Introduction
The emerald ash borer (EAB), Agrilus planipennis Fairmaire (Coleoptera: Buprestidae), according to dendrochronological evidence, was accidentally introduced into North America sometime in the mid-1990s (Siegert et al. Reference Siegert, McCullough, Liebhold and Telewski2014). The species is native to eastern Asia (China, Japan, South Korea, Mongolia, Taiwan, and the Russian Far East) where it is an uncommon species (Haack et al. Reference Haack, Jendek, Liu, Marchant, Petrice and Poland2002; Lui et al. Reference Liu, Bauer, Gao, Zhao, Petrice and Haack2003), but has become an occasional pest of introduced North American ash in China and the Russian Far East (Wei et al. Reference Wei, Wu, Reardon, Sun, Lu and Sun2007; Baranchikov et al. Reference Baranchikov, Mozolevskaya, Yurchenko and Kenis2008). This invasive beetle went undetected until it was discovered in the environs of Detroit, Michigan, United States of America and Windsor, Ontario, Canada in 2002 (Haack et al. Reference Haack, Jendek, Liu, Marchant, Petrice and Poland2002). Consequently, for approximately a decade the species was able to disperse unchecked from its putative invasion locale near Detroit. The insect is a devastating pest of ash, Fraxinus Linnaeus (Oleaceae) species, and has killed millions of these host trees since its invasion (Herms and McCullough Reference Herms and McCullough2014). All 20 species (Wallander Reference Wallander2008) of North American ashes are probably at risk. The realised and potential ecological (Gandhi and Herms Reference Gandhi and Herms2010a, 2010Reference Gandhi and Hermsb; Hausman et al. Reference Hausman, Jaeger and Rocha2010; Fissore et al. Reference Fissore, McFadden, Nelson, Peters, Hobbie and King2012; Flower et al. Reference Flower, Knight and Gonzalez-Meler2013) and economic (Kovacs et al. Reference Kovacs, Haight, McCullough, Mercader, Siegert and Liebhold2010, Reference Kovacs, Mercader, Haight, Siegert, McCullough and Liebhold2011, Reference Kovacs, Haight, Mercader and McCullough2014; McKenney et al. Reference McKenney, Pedlar, Yemshanov, Lyons, Campbell and Lawrence2012) impacts of this invasive beetle are staggering.
The eggs of A. planipennis are laid in bark crevices and under bark scales on the host tree (Wang et al. Reference Wang, Yang, Gould, Zhang, Liu and Liu2010). The neonate larvae chew their way out of the sides of the disc-like eggs that are in contact with the bark substrate. The larvae burrow through the bark to the cambial layer and feed on the inner phloem, cambium, and outer xylem where they form serpentine-shaped galleries (Cappaert et al. Reference Cappaert, McCullough, Poland and Siegert2005). The larvae pass through four stadia in the galleries and the mature fourth-instar excavate pupal chambers in the outer xylem tissue or in the inner phloem tissue (Cappaert et al. Reference Cappaert, McCullough, Poland and Siegert2005). Within the pupal chambers the mature larvae fold over on themselves and become “J” shaped. In the region of invasion, the majority of the population of A. planipennis is univoltine and overwinters in the “J”-shaped larval stage in the pupal chambers. The remainder of the population (20–25%) overwinter as earlier feeding-stage larvae in the feeding galleries (Cappaert et al. Reference Cappaert, McCullough, Poland and Siegert2005). These latter individuals will overwinter a second time resulting in a two-year life cycle. Voltinism in A. planipennis seems to be regulated by tree resistance and climate (Duan et al. Reference Duan, Ulyshen, Bauer, Gould and Van Driesche2010). As temperatures rise in spring, the prepupae lose their “J” shape and constrict in length in preparation for pupation (Duan et al. Reference Duan, Ulyshen, Bauer, Gould and Van Driesche2010). Following eclosion within these chambers, adults chew their way out of the tree through the bark via a characteristic “D”-shaped hole (Haack et al. Reference Haack, Jendek, Liu, Marchant, Petrice and Poland2002). The adults undergo a maturation period during which they feed on the foliage of the host trees. Following this maturation period, they mate and females begin to oviposit (Cappaert et al. Reference Cappaert, McCullough, Poland and Siegert2005).
If the beetle had arrived on this continent with a complement of its Asian natural enemies, then populations of these organisms would have had time to build up in the beetle’s population. There was also ample opportunity between the introduction of the beetle and its discovery for natural enemies that attack native species of Buprestidae to make the host shift to A. planipennis if they were capable of the transition. Preliminary surveys conducted in Michigan shortly after its discovery indicated that parasitism of A. planipennis by parasitoids amounted to <1%, while parasitisation by fungal pathogens was ~3% (Liu et al. Reference Liu, Bauer, Gao, Zhao, Petrice and Haack2003; Bauer et al. Reference Bauer, Liu, Haack, Petrice and Miller2004). No parasitoids that had been previously associated with A. planipennis in Asia were encountered during these surveys. Several subsequent regional surveys of parasitoids and pathogens have been undertaken since that time and will be discussed below.
Herein, mortality factors that influence the population dynamics of A. planipennis are reviewed. Investigations to date, chronicling native natural enemies attacking A. planipennis will be discussed. Evidence for other mortality factors (overwintering mortality, host plant resistance, competition, and cannibalism) in the population dynamics of A. planipennis will be detailed.
North American natural enemies
Parasitoids
Surveys were conducted from 2002 to 2004 in Michigan to determine if North American natural enemies were having any impact on introduced populations of A. planipennis, as a prerequisite to undertaking classical biological control of A. planipennis using exotic parasitoids from China (Lui et al. Reference Liu, Bauer, Gao, Zhao, Petrice and Haack2003; Bauer et al. Reference Bauer, Liu, Haack, Petrice and Miller2004). Several species of Hymenoptera parasitoids were reared from larvae of A. planipennis during these surveys including: Atanycolus hicoriae Shenefelt (Hymenoptera: Braconidae), Atanycolus simplex (Cresson) (Hymenoptera: Braconidae), Spathius floridanus Ashmead (Hymenoptera: Braconidae) (=S. simillimus Ashmead; synonymised by Marsh and Strazanac Reference Marsh and Strazanac2009), Phasgonophora sulcata Westwood (Hymenoptera: Chalcididae), and Balcha Walker (Hymenoptera: Eupelmidae) species. Less than 1% of the A. planipennis larvae sampled were parasitised. In spite of having examined large numbers of eggs of A. planipennis, no egg parasitoids have been reported in A. planipennis populations in Michigan or Pennsylvania, United States of America (Liu et al. Reference Liu, Bauer, Gao, Zhao, Petrice and Haack2003; Bauer et al. Reference Bauer, Liu, Gould and Reardon2007; Duan et al. Reference Duan, Fuester, Wildonger, Taylor, Barth and Spichiger2009).
