Introduction
Non-native plant invasions have provided remarkable insights into rapid adaptation in defense allocation (Callaway & Maron, 2006; Waller et al., 2020). In part, this is because non-native plant species are typically released from specialist enemies in introduced ranges but continue to be attacked by generalists - the Enemy Release Hypothesis (Keane & Crawley, 2002). Thus, the most predictable change in herbivore communities in non-native ranges, at least until biological control agents are introduced, is the loss of specialists and disproportionate exposure to naive generalists. Generalist-dominated communities in non-native ranges commonly correspond with exotic species developing greater qualitative chemical defenses against generalists, relative to conspecifics in the native range - the Shifting Defense Hypothesis (Joshi & Vrieling, 2005; Zhang et al., 2018). These qualitative defenses may be constitutive or induced, but little is known about selection on constitutive versus induced defenses in non-native ranges. Theory and hypotheses that specifically focus on how constitutive and induced defenses might respond to generalists in native and non-native ranges have not led to clear predictions (Orians & Ward, 2010), which is barrier to understanding successful invasion.
Plants are thought to maximize their fitness by balancing resource allocation to defense with other functions (Rotter & Holeski, 2018; Mertens et al., 2021), and this allocation can be affected by herbivore pressure (Stamp, 2003). Constitutive defenses protect plants from attack without a time lag, but constitutive defenses are thought to be costly in the absence of herbivores (Wittstock & Gershenzon, 2002; Ali & Agrawal, 2012; Aljbory & Chen, 2018). Thus, when herbivore pressure is low, plant fitness might be maximized by investing less in constitutive defenses. In contrast, induced defenses appear to be cost-saving strategies where defenses are expressed only in response to herbivore pressure (Gatehouse, 2002; Agrawal & Hastings, 2019).
From the “quagmire” (Stamp, 2003) of hypotheses for defense allocation in plants, a few possibilities emerge for how induced and constitutive defenses might evolve in plant species in their non-native ranges. In a review of invasive plants, Orians & Ward (2010) suggested that induced defense should evolve to be more prevalent in non-native ranges where the cost of defense is high and plants are attacked infrequently (Bixenmann et al., 2016; Agrawal & Karban, 1999). Different non-native species often experience very different degrees of generalist herbivore pressure (Zou, 2008; Sims-Chilton et al., 2009), providing opportunities to test Orian & Ward’s idea. For example, a common garden survey found that different non-native species suffered from 0% to more than 40% leaf herbivory from generalists (Agrawal & Kotanen, 2003). Variation in herbivore pressure should lead to variation in defense traits (Coverdale & Agrawal, 2022; Wan et al., 2022) and less intense and less frequent herbivory is theorized to select for allocation to induced defenses (Bixenmann et al., 2016). However, how this variation in generalist herbivore pressure affects any potential tradeoffs in constitutivevs. induced defensive strategies among non-native plant species is unknown.
Growth-defense tradeoffs appear to derive from plasticity or adaptive variation in allocation that maintains fitness in a variable environment (Monson et al., 2022). Increased investment in defense may result in reduced growth (Hahn et al., 2021), leading to differential investment in growth or defense by species based on their evolutionary responses to environmental conditions - the Resource Availability Hypothesis (Coley et al., 1985). Most studies have focused on how variation in herbivore pressure across geographical clines shapes evolution of plant defense (Woods et al., 2012; Moreira et al., 2018). However, these studies have not generally considered how biogeographical shifts in defense strategies might constrain growth.
Non-native species provide a good system for studying induced and constitutive defenses. If non-native species are attacked less by generalists than native species, we can test the prediction that (i) non-natives increase expression of cheaper induced defenses and decrease expression of expensive constitutive defenses (Figure S1A and S1B). If non-native species are attacked disproportionately by generalists, then this disentangling of generalist and specialist effects might permit a cleaner opportunity to test another prediction (ii), that the herbivore pressure a species experiences will increase expression of constitutive defenses and decrease induced defenses (Figure S1A and S1B). We can then test the prediction (iii) that allocation to less expensive induced defense allow allocation to greater growth (Figure S1C).
To explore these predictions, we first carried out large spatial scale measurements of herbivore intensity on 27 non-native and 59 co-occurring native species in a field survey in northern China (Table S1, S2 and S3). We then quantified herbivore intensity on 12 pairs of non-native species and native congeners in a common garden. Finally, we conducted measurements of constitutive and induced defenses as well as plant growth on these pairs of common garden species (Table S4).