Predicting spillover risk to non-target plants pre-release: Bikasha collaris a potential biological control agent of Chinese tallowtree (Triadica sebifera)
Graphical abstract
Introduction
Insect herbivores influence plant performance, regulate plant populations, and shape communities (Maron and Crone, 2006). Classical biological control of weeds, the deliberate introduction of exotic agents to control exotic invasive weeds, seeks to capitalize on these attributes while assisting in the restoration of invaded habitats. Biological control of weeds can be a cost-effective, self-sustaining means of controlling invasive species (van Wilgen et al., 2013). While a few authors have viewed the implementation of biological control of weeds with scepticism and described it as risky (Louda et al., 2003, Simberloff and Stiling, 1996), others claimed it is an underutilized tool that should play a larger role in the control of invasive species in natural areas (Seastedt, 2015). However, biological control should only be implemented with proper assessments of risks and benefits (Hinz et al., 2014, Pemberton, 2000).
Concerns for the safety of biological control focus on two key risks, direct effects where a potential agent may cause significant harm to non-target species and indirect effects where broader ecological impacts may occur (Fowler et al., 2012). These direct effects include transitory damage to non-target species upon which the agent is unable to complete development. Predictions of risks from direct effects are generally determined through no-choice starvation testing which distinguishes hosts from non-hosts (McClay and Balciunas, 2005, Van Klinken, 2000). Although this research has an excellent track record (Balciunas and Smith, 2006), it may overestimate host range and can produce false-positives, excluding otherwise safe agents (Van Klinken, 2000). An approved agent may exhibit a small amount of feeding on a non-target species, and this may be considered acceptable as the agent may be unable to complete development and sustain a population on any species but the target weed. Such direct effects may occur as transient damage in the form of spillover onto non-target species that grow in association with the weed. Distinguishing these short term ephemeral effects from more sustainable non-target damage would be helpful while making decisions pre-release about the relative risk of potential agents. Predicting both direct and indirect risks a priori is a major goal and challenge of weed biological control.
Spillover in weed biological control is a direct effect where a non-target is used after an agent builds up high numbers leading to the collapse of the target weed population (Schooler et al., 2003). While herbivores may commonly restrict their host range to sub-family taxa (e.g., genus) (Forister et al., 2015, Jaenike, 1990, Novotny and Basset, 2005, Novotny et al., 2002), strictly monophagous species may be only rarely available as biological control agents (Sheppard et al., 2005). Consequently, unintended short term spillover onto non-target species may occur especially when oligophagous agents become over abundant and decimate the target weed (Holt and Hochberg, 2001, Lynch et al., 2002). When discovered, this spillover may be of great concern (Dhileepan et al., 2006, Diehl and McEvoy, 1990, Johnson and Stiling, 1998, Rand and Louda, 2004, Stiling et al., 2004). Population outbreaks have been reported following initial release of host specific agents, however these effects are transitory and have not led to long term population level non-target impacts (Catton et al., 2015, Hoddle, 2004a, Hoddle, 2004b, Suckling and Sforza, 2014, Taylor et al., 2007).
Chinese tallow (Triadica sebifera (L.); hereafter ‘tallow’) is one of the most damaging invasive weeds in the southeastern U.S.A., impacting wetlands, forests, and natural areas (Bruce et al., 1997). Classical biological control research of tallow began in 2006 (Wheeler and Ding, 2014) with overseas and quarantine host testing studies resulting in a petition to regulatory authorities requesting field release of a flea beetle, Bikasha collaris (Baly) (Coleoptera: Chrysomelidae) (Huang et al., 2011, Wheeler et al., 2017).
The larvae and adults of B. collaris feed on tallow roots and leaves, respectively. The results of host range studies indicated a high degree of specificity where B. collaris was unable to sustain a population on any non-target species. However, adult no-choice tests showed a limited amount of foliage feeding on two related species of Euphorbiaceae, Ditrysinia fruticosa (Bartram) Govaerts & Frodin and Gymnanthes lucida Sw. Oviposition by B. collaris only occurred on the target weed, tallow and on G. lucida where an average of 4.6 eggs (all non-viable) were produced (Wheeler et al., 2017). Another species of special concern was Hippomane mancinella L., a close relative of tallow and listed as endangered in Florida (Coile and Garland, 2003, Weaver and Anderson, 2010). Our host range research indicated little risk from direct effects by B. collaris on any of these non-targets. However, the risk of spillover damage onto non-targets after B. collaris had fed on tallow was unknown.
Results reported by Wheeler et al. (2017) indicated that, when B. collaris adults had fed on T. sebifera for two weeks, they were able to feed and had extended longevity on two non-target species, D. fruticosa and G. lucida. Possibly, this feeding and extended longevity could be explained by the use of well-fed adults that had previously developed on their primary host on which they acquired sufficient resources to continue to feed after being switched to non-target plants. Thus, we predicted that adult feeding and longevity would decrease if adults were fed on these non-targets without the benefit of prior exposure to T. sebifera.
To examine the risk of spillover pre-release, we compared B. collaris adult performance when they were naïve, or when they were fed tallow for 2 or 4 weeks and then switched to the non-targets in question. These responses were compared with adults fed continuously on tallow and also adults provided with water only. We determined the effects of these manipulations on adult feeding, longevity and oviposition to show pre-release to what extent B. collaris adults could spillover and cause significant and sustained damage on the non-target species.