Natural enemies of Agrilus species related to the pest or parasitoids attacking populations of A. planipennis in North America have been sought for an augmentative/inundative biocontrol strategy (Cappaert and McCullough Reference Cappaert and McCullough2009; Duan et al. Reference Duan, Fuester, Wildonger, Taylor, Barth and Spichiger2009; Lyons Reference Lyons2010; Kula et al. Reference Kula, Knight, Rebbeck, Bauer, Cappaert and Gandhi2010). Five species of Hymenoptera parasitoids were reared from A. planipennis life stages collected in Pennsylvania in a preliminary survey with a total parasitism rate of 3.6% (Duan et al. Reference Duan, Fuester, Wildonger, Taylor, Barth and Spichiger2009). A subsequent survey in Pennsylvania produced six species of parasitoids causing a combined total of 5.6% parasitism (Duan et al. Reference Duan, Taylor, Fuester, Kula and Marsh2013b). These natural enemies could potentially be mass reared and introduced into A. planipennis populations.
Extremely low levels of A. hicoriae Shenefelt and A. simplex (Cresson) (Hymenoptera: Braconidae), that were associated with A. planipennis larvae were collected during preliminary surveys for native parasitoids in Michigan (Liu et al. Reference Liu, Bauer, Gao, Zhao, Petrice and Haack2003). A total of five species of ectoparasitoids in the genus Atanycolus have been reported from A. planipennis including: A. nigropyga Shenefelt, A. hicoriae Shenefelt, A. simplex Cresson, A. tranquebaricae Shenefelt (Cappaert and McCullough Reference Cappaert and McCullough2009) and a new species, A. cappaerti Marsh and Strazanac, which was described by Marsh et al. (Reference Marsh, Strazanac and Laurusonis2009). Atanycolus nigropyga caused 0.26–1.03% parasitism of A. planipennis in Pennsylvania (Duan et al. Reference Duan, Taylor, Fuester, Kula and Marsh2013b). In study sites near Lansing, Michigan, parasitism of A. planipennis by Atanycolus species increased from nine (<0.3% parasitism of A. planipennis larvae) to 774 (19.4% parasitism of A. planipennis larvae) individuals in one year (Duan et al. Reference Duan, Bauer, Abell and Van Driesche2012b). About 93% of these parasitoids were A. cappaerti and the remainder were A. hicoriae (5%), A. tranquebaricae (1%), and A. disputabilis (<1%). Atanycolus cappaerti, A. hicoriae, and A. longicauda Shenefelt (identified by P. Marsh, United States National Museum, Washington, District of Columbia, United States of America; retired), were reared from logs infested by A. planipennis from southwestern Ontario, Canada (Lyons Reference Lyons2010). Two of these species, A. cappaerti and A. hicoriae, plus an additional species, A. disputabilis (Cresson), were also collected from white birch logs, infested with Agrilus anxius Gory, from in the vicinity of Sudbury, Ontario. Cappaert and McCullough (Reference Cappaert and McCullough2009) investigated the natural history and biocontrol potential of A cappaerti. Observed parasitism rates ranged from 9% to 71% in two sites over two years suggesting that this species might be an effective biological control agent for A. planipennis. This idiobiont ectoparasitoid has a bivoltine life cycle with a generation time of <30 days. The sex ratio (male:female) of the parasitoids was 1:6 with similar median adult emergence dates for the sexes in late May for overwintering individuals. Atanycolus cappaerti larvae were found on second-instar to fourth-instar hosts. Duan et al. (Reference Duan, Bauer, Abell and Van Driesche2012b) reported a very different sex ratio for A. cappaerti of 1:0.6 (male:female). Males and females of Atanycolus species reared from log bolts collected in Ontario were extremely long lived with both sexes living for an average of more than 100 days at 21 °C (Lyons Reference Lyons2010). Tetrastichus planipennisi Yang (Hymenoptera: Eulophidae) is one of two larval parasitoids introduced into North America from China as a classical biological control agent against A. planipennis. In recent investigations, it was determined that Atanycolus species are able to attack larvae of A. planipennis under much thicker bark than can T. planipennisi (Abell et al. Reference Abell, Duan, Bauer, Lelito and Van Driesche2012). Atanycolus species can parasitise A. plannipennis larvae in bark up to 8.8 mm thick.
The most abundant parasitoid in Delaware County, Ohio, United States of America, which was positively associated with A. planipennis cadavers, was Leluthia astigma (Ashmead) (Hymenoptera: Braconidae) with a parasitism rate of 2.1% (Kula et al. Reference Kula, Knight, Rebbeck, Bauer, Cappaert and Gandhi2010). This species is broadly distributed in North America from California, United States of America and Guadalajara, Mexico to North Carolina, United States of America and Québec, Canada (Kula et al. Reference Kula, Knight, Rebbeck, Bauer, Cappaert and Gandhi2010). One specimen was collected from green ash (Fraxinus pennsylvanica Marshall) infested with A. planipennis near Windsor, Ontario and another specimen was collected from red oak (Quercus rubra Linnaeus (Fagaceae)) infested by Agrilus bilineatus (Weber) near Midland, Ontario (Lyons Reference Lyons2010). Although reported from Québec (Marsh Reference Marsh1979) these are the first specimens collected from elsewhere in Canada. Adults of this solitary idiobiont ectoparasitoid, which emerged from cocoons, successfully produced a new generation on A. planipennis larvae inserted into ash sticks (Kula et al. Reference Kula, Knight, Rebbeck, Bauer, Cappaert and Gandhi2010).
