing season when like spotted knapweed (Centaurea maculosa) (Sheley et al. they are readily used by native spiders. 1998), which produce flowering stems that are larger, more The dominant web building spiders in these grasslands structurally complex and rigid, and far more persistent than are D. major and D. coloradensis. These are small (cara- those of most natives. As a result, exotic plant invasions are pace length B1 mm), univoltine spiders that overwinter in transforming vegetation architecture over vast regions, plant litter during their 4th and 5th instars (Wheeler et al. with important implications for predator–prey interactions. 1990). Dictyna emerge from April to May and seek out In western Montana in the United States, Dictyna col- dried, upright flower stalks from the previous year’s growth oradensis and Dictnya major occur as small, uncommon as scaffolding on which to construct webs. Spiderlings native spiders that build persistent, irregular webs in the develop into adults, breed, and begin laying eggs by mid- upper one-third of the vegetation. Dictyna spp. appear to be to-late June. Young spiders begin to disperse in July. Both limited by their requirement for the more rigid, complex adult and young spiders travel by ballooning from plant to architecture offered by the flowering stalks of certain plant at the vegetation surface about 0.5 m above the 123 Oecologia (2009) 159:549–558 551 ground. Dictyna construct irregular webs that are built up along each transect and assigned webs either to Dictyna over time, but do not reconstruct webs daily like some orb spp or other web spiders. Other web spiders were com- weavers (Jackson 1978). These spiders retain prey items in prised primarily of the orb weaver