Section snippets
Insects
In its native range, the flea beetle B. collaris has a temperate to subtropical distribution and was collected in Hubei, Ghizhou, Guangxi, and Hunan provinces ranging from latitudes 31.6°–24.8° North. Quarantine colonies of B. collaris were initiated from two shipments made in November 2008 and October 2009 from Wuhan Botanical Garden, Wuhan, Hubei, China. Upon arrival in the US, the B. collaris collections were housed in the quarantine facility at the Invasive Plant Research Laboratory,
Adult longevity
Mean longevity was significantly greater for naïve B. collaris adult females fed tallow continuously compared with those fed only water or the leaves of non-target species (F4,10 = 15.87; P = 0.0002) (Fig. 1). Naïve adult females (mean (±SEM)) fed tallow lived significantly longer (63.2 ± 10.9 days) compared with those fed only water (2.3 ± 0.3 days), D. fruticosa (5.5 ± 0.3 days), G. lucida (4.0 ± 0.3 days), or leaves of H. mancinella (8.3 ± 6.1 days) (Fig. 1). All naïve adults fed only water or non-target
Discussion
The results presented here confirm previous research which indicated that B. collaris will not sustain populations on the non-target species (Wheeler et al., 2017). Further, this study indicated that populations of B. collaris will not be sustained without continued access to the target weed. Our results showed that egg production is dependent upon sustained adult feeding on tallow leaves. Further, when fed tallow for 2 or 4 weeks and then switched to non-target leaves, they produced no more
Acknowledgments
We thank Lollis, J.A., Steininger, M. S., (USDA-ARS-IPRL) for technical assistance and J. Ding (Chinese Academy of Science) for field assistance. Additionally, voucher collections of these flea beetles are deposited in US National Museum, Washington DC; Entomology Department, and the Florida State Collection of Arthropods, Gainesville, FL. Our quarantine collections were identified morphologically by Dr. Alexander S. Konstantinov, Systematic Entomology Laboratory, National Museum of Natural
References (61)
- et al.
Prerelease efficacy assessment, in quarantine, of a tephritid gall fly being considered as a biological control agent for Cape-ivy (Delairea odorata)
Biol. Control
(2006) - et al.
The host range and impact of Bikasha collaris (Coleoptera: Chrysomelidae), a promising candidate agent for biological control of Chinese tallow, Triadica sebifera (Euphorbiaceae) in the United States
Biol. Control
(2011) - et al.
Ecological host-range of Lilioceris cheni (Coleoptera: Chrysomelidae), a biological control agent of Dioscorea bulbifera Biol
Control
(2015) - et al.
The role of pre-release efficacy assessment in selecting classical biological control agents for weeds – applying the Anna Karenina principle
Biol. Control
(2005) - et al.
Predicting risk and benefit a priori in biological control of invasive plant species: a systems modelling approach
Ecol. Model.
(2007) - et al.
Scientific advances in the analysis of direct risks of weed biological control agents to nontarget plants
Biol. Control
(2005) - et al.
Nontarget effects of a weed biological control agent on a native plant in Northern Australia
Biol. Control
(2007) - et al.
Biological control and indirect effects
Trend. Ecol. Evol.
(2004) - et al.
Amino-acids in nectar and their evolutionary significance
Nature
(1973) Physiologically induced changes in resource-oriented behavior
Annu. Rev. Entomol.
(1993)
Extrafloral nectaries and protection by pugnacious bodyguards
Annu. Rev. Ecol. Syst.
Host-Plant Selection by Phytophagous Insects
Nontarget feeding of leaf-beetles introduced to control purple loosestrife (Lythrum salicaria L.)
Nat. Area J.
Introduction, impact on native habitats, and management of a woody invader, the Chinese tallow tree, Sapium sebiferum (L.) Roxb
Nat. Area J.
Nontarget herbivory by a weed biocontrol insect is limited to spillover, reducing the chance of population-level impacts
Ecol. Appl.
Field efficacy and predicted host range of the pickerelweed borer, Bellura densa, a potential biological control agent of water hyacinth
Biocontrol
Should the mirid, Eccritotarsus catarinensis (Heteroptera: Miridae), be considered for release against water hyacinth in the United States of America?
Biocontrol Sci. Tech.
Notes on Florida's Endangered and Threatened Plant
Mechanism of host-plant recognition
Entomol. Exp. Appl.
Butterfly exploitation of an ant-plant mutualism: adding insult to herbivory
J. N. Y. Entomol. Soc.
Temporal patterns in incidence and abundance of Aconophora compressa (Hemiptera: Membracidae), a biological control agent for Lantana camara, on target and nontarget plants
Environ. Entomol.
A revision of Triadica Lour. (Euphorbiaceae)
Harvard Pap. Bot.
The influence of a shortage of hosts on the specificity of oviposition behaviour in species of Dacus (Diptera, Tephritidae)
Physiol. Entomol.
The global distribution of diet breadth in insect herbivores
Proc. Nat. Acad. Sci.
How can ecologists help practitioners minimize non-target effects in weed biocontrol?
J. Appl. Ecol.
Extrafloral nectar at the plant-insect interface: a spotlight on chemical ecology, phenotypic plasticity, and food webs
Annu. Rev. Entomol.
Successes we may not have had: a retrospective analysis of selected weed biological control agents in the United States
Invasive Plant Sci. Manag.
Restoring balance: using exotic species to control invasive exotic species
Conserv. Biol.
Cited by (8)
Predicting non-target impacts
2020, Current Opinion in Insect ScienceCitation Excerpt :The realized host range of herbivorous insects can increase at times of population outbreaks [9], which are expected in the period following the release of effective biocontrol agents. Laboratory experiments have been developed specifically to predict spill-over damage on plant species that could not support sustained agent populations [10,11]. Candidate agents, when confined in laboratory conditions, may complete development on plants they would not attack in the field, making interpretation of test results challenging and potentially resulting in the rejection of safe agents [12].