Spathius floridanus was encountered by Bauer et al. (Reference Bauer, Liu, Haack, Petrice and Miller2004) and Duan et al. (Reference Duan, Bauer, Abell and Van Driesche2012b) during surveys for native parasitoids of A. planipennis in Michigan. The known distribution of S. floridanus extends from New Brunswick to Ontario in Canada and from Wisconsin to Texas, United States of America (Marsh and Strazanac Reference Marsh and Strazanac2009). Known hosts for S. floridanus include A. anxius, A. bilineatus (Weber), Chrysobothris femorata (Olivier) (Coleoptera: Buprestidae), Magdalis olyra (Herbst) (Coleoptera: Curculionidae), Phymates aereum (Newman), and Xylotrechus colonus (Fabricius) (Coleoptera: Cerambycidae) (Marsh and Strazanac Reference Marsh and Strazanac2009). This species was considered the most promising native Spathius as a candidate for biological control of A. planipennis (Marsh and Strazanac Reference Marsh and Strazanac2009). Fifteen specimens of this parasitoid were reared from red oak infested with A. bilineatus near Midland, Ontario, but the species has not been reared from A. planipennis in Ontario (Lyons Reference Lyons2010). This record may represent a small number of oviposition events because Spathius species are gregarious (Marsh and Strazanac Reference Marsh and Strazanac2009). Spathius laflammei Provancher caused a maximum of 2.6% parasitism of A. planipennis in western Pennsylvania and readily attacked third and fourth instars of A. planipennis in the laboratory (Duan et al. Reference Duan, Taylor, Fuester, Kula and Marsh2013b). An undescribed species of Spathius has also recently been recorded from A. planipennis in Michigan (Marsh and Strazanac Reference Marsh and Strazanac2009). A congeneric species, Spathius agrili Yang is a larval parasitoid from China that has been introduced into North America for biological control of A. planipennis (Duan et al. Reference Duan, Ulyshen, Bauer, Gould and Van Driesche2010). Kuhn et al. (Reference Kuhn, Duan and Hopper2013) have developed molecular tools for differentiating species of Spathius.
Balcha indica (Mani and Kaul) (Hymenoptera: Eupelmidae) was one of the parasitoids encountered during surveys for native parasitoids of A. planipennis in Michigan (Gibson Reference Gibson2005; Duan et al. Reference Duan, Bauer, Abell and Van Driesche2012b), Pennsylvania (Duan et al. Reference Duan, Fuester, Wildonger, Taylor, Barth and Spichiger2009, Reference Duan, Taylor, Fuester, Kula and Marsh2013b), and Ontario (Lyons Reference Lyons2010). Balcha indica is itself an alien species, which is native to southeastern Asia and has been collected from Burma, India, Thailand, and Vietnam. It probably arrived in North America on some host other than A. planipennis because its discovery in Virginia, United States of America in 1994 pre-dates the estimated arrival of A. planipennis in that area (Gibson Reference Gibson2005). Males of B. indica are known to occur in Asia but only females are known from North America (Gibson Reference Gibson2005). This parasitoid was the most abundant species reared from A. planipennis in Pennsylvania (Duan et al. Reference Duan, Fuester, Wildonger, Taylor, Barth and Spichiger2009). It successfully reproduced parthenogenetically on larvae of A. planipennis in the laboratory (Duan et al. Reference Duan, Fuester, Wildonger, Taylor, Barth and Spichiger2009, Reference Duan, Taylor, Fuester, Kula and Marsh2013b). The life history and stages of this solitary ectoparasitoid was described by Duan et al. (Reference Duan, Taylor and Fuester2011b). Females begin to oviposit during the first week after emergence with peak emergence occurring several weeks later. The wasps paralysed and parasitised larvae (second to fourth instar), prepupae, and pupae of the host. Eggs are attached to the host by sticky strands. Mean fecundity was 35.8 eggs. First instars have sclerotised head capsules and mandibles while later instars do not. A generation takes about three months to complete at 25 °C. The long generation time for the univoltine B. indica may hamper effective rearing methods and thus limit its use as an augmentative biological control agent against A. planipennis (Duan et al. Reference Duan, Taylor and Fuester2011b). Small numbers of B. indica have routinely been collected from logs infested by A. planipennis in Ontario (Lyons Reference Lyons2010).
A single specimen of Metapelma spectabile Westwood (Hymenoptera: Eupelmidae) was reared from an ash bolt infested by A. planipennis collected near Windsor, Ontario (Lyons Reference Lyons2010). This species has been reported from Agrilus angelicus Horn but there is no previous record from Canada (Burks Reference Burks1979).
An unprecedented number of P. sulcata Westwood (Hymenoptera: Chalcididae) emerged from green ash log bolts collected from a site in southwestern Ontario that was in the later stages of an A. planipennis outbreak (Lyons Reference Lyons2010). Of the 215 insects reared from these bolts 146 were A. planipennis and 54 were P. sulcata. Eight P. sulcata that emerged from bolts from a nearby location were greatly outnumbered by 648 emerging adults of A. planipennis, suggesting a parasitism rate of only about 1.2%. Parasitism rates by P. sulcata apparently vary spatially and probably temporally. Sticky-band traps were placed in the woodlot in Essex County, Ontario in 2007 from which the high number of P. sulcata had been obtained. Six hundred adults of A. planipennis and 407 adults of P. sulcata were collected from the traps suggesting a parasitism rate of 40.7% (Lyons Reference Lyons2010). Mean emergence of P. sulcata was after the mean emergence of adults of A. planipennis in the laboratory. This corresponds with the observed late flight period of the parasitoid relative to the host observed in the field in Ontario (Lyons Reference Lyons2010). The flight period of the parasitoid seems to be synchronised with the egg laying period of A. planipennis and supports the observation by Haack et al. (Reference Haack, Benjamin and Schuh1981) that the parasitoid may lay eggs near the eggs of the host. Mean longevities for P. sulcata male and female adults were ~25 and 39 days, respectively (Lyons Reference Lyons2010). This parasitoid species was also reared from A. planipennis in Michigan (Liu et al. Reference Liu, Bauer, Gao, Zhao, Petrice and Haack2003) and has been reported from the two-lined chestnut borer, A. bilineatus (Haack et al. Reference Haack, Benjamin and Schuh1981), A. anxius (Akers and Nielsen Reference Akers and Nielsen1990), and the bronze poplar borer, A. granulatus liragus Barter and Brown (Barter Reference Barter1965).
