Biol Invasions DOI 10.1007/s10530-017-1543-9 ORIGINAL PAPER Biogeography and phenology of oviposition preference and larval performance of Pieris virginiensis butterflies on native and invasive host plants Kate E. Augustine . Joel G. Kingsolver Received: 10 February 2017 / Accepted: 14 August 2017 Ó Springer International Publishing AG 2017 Abstract In invaded environments, formerly reli- driven by the early senescence of C. diphylla and able cues might no longer be associated with adaptive suggests a seasonal component to the impact of A. outcomes and organisms can become trapped by their petiolata. Therefore, the already short flight season of evolved responses. The invasion of Alliaria petiolata P. virginiensis could become further constrained in (garlic mustard) into the native habitat of Pieris invaded populations. virginiensis (West Virginia White) is one such exam- ple. Female butterflies oviposit on the invasive plant Keywords West Virginia White Á Alliaria petiolata Á because it is related to their preferred native host plant Evolutionary trap Á Novel plant–insect interactions Cardamine diphylla (toothwort), but larvae are unable to complete development. We have studied the impact of the A. petiolata invasion on P. virginiensis butter- flies in the Southeastern USA by comparing oviposi- Introduction tion preference and larval survival on both plants in North Carolina (NC) populations without A. petiolata Invading species can lead to novel ecological com- and West Virginia (WV) populations where A. peti- munities in which existing biotic interactions are olata is present. Larval survival to the 3rd instar was altered and new interactions are created (van der equally low in both populations when raised on A. Putten et al. 2004). Specifically, the introduction of petiolata. Mean oviposition preference on the two invasive plant species often changes environments by plants also did not differ between populations. How- altering biogeochemical cycles, decreasing commu- ever, we found a seasonal effect on preference nity diversity, and increasing competition with already between early and late season flights within WV established plants for nutrients, light, and pollinators populations. Late season females laid 99% of total (Gordon 1998). Invasive plant species can also impact eggs on A. petiolata while early season females native herbivores by altering plant–herbivore interac- utilized both host plants. Late season females were tions, including interactions with insect herbivores also less likely to lay eggs than early season females. (Pimentel et al. 2005). This change in preference toward A. petiolata could be Many plant–insect interactions have co-evolved as insect herbivores adapt to specific secondary defensive compounds of their native host plant species (Cornell & K. E. Augustine ( ) Á J. G. Kingsolver and Hawkins 2003). For example, plants are able to Department of Biology, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA perceive a wide range of herbivore-associated cues to e-mail: [email protected] elicit the release of defensive secondary compounds 123 K. E. Augustine, J. G. Kingsolver (Ali and Agrawal 2012). Many specialized insect Brassica species (Frisch et al. 2014). Alliarinoside has herbivores use these plant chemical cues to locate and been shown to function as a feeding deterrent to young identify preferred plant species for oviposition (Ren- larval instars of P. virginiensis (Davis et al. 2015). In wick 1989; Schoonhoven et al. 2005). Additionally, invaded habitats, A. petiolata has been documented as plant defensive compounds act as feeding stimulants an oviposition site for several Pierid species in North and deterrents during larval development and the America despite its unique glucosinolate composition balance of these stimulatory and inhibitory com- (Huang et al. 1995; Keeler and Chew 2008), including pounds controls larval acceptance or rejection of the P. virginiensis (Davis and Cipollini 2014a). Pieris host plant initially chosen by the ovipositing female rapae is native to Europe, where A. petiolata is (Schoonhoven et al. 2005). Therefore, plant invasions endemic, and is able to successfully use A. petiolata as have the potential to ‘‘trap’’ ovipositing females into a host plant (Huang et al. 1995). Evidence also shows laying their eggs onto host plants that are unsuitable to that a native cogener, Pieris oleracea, has been larvae depending on the chemical similarity of the successfully utilizing A. petiolata as a host as the invasive plant species to native host plants. Oviposi- plant has invaded and spread throughout New England tion mistakes such as these have the potential to (Huang et al. 1995; Keeler and Chew 2008). Keeler threaten and endanger the persistence of insect and