A few other species of Hymenoptera parasitoids have been reared from A. planipennis including: Eupelmus pini Taylor (Eupelmidae); three ichneumonids, Dolichomitus vitticrus Townes, Orthizema Förster species, and Cubocephalus Ratzeburg species (Duan et al. Reference Duan, Fuester, Wildonger, Taylor, Barth and Spichiger2009); and unknown species of Dolichomitus Smith (Ichneumonidae) and Eurytoma Illiger (Eurytomidae) (Duan et al. Reference Duan, Bauer, Abell and Van Driesche2012b). Duan et al. (Reference Duan, Fuester, Wildonger, Taylor, Barth and Spichiger2009) successfully reared the thelytokous parthenogenetic E. pini on A. planipennis larvae, prepupae, and pupae inserted into ash sticks in the laboratory.
Predators
A few species of North American invertebrate predators have made the successful transition to preying on populations of A. planipennis. The predatory beetles (Coleoptera) Enoclerus Gahan (Cleridae) species, Catogenus rufus (Fabricius) (Passandridae), and Tenebroides Piller and Mitterpacher (Trogossitae) species were reported feeding on life stages of A. planipennis under the bark of host trees during natural enemy surveys conducted in Michigan in 2002–2004 (Liu et al. Reference Liu, Bauer, Gao, Zhao, Petrice and Haack2003; Bauer et al. Reference Bauer, Liu, Haack, Petrice and Miller2004).
The predatory wasp, Cerceris fumipennis Say (Hymenoptera: Crabronidae), collects adults of A. planipennis to provision its nest. The wasp is unable to regulate A. planipennis populations because it is not specific in its prey choice, capturing many species of buprestid adults and occasionally adults of other beetle families (Careless et al. Reference Careless, Marshall and Gill2013). This species has proven to be a useful biosurveillance tool for detecting low-density populations of A. planipennis even though the impact of the wasp on A. planipennis populations is not significant (Marshall et al. Reference Marshall, Paiero and Buck2005; Careless et al. Reference Careless, Marshall, Gill, Appleton, Favrin and Kimoto2009, Reference Careless, Marshall and Gill2013).
A large percentage (37–52%) of sentinel eggs of A. planipennis, used to monitor the exotic parasitoid Oobius agrili Zhang and Huang (Hymenoptera: Encyrtidae) released as a biological control agent against A. planipennis, disappeared possibly resulting from ant predation (Duan et al. Reference Duan, Bauer, Ulyshen, Gould and Van Driesche2011a). Ants might be a factor in the establishment of O. agrili because parasitised eggs are available for a prolonged period in the field. Predation of naturally occurring A. planipennis eggs in the field has never been quantified.
Woodpecker (Aves: Picidae) predation is probably the most important source of mortality in A. planipennis populations in Michigan and has accounted for 9–95% mortality (Cappaert et al. Reference Cappaert, McCullough, Poland and Siegert2005). Woodpecker predation (up to 41.1% mortality) was the second highest source of mortality, after host tree defences, for experimental populations of A. planipennis (Duan et al. Reference Duan, Ulyshen, Bauer, Gould and Van Driesche2010), while woodpecker predation reached 32–42% in a later study in natural populations (Duan et al. Reference Duan, Bauer, Abell and Van Driesche2012b). Results for Ohio forests were comparable with a woodpecker predation rate of ~37% (Flower et al. Reference Flower, Long, Knight, Rebbeck, Brown and Gonzalez-Meler2014). Three common species of woodpeckers, namely: hairy (Picoides villosus (Linnaeus)), downy (P. pubescens (Linnaeus)), and red-bellied (Melanerpes carolinus (Linnaeus)) woodpeckers, have been observed preying on A. planipennis (Lindell et al., Reference Lindell, McCullough, Cappaert, Apostolou and Roth2008). The maintenance of conditions (e.g., nest sites) that attract woodpeckers was recommended by those authors as a means of reducing A. planipennis numbers. Woodpecker predation was positively correlated with A. planipennis density. There were more woodpecker attacks on white ash than on green ash. Woodpecker predation was also higher in less forested sites. Woodpecker predation occurred almost exclusively on prepupae (including “J”-shaped larvae) and pupae (98.8%) (Jennings et al. Reference Jennings, Gould, Vandenberg, Duan and Shrewsbury2013; Duan et al. Reference Duan, Abell, Bauer, Gould and Van Driesche2014). Almost all woodpecker attacks resulted in the death of prepupae and pupae of A. planipennis. Woodpecker predation has been reported as high as 26% (average 4%) in the native range of A. planipennis in China (Wang et al. Reference Wang, Yang, Gould, Zhang, Liu and Liu2010). However, with the exception of one observation of 42% predation of A. planipennis by woodpeckers at an arboretum in the Russian Far East, woodpecker predation is insignificant in that region (Duan et al. Reference Duan, Yurchenko and Fuester2012a). Bird-feeding activities that attract woodpeckers into an area infested by A. planipennis may have some impact on the borer’s populations. Providing suet in summertime when the birds are feeding their offspring may assist in increasing woodpecker populations (Poland and McCullough Reference Poland and McCullough2010). Woodpecker activity or signs of feeding can be used as a means of detecting and delineating high-density populations of A. planipennis. Strategies for attracting woodpeckers into an area would be characterised as augmentative biological control while strategies for maintaining them in an area would be characterised as conservation biological control. Woodpeckers prefer to forage on trees in poorer condition, which have higher densities of A. planipennis, resulting in attacks on trees that are beyond saving (Jennings et al. Reference Jennings, Gould, Vandenberg, Duan and Shrewsbury2013). Moreover, dead and dying trees that provide nesting sites for woodpeckers only occur abundantly in old A. planipennis outbreaks. These two factors result in delayed population buildup of these predators (Jennings et al. Reference Jennings, Gould, Vandenberg, Duan and Shrewsbury2013).
Dodds and Otis (Reference Dodds and Otis2009) collected baseline data on populations of birds in Point Pelee National Park, Ontario, in advance of the A. planipennis population explosion. They predicted that species of birds that required large tracts of forested land might be eliminated, while species that required patchy forest might benefit from ash tree loss. Woodpeckers might benefit temporarily from the abundance of A. planipennis prey, while herbivorous birds that feed on ash seeds may decline.
Koenig et al. (Reference Koenig, Liebhold, Bonter, Hochachka and Dickinson2013) examined the effects of infestations by A. planipennis on the three woodpeckers listed above and the white-breasted nuthatch, Sitta carolinensis Latham (Aves: Sittidae), which gleans insects from bark; using regional data from the Project FeederWatch, a volunteer science programme. In areas heavily affected by A. planipennis, populations of all four avian predators increased (albeit delayed for two of the species), probably as a result of increased survival and reproduction. All four bird species are cavity nesters and may benefit from increased nest site availability from dead and dying trees. In stands infested by A. planipennis, the three dominant woodpeckers shift their foraging activities to ash and concentrate their efforts on trees with declining canopies (Flower et al. Reference Flower, Long, Knight, Rebbeck, Brown and Gonzalez-Meler2014).
Pathogens
Native fungal pathogens resulted in <3% mortality in Michigan populations of A. planipennis (Liu et al. Reference Liu, Bauer, Gao, Zhao, Petrice and Haack2003). The most prevalent fungus collected was Beauveria bassiana (Balsamo) Vuillemin (Hypocreales: Clavicipitaceae) (Bauer et al. Reference Bauer, Liu, Haack, Petrice and Miller2004). The other four species recovered included Isaria farinosus (Holmskjold) Fries, I. fumosoroseus Wize (Eurotiales: Trichocomaceae), Lecanicillium lecanii (Zimmerman) Viegas, and Metarhizium anisopliae (Metchnikoff) Sorokin (Hypocreales: Clavicipitaceae). Similarly, fungal pathogens accounted for <3% mortality of wild and experimental cohorts of A. planipennis in a subsequent experimental site in Michigan (Duan et al. Reference Duan, Ulyshen, Bauer, Gould and Van Driesche2010). Molecular characterisation of the strains of B. bassiana collected by Bauer et al. (Reference Bauer, Liu, Haack, Petrice and Miller2004) suggested they were indigenous strains that were found in bark and soil samples in the same sites (Castrillo et al. Reference Castrillo, Bauer, Liu, Griggs and Vandenberg2010).
In Ontario, collections of mycosed A. planipennis prepupal and adult cadavers, and frass containing fungi yielded 78 isolates of B. bassiana and B. pseudobassiana, as the predominant “natural” pathogens (Kyei-Poku and Johny Reference Kyei-Poku and Johny2013). Lecanicillium Gams and Zare (Cordycipitaceae) species, Metarhizium Sorokīn (Clavicipitaceae) species, and Paecilomyces Bainier (Trichocomaceae) species were also collected but very low levels (Lyons Reference Lyons2010). The isolates of B. bassiana were characterised using molecular techniques and their virulence and conidia production were assessed (Johny et al. Reference Johny, Kyei-Poku, Gauthier, van Frankenhuyzen and Krell2012b). Eight of the 23 isolates screened resulted in more than 90% mortality of EAB in a single dose bioassay. In laboratory bioassays, the L49-1AA isolate of B. bassiana was the most virulent. It was even more virulent than the commercial GHA isolate and killed A. planipennis adults much more quickly (Johny et al. Reference Johny, Kyei-Poku, Gauthier, van Frankenhuyzen and Krell2012b). The L49-1AA isolate can be separated from other isolates in the field by a molecular diagnostic tool developed by Kyei-Poku and Johny (Reference Kyei-Poku and Johny2013).
Two native isolates of B. bassiana, INRS-CFL, characterised by Sabbahi et al. (Reference Sabbahi, Lavallee, Merzouki and Guertin2009) and L49-1AA, characterised by Johny et al. (Reference Johny, Kyei-Poku, Gauthier, van Frankenhuyzen and Krell2012b), were experimentally released into A. planipennis populations in southwestern Ontario in 2011 using auto-dissemination techniques (Lyons et al. Reference Lyons, Lavallee, Kyei-Poku, van Frankenhuyzen, Johny and Guertin2012). Proof of concept was demonstrated by the development of an auto-contamination trap that successfully contaminated EAB adults in the field and provided viable fungal conidia throughout the flight period of the beetle (Lyons et al. Reference Lyons, Lavallee, Kyei-Poku, van Frankenhuyzen, Johny and Guertin2012).
Isaria farinosa and Purpureocillium lilacinum (Thom) Luangsa-ard, Hou-braken, Hywel-Jones, and Samson (Hypocreales: Ophiocordycipitaceae) have been isolated and characterised from A. planipennis cadavers collected in southwestern Ontario (Johny et al. Reference Johny, Kyei-Poku, Gauthier and van Frankenhuyzen2012a). Bioassays indicated that they were not as virulent to EAB adults as were the commercial isolates of Isaria fumosorosea (LRC176), Metarhizium brunneum (LRC187), and B. bassiana (GHA) (Johny et al., Reference Johny, Kyei-Poku, Gauthier and van Frankenhuyzen2012a).
Forest managers may be hesitant to use broad-scale applications of generalist entomopathogens for fear of impacting non-target insects. Entomopathogens with broad host ranges may have limited use in integrated control programmes for A. planipennis, because of deleterious impacts on parasitoids used in classical and augmentative biological control programmes. Fortuitously, T. planipennisi appears to have very low to no susceptibility to B. bassiana, while S. agrili is only slightly susceptible when exposed to fungus-inoculated ash twigs that resulted in 100% mortality of EAB (Dean et al. Reference Dean, Vandenberg, Griggs, Bauer and Fierke2012). However, the native P. sulcata and Atanycolus species, appear to be susceptible to the fungus at high exposure levels (D.B.L., personal observation), but auto-contamination traps placed in the canopy for the dissemination of fungi (Lyons et al. Reference Lyons, Lavallee, Kyei-Poku, van Frankenhuyzen, Johny and Guertin2012) that are designed to target A. planipennis may mitigate their exposure.