Chew (2008) found that P. oleracea populations species. Here we explore how the ongoing range where A. petiolata is well established have improved expansion of an invasive plant, garlic mustard (Al- larval performance and increased adult female ovipo- liaria petiolata), is altering oviposition and larval sition preference for A. petiolata relative to popula- success of a native herbivore, the West Virginia White tions that do not co-occur with A. petiolata, suggesting butterfly (Pieris virginiensis), by comparing herbivore invaded populations are adapting to the invasive plant. populations inside and outside the current range of Like P. oleracea, P. virginiensis is native to eastern garlic mustard. North America and has also been shown to lay eggs on A. petiolata is an invasive biennial crucifer that was A. petiolata (Courant et al. 1994; Porter 1994; Davis introduced to North America from Europe in the late and Cipollini 2014a). It is a relatively rare, univoltine 1800s (Cavers et al. 1979; Rodgers et al. 2008). It is butterfly found where native crucifer species are highly successful as an invasive plant as it competes abundant in beech-maple-hemlock woods (Courant with native plants for nutrients and light and uses et al. 1994; Porter 1994). Most studies on this species allelopathy to reduce native seed germination have been conducted on populations throughout New (Vaughn and Berhow 1999; Meekins and McCarthy England, Pennsylvania, and Ohio, but the range of P. 1999; Prati and Bossdorf 2004), and negatively affects virginiensis extends throughout the Appalachians beneficial soil microbes which indirectly affects native southward into Virginia and North Carolina (Mather plant health (Roberts and Anderson 2001; Callaway 1964). The native ephemeral forb Cardamine diphylla et al. 2008; Wolfe et al. 2008; Burke 2008). As a result, is the most common larval host plant of P. virginiensis, A. petiolata is now common in many habitats across but there are also occasional small populations that use the eastern United States. It has become abundant as Cardamine concatenata, Cardamine dissecta,or far south as Virginia during the past 20 years, and it is Boechera laevigata when C. diphylla is absent (Cal- continuing to spread southward into North Carolina houn and Iftner 1988; Shuey and Peacock 1989). C. where it is now present in some habitats but not yet diphylla emerges in early April, completes leaf widespread or abundant (Rodgers et al. 2008). expansion by May, and senesces by early June when Most Pieris butterflies utilize hostplants in the the tree canopy begins to shade out the lower family Brassicaceae, and glucosinolates produced by understory. The life cycle of P. virginiensis is tightly these plants are important stimuli for female oviposi- coupled to the early spring phenology of C. diphylla. tion and larval feeding. A. petiolata produces similar After overwintering as pupae, the adult butterflies glucosinolates to native crucifer species, but the emerge in early to mid-April, flying 4–5 weeks specific glucosinolates produced are unique and only through May, and new larvae must complete devel- found in A. petiolata (Barto et al. 2010). Additionally, opment and pupate before senescence of the hostplant A. petiolata also produces the novel hydroxynitrile in June (Shapiro 1971; Cappuccino and Kareiva glucoside alliarinoside which is unknown from other 1985). Research on P. virginiensis in Connecticut by 123 Biogeography and phenology of oviposition preference and larval performance Doak et al. (2006) in the 1980 s suggests that low egg and avoidance of A. petiolata was quantified and loads and time limitation in this species contribute to compared by conducting ovipositional assays on both the ‘‘choosiness’’ of females’ oviposition sites. host plants between the North Carolina and West The northern distribution of P. virginiensis (New Virginia populations. Similarly, larval survival was York, Ohio, Pennsylvania) overlaps strongly with the quantified and compared by hatching and raising A. petiolata invasion. Adult female butterflies from larvae on both host plants and comparing between these populations readily oviposit on the invasive A. both populations. Additionally, because relative petiolata (Davis et al. 2015), and lab and field attractiveness between host plants is unlikely to be experiments show that oviposition is significantly static throughout the flight season due to differences in greater on invasive A. petiolata than on native C. phenology between the two species we assayed P. diphylla (Davis and Cipollini 2014a). However, larvae virginiensis oviposition preference in West Virginia are unable to complete development on these novel twice within one flight season—once during the first host plants (Bowden