Samples of A. planipennis were also infected with a nematode characterised using molecular techniques as a Rhabditis (Oscheius) Andrássy (Rhabditida: Rhabditidae) species (see Lyons Reference Lyons2010). Adults of the bronze birch borer, A. anxius Gory, were infected with a microsporidian, Cystosporogenes Canning, Barker, Nicholas, and Page (Pansporoblastina: Glugeidae) species with outbreak population infections rates as high as 85.4% (Kyei-Poku et al. Reference Kyei-Poku, Gauthier, Schwarz and van Frankenhuyzen2011). Microsporidia are pathogens that reduce the reproductive potential of the host (i.e., fecundity and longevity) but typically do not kill their host. Microsporidia are released as spores into the environment in the faeces of the beetle. Cross infectivity tests are being conducted to determine if this parasite has potential to affect populations of the closely related A. planipennis. The majority of bacteria found in the guts of EAB may be symbiotic, some digesting cellulose, but others such as Paenibacillus Ash, Priest, and Collins (Bacillales: Paenibacillaceae) species may be pathogenic and potentially serve as biological control agents (Vansanthakumar et al. Reference Vansanthakumar, Handelsman, Schloss, Bauer and Raffa2008).
Climate/weather
Low winter temperatures
The potential northern and high altitude distribution of an invading poikilotherm species is usually constrained by its ability to survive low winter temperatures, which is considered a mortality factor (Bale Reference Bale2002). The question often arises as to whether or not A. planipennis can survive the low winter temperatures that occur in parts of North America where ash grow naturally or has been planted. Insects have three modes in which they respond to cold temperatures (Lee Reference Lee2010). They can be chill intolerant whereby they succumb to temperatures above their freezing point. They can be freeze tolerant in which case they can survive freezing or they can be freeze intolerant and succumb when their tissues freeze. Typically, EAB overwinters in the prepupal (“J”-shaped larvae) stage. Prepupae of A. planipennis are freeze intolerant (Crosthwaite et al. Reference Crosthwaite, Sobek, Lyons, Bernards and Sinclair2011). The lower lethal temperature for freeze-intolerant species is usually their supercooling point (SCP). The SCP of A. planipennis in China was reported to range from −29.6 °C to −23.0 °C (Wu et al. Reference Wu, Li, Yang, Wang, Wang and Bai2007). In North America, the mean SCP for prepupae of A. planipennis was determined to be approximately −30 °C, with a minimum observed SCP of −35 °C (Crosthwaite et al. Reference Crosthwaite, Sobek, Lyons, Bernards and Sinclair2011). The borer achieves these low SCPs by producing large quantities of glycerol and smaller amounts of antifreeze proteins as cryoprotectants (Crosthwaite et al. Reference Crosthwaite, Sobek, Lyons, Bernards and Sinclair2011). In addition, A. planipennis prepupae have waxy cuticles that inhibit the penetration of ice crystals. The SCP was lowest during the coldest part of winter (January–February), decreasing in late fall and increasing in early spring (Crosthwaite et al. Reference Crosthwaite, Sobek, Lyons, Bernards and Sinclair2011). The maximum production of cryoprotectants also occurs during January and February. If prepupae of A. planipennis experienced a warm period (i.e., 10–15 °C) in winter, their SCP increased and was not reversible when they returned to cold temperatures (Sobek-Swant et al. Reference Sobek-Swant, Crosthwaite, Lyons and Sinclair2012a). They also lost antifreeze compounds. However, only extreme warming events for several days followed by extreme cold periods would affect survival of populations.
The temperature that insects experience in the microhabitat where they dwell is frequently very different from the ambient air temperatures recorded in meteorological stations. EAB overwinter as prepupae in pupal chambers in the inner bark (phloem) or the outer sapwood (xylem) of host trees, or as feeding-stage larvae in galleries in the cambial layer between the sapwood and bark. Mean daily temperature minimums were significantly warmer under the bark than comparable ambient air temperatures based on winter temperature recordings from data-logger probes in the under-bark microhabitats in woodlot and urban ash trees (Vermunt et al. Reference Vermunt, Cuddington, Sobek-Swant, Crosthwaite, Lyons and Sinclair2012b). Although low temperatures were reported from local weather stations, the insects under the bark did not always experience these temperature minimums, although on a few occasions the opposite was true where under-bark temperatures were lower than ambient air temperatures. Maximum temperatures on the north sides of ash trees were not significantly different from the ambient temperatures in spring, but the south sides of the trees were significantly warmer especially for open grown urban trees (Vermunt et al. Reference Vermunt, Cuddington, Sobek-Swant, Crosthwaite, Lyons and Sinclair2012b). These observations would affect predicted mortality resulting from low lethal temperatures or development times for EAB in the spring if predictions were based solely on ambient air temperatures. To overcome these problems, Vermunt et al. (Reference Vermunt, Cuddington, Sobek-Swant and Crosthwaite2012a) developed a model to predict the under-bark minimum temperature (T) for ash trees across a range of latitudes and longitudes in urban and woodlot environs based on ambient air temperature (A) where: T t+Δt=T t+0.11(A t+Δt–T t), t=time and the time step Δt=one hour. Some A. planipennis overwinter in the lower bole of host trees, above and below ground level, and may be insulated from extreme temperature minimums by snow cover. Models predicting temperatures under snow cover indicate that they are usually well above ambient temperature minimums (DeSantis et al. Reference DeSantis, Moser, Gormanson, Bartlett and Vermunt2013). DeSantis et al. (Reference DeSantis, Moser, Gormanson, Bartlett and Vermunt2013), using historical weather data, ash distribution data, models for under bark and under-snow temperatures, and lethal SCPs, predicted that there are areas in southern Canada (e.g., Manitoba, Saskatchewan, and Alberta) and northern United States of America (e.g., Minnesota, North Dakota, and Montana) where low winter temperatures might suppress populations of A. planipennis to below damaging thresholds. Liang and Fei (Reference Liang and Fei2014) made a case for the divergence of the range of ash trees in North America and the future distribution of A. planipennis under a climate change scenario. Under predicted climate-change outcomes the distribution of A. planipennis will shift northward and the distribution of ash trees will remain static, resulting in refugia in the south where ash trees will avoid attack. This idea is predicated on the supposition that management of A. planipennis will slow the spread of the beetle so that climate change will take place prior to ash in the south being destroyed by the pest. However, if the predictions by DeSantis et al. (Reference DeSantis, Moser, Gormanson, Bartlett and Vermunt2013) are correct, A. planipennis would have a greater impact on ash in the north and reduced mortality from lethal cold conditions. These studies examined the effect of winter temperatures on prepupae, but a portion of the population does not reach the prepupal stage and overwinters as feeding-stage larvae (Cappaert et al. Reference Cappaert, McCullough, Poland and Siegert2005). These are the larvae that will undergo a two-year life cycle. In more northern latitudes, we suspect that a greater portion of the population of A. planipennis will require two years to complete development. Consequently, more early instars will be exposed to extreme low temperatures. Little is known about the effects of low temperatures on these feeding-stage larvae.
Host resistance
Agrilus planipennis is rare in Asia where its primary natural hosts include Manchurian ash, Fraxinus mandshurica Ruprect and Chinese ash, F. chinensis Roxburgh (Liu et al. Reference Liu, Bauer, Gao, Zhao, Petrice and Haack2003; Rebek et al. Reference Rebek, Herms and Smitley2008; Wang et al. Reference Wang, Yang, Gould, Zhang, Liu and Liu2010). In China, A. planipennis is considered a secondary pest of its natural hosts and only colonises declining or stressed trees (Haack et al. Reference Haack, Jendek, Liu, Marchant, Petrice and Poland2002). In its native range, outbreaks of A. planipennis frequently occur in ash species (e.g., F. pennsylvanica and F. velutina Torrey) that have been introduced from North America (Wang et al. Reference Wang, Yang, Gould, Zhang, Liu and Liu2010). Unlike native Agrilus species like bronze birch borer, A. anxius Gory, and two-lined chestnut borer, A. bilineatus (Weber), which attack stressed trees (Anderson Reference Anderson1944; Dunn et al. Reference Dunn, Kimmerer and Nordin1986), A. planipennis apparently attacks and kills healthy trees in North America (Poland and McCullough Reference Poland and McCullough2006; Anulewicz et al. Reference Anulewicz, McCullough and Cappaert2007). Beetles may be initially attracted to stressed trees but once beetle populations increase, healthy trees are overcome (Poland and McCullough Reference Poland and McCullough2006). All North American ash may be susceptible to some degree. Blue ash, F. quadrangulata Michaux, may exhibit some inherent resistance (Anulewicz et al. Reference Anulewicz, McCullough and Cappaert2007; Pureswaren and Poland Reference Pureswaran and Poland2009; Tanis and McCullough Reference Tanis and McCullough2012). Anulewicz et al. (Reference Anulewicz, McCullough and Cappaert2007) reported that blue ash was less suitable for larvae of A. planipennis compared with green and white ash. This resulted from the formation of callous tissue around larval galleries in blue ash and not the other two species, and the induction of two-year life cycle larvae by blue ash. In a young ash plantation, the susceptibility of native F. pennsylvanica and F. americana cultivars to A. planipennis was compared with that of the exotic F. mandshurica (Rebek et al. Reference Rebek, Herms and Smitley2008). Based on mortality data and adult emergence hole data gathered over a three-year period, Manchurian ash was much more resistant than the North American species. The mechanisms by which some ash species are resistant to A. planipennis are poorly understood (Hill et al. Reference Hill, Whitehill, Opiyo, Phelan and Bonello2012). Variable resistance between ash species might be achieved by differences in physical properties or chemistry of phloem and foliage tissues such as constitutive and induced phytochemical defenses (Eyles et al. Reference Eyles, Jones, Riedl, Cipollini, Schwartz and Chan2007; Cipollini et al. Reference Cipollini, Wang, Whitehall, Powell, Bonello and Herms2011; Whitehill et al. Reference Whitehill, Popova-Butler, Green-Church, Koch, Herms and Bonello2011), expressed proteins (proteome) (Whitehill et al. Reference Whitehill, Opiyo, Koch, Herms, Cipollini and Bonello2012), and/or water and nutritional chemicals (Chen et al. Reference Chen, Ciaramitaro and Poland2011; Hill et al. Reference Hill, Whitehill, Opiyo, Phelan and Bonello2012). Host tree resistance to A. planipennis is generally thought to be confined to Asian species of ash with which A. planipennis has coevolved while North American species are deemed susceptible (Liu et al. Reference Liu, Bauer, Gao, Zhao, Petrice and Haack2003; Rebek et al. Reference Rebek, Herms and Smitley2008; Whitehill et al. Reference Whitehill, Popova-Butler, Green-Church, Koch, Herms and Bonello2011; Tanis and McCullough Reference Tanis and McCullough2012; DeSantis et al. Reference DeSantis, Moser, Gormanson, Bartlett and Vermunt2013). However, host tree defense mechanisms are the major cause of mortality of young EAB larvae (9.9–41.1%) in North American ash trees (Duan et al. Reference Duan, Ulyshen, Bauer, Gould and Van Driesche2010). These mortality effects were inversely proportional to larval density suggesting that increased larval numbers overcome the defence mechanisms. These mechanisms are most effective in spring and early summer during the period of maximum tree growth.
Competition and cannibalism
During surveys conducted in Michigan from 2002 to 2004 to determine mortality factors for A. planipennis, apart from natural enemies, larvae succumbed from starvation, desiccation, and cannibalism, especially in heavily infested logs (Bauer et al. Reference Bauer, Liu, Haack, Petrice and Miller2004). Anulewicz et al. (Reference Anulewicz, McCullough and Cappaert2007) observed ash trees that exhibited extensive dieback, but lacked A. planipennis emergence holes and woodpecker feeding holes. Dissection of the trees revealed almost 100% consumption of the phloem tissue and complete mortality of the second-instars and third-instars, which they attributed to intraspecific competition. Tluczek et al. (Reference Tluczek, McCullough and Poland2011) also observed high larval mortality resulting from nearly complete phloem consumption in girdled trees. Cannibalism in A. planipennis occurs when larvae encounter one another and results in strong density-dependent mortality (Duan et al. Reference Duan, Larson, Watt, Gould and Lelito2013a). On logs of F. pennsylvanica and F. uhdei Wenzig (Lingelsheim), cannibalism increased with increased population density and larval biomass decreased with increased larval density (Duan et al. Reference Duan, Larson, Watt, Gould and Lelito2013a). Cannibalism was classified as a larva that had been cut through or damaged by another larva, while starvation was classified as a larva that died without apparent damage. Larval survival declined from 68% to 10% in F. pennsylvanica and from 86% to 55% in F. uhdei as egg densities were systematically increased from 200 to 1600 eggs/m2 (Duan et al. Reference Duan, Larson, Watt, Gould and Lelito2013a). Intraspecific competition (cannibalism and starvation) by EAB larvae was greater on logs of F. pennsylvanica than on logs of F. uhdei, suggesting that tree species or condition can influence processes of population dynamics (Duan et al. Reference Duan, Larson, Watt, Gould and Lelito2013a).
Conclusions
When A. planipennis was first detected in North America we knew very little about the natural history and population dynamics of the insect in Asia or in its introduced range in North America (Haack et al. Reference Haack, Jendek, Liu, Marchant, Petrice and Poland2002). Since that time, considerable information has been gathered about many aspects of the life history of this species. Considerable knowledge about mortality factors was also gathered in the context of using imported natural enemies from China for classical biological control of the pest. Preliminary surveys of parasitoids, predators, and pathogens (Liu et al. Reference Liu, Bauer, Gao, Zhao, Petrice and Haack2003; Bauer et al. Reference Bauer, Liu, Haack, Petrice and Miller2004) were undertaken to assess the impact of North American natural enemies on A. planipennis and to demonstrate their inability to regulate populations of the pest and thus justify importation of Asian biological control agents. Subsequent regional surveys for parasitoids (Cappaert and McCullough Reference Cappaert and McCullough2009; Duan et al. Reference Duan, Fuester, Wildonger, Taylor, Barth and Spichiger2009; Reference Duan, Taylor, Fuester, Kula and Marsh2013b; Kula et al. Reference Kula, Knight, Rebbeck, Bauer, Cappaert and Gandhi2010; Lyons Reference Lyons2010), predators (Lindell et al. Reference Lindell, McCullough, Cappaert, Apostolou and Roth2008), and pathogens (Castrillo et al. Reference Castrillo, Bauer, Liu, Griggs and Vandenberg2010; Johny et al. Reference Johny, Kyei-Poku, Gauthier and van Frankenhuyzen2012a, Reference Johny, Kyei-Poku, Gauthier, van Frankenhuyzen and Krell2012b; Kyei-Poku and Johny Reference Kyei-Poku and Johny2013) were conducted to assess North American natural enemies of A. planipennis as augmentative or inundative biological control agents. These regional surveys produced some natural enemies that may have some potential for reducing A. planipennis populations (e.g., P. sulcata, A. cappaerti, B. bassiana) and have inspired additional and ongoing research (e.g., Lyons et al. Reference Lyons, Lavallee, Kyei-Poku, van Frankenhuyzen, Johny and Guertin2012).
The question about the impact of low winter temperatures on populations of A. planipennis, in more northern latitudes, often arises especially after extreme low temperature events. Investigations into the overwintering physiology of A. planipennis (Crosthwaite et al. Reference Crosthwaite, Sobek, Lyons, Bernards and Sinclair2011; Sobek-Swant et al. Reference Sobek-Swant, Crosthwaite, Lyons and Sinclair2012a) have explained the pest’s ability to survive via supercooling mechanisms. Investigations into the winter microclimate in the beetle’s microhabitat (Vermunt et al. Reference Vermunt, Cuddington, Sobek-Swant and Crosthwaite2012a, 2012Reference Vermunt, Cuddington, Sobek-Swant, Crosthwaite, Lyons and Sinclairb) in conjunction with the physiological results have allowed predictions of A. planipennis distribution at a landscape level (Sobek-Swant et al. Reference Sobek-Swant, Kluza, Cuddington and Lyons2012b; DeSantis et al. Reference DeSantis, Moser, Gormanson, Bartlett and Vermunt2013; Liang and Fei Reference Liang and Fei2014). However, we still do not know the effects of low winter temperature of feeding-stage larvae that will have a two-year life cycle and may be abundant in northern latitudes.
Assumptions are often made about the lack of resistance in native plants that are attacked by invasive insects, alluding to the lack of coevolutionary history between the species. We now know that tree resistance by North American ashes is a major source of mortality to young larvae of A. planipennis (Duan et al. Reference Duan, Ulyshen, Bauer, Gould and Van Driesche2010). The resistance mechanisms, however, are overcome by increased population pressure. Reduction of development rates in young larvae by these resistance mechanisms may also explain the preponderance of two-year life cycle larvae in new infestations. We now know that A. planipennis populations are constrained by phloem availability (McCullough and Siegert Reference McCullough and Siegert2007) and mortality by cannibalism is a function of competition for space (Duan et al. Reference Duan, Larson, Watt, Gould and Lelito2013a).
Future research into the population dynamics of A. planipennis should include the long-term role of native mortality agents in regulating populations and their potential exploitation as biological control agents. The interaction of these North American mortality agents with introduced classical biological control agents must also be understood. In peripheral parts of the potential and realised invaded distribution of A. planipennis we must be able to predict the effect of weather/climate on voltinism and survival of the beetle. To date, SCP investigations on A. planipennis have been limited to “J”-shaped larvae/prepupae (e.g., Crosthwaite et al. Reference Crosthwaite, Sobek, Lyons, Bernards and Sinclair2011; Sobek-Swant et al. Reference Sobek-Swant, Crosthwaite, Lyons and Sinclair2012a). The effect of low winter temperatures on overwintering feeding-stage larvae is unknown. With an increased preponderance of semivoltinism in A. planipennis populations in northern latitudes, the impact of low temperatures on these feeding-stage larvae must be taken into consideration when predicting overwintering survival. Increased incidence of semivoltinism might also increase the exposure of the life stages of the beetle to mortality agents. This is especially critical under a climate change scenario wherein moderating climates may result in northward range expansion and enhanced overwintering survival, but also begs the following questions. Are there southern limits to the potential distribution of the beetle resulting from high temperatures and can high overwintering temperatures interfere with diapause requirements? Will extreme weather events under climate change mitigate or exacerbate the impact of this invasive species? We also must understand the role of climate change on trophic-level relationships (e.g., synchrony, host resistance) in the A. planipennis system (e.g., Liang and Fei Reference Liang and Fei2014). Host resistance to A. planipennis is exhibited by North American ash trees and results in larval mortality. Selective breeding of ash trees that have survived the course of an A. planipennis outbreak (e.g., lingering ash) may produce more resistant cultivars.
Acknowledgements
The author would like to thank the organisers of the Emerald Ash Borer Workshop at the Entomological Society of Canada and Ontario Joint Annual Meeting in Guelph, Ontario, Canada (2013) for the opportunity to make the presentation that served as the foundation for this paper.
