Exotic Xylosandrus spp. ambrosia beetles established in non-native habitats have been associated with sudden and extensive attacks on a diverse range of living trees, but factors driving their shift from dying/dead hosts to living and healthy ones are not well understood. We sought to characterize the role of host physiological condition on preference and colonization by two invaders, Xylosandrus germanus and Xylosandrus crassiusculus. When given free-choice under field conditions among flooded and non-flooded deciduous tree species of varying intolerance to flooding, beetles attacked flood-intolerant tree species over more tolerant species within 3 days of initiating flood stress. In particular, flood-intolerant flowering dogwood (Cornus florida) sustained more attacks than flood-tolerant species, including silver maple (Acer saccharinum) and swamp white oak (Quercus bicolor). Ethanol, a key host-derived attractant, was detected at higher concentrations 3 days after initiating flooding within stems of flood intolerant species compared to tolerant and non-flooded species. A positive correlation was also detected between ethanol concentrations in stem tissue and cumulative ambrosia beetle attacks. When adult X. germanus and X. crassiusculus were confined with no-choice to stems of flood-stressed and non-flooded C. florida, more ejected sawdust resulting from tunneling activity was associated with the flood-stressed trees. Furthermore, living foundresses, eggs, larvae, and pupae were only detected within galleries created in stems of flood-stressed trees. Despite a capability to attack diverse tree genera, X. germanus and X. crassiusculus efficiently distinguished among varying host qualities and preferentially targeted trees based on their intolerance of flood stress. Non-flooded trees were not preferred or successfully colonized. This study demonstrates the host-selection strategy exhibited by X. germanus and X. crassiusculus in non-native habitats involves detection of stress-induced ethanol emission and early colonization of living but weakened trees.
Citation: Ranger CM, Schultz PB, Frank SD, Chong JH, Reding ME (2015) Non-Native Ambrosia Beetles as Opportunistic Exploiters of Living but Weakened Trees. PLoS ONE 10(7): e0131496. https://doi.org/10.1371/journal.pone.0131496
Editor: Garret Suen, University of Wisconsin-Madison, UNITED STATES
Received: March 22, 2015; Accepted: June 1, 2015; Published: July 2, 2015
This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This research was funded by the USDA Floriculture and Nursery Research Initiative and base funds associated with USDA ARS Research Project 3607-22000-012-00D (National Program 304-Crop Protection and Quarantine), and the Horticultural Research Institute. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Non-native ambrosia beetles (Coleoptera: Curculionidae) in the tribe Xyleborini have been regularly intercepted at ports of entry and several species are now established and abundant in North America [1–9]. A cryptic nature, diet of symbiotic fungi, sib-mating, haplodiploidy, broad host range, and interactions with naïve hosts likely aid in their introduction and establishment in non-native habitats [1–9]. In particular, two species native to Southeast Asia, Xylosandrus germanus (Blandford) and Xylosandrus crassiusculus (Motschulsky), have become widely established across regions of North America [2,10]. As of 2010, X. germanus and X. crassiusculus were found in 32 and 29 of 50 states, respectively . Xylosandrus germanus has also become established throughout much of Europe [11–16]. Xylosandrus crassiusculus currently has a narrower distribution in Europe, but is established in Central America and the Caribbean, East and West Africa, and Oceania [17,18].
Xylosandrus germanus and X. crassiusculus exhibit a capability to attack a diverse range of >200 and >120 species, respectively [19,20], and have extensively attacked living and apparently-healthy trees growing within non-native ornamental, horticultural, and forested settings [9,11,12,21–26]. However, the basis for these attacks is not well understood, particularly since both species colonize dying and dead trees in their native habitats . Over a dozen additional species of ambrosia beetles that normally attack dying or dead trees have also shifted to attacking living trees in non-native habitats; thereby leading to speculation about the mechanisms underlying their selection of living trees as hosts [4,9,27].
As with other Scolytinae, host-derived olfactory cues, particularly ethanol, play an important role during host selection by X. germanus and X. crassiusculus [9,28–30]. Ethanol is a key host-derived attractant and baiting, irrigating, or injecting trees with ethanol induced attacks on specific trees [9,28,31–33]. Xylosandrus germanus also efficiently located and attacked ethanol-injected trees, but rarely landed on and did not attack adjacent trees not emitting ethanol . A positive correlation was also demonstrated between concentration of ethanol injected into a tree and cumulative ambrosia beetle attacks , along with a positive correlation between emission rates from lures and corresponding trap captures . In addition to using host-derived volatiles, some Scolytinae and Platypodinae ambrosia beetles produce aggregation pheromones to attract conspecifics and overwhelm tree defenses [34,35]. However, this strategy is uncommon among the Xyleborini ambrosia beetles and X. germanus and X. crassiusculus are not anticipated to produce aggregation pheromones .
Ethanol is rapidly induced by trees in response to a variety of abiotic and biotic stressors, including flood and drought stress, girdling, freezing, pathogens, root and crown disturbance, and pollutants [36–43]. In the case of flooding, roots subjected to little or no oxygen will switch from aerobic to anaerobic cellular respiration [41,42]. Pyruvate formed during glycolysis is then converted into ethanol and either metabolized by alcohol dehydrogenase or emitted from the epidermis. Field and experimental observations have documented the potential for flood stress to predispose trees to attack by non-native ambrosia beetles. For instance, extensive ambrosia beetle attacks were detected in May 2011 on field grown dogwood (Cornus florida × Cornus kousa) trees that were subjected to water logging and poor soil drainage at two ornamental nurseries following record setting precipitation . Subsequent experiments demonstrated flood-stressed C. florida, which is considered intolerant of flooding, were more attractive and preferentially attacked by X. germanus and X. crassiusculus . Ethanol was also detected within the vascular tissue and emitted from the epidermis of C. florida, but not from non-flooded trees.
Comparatively high concentrations of ethanol can be associated with flood intolerant trees, but some tolerant species delay or avoid accumulating ethanol [44,45]. Thus, allowing X. germanus and X. crassiusculus to choose among tree species varying in their tolerance of flood stress could provide insight into their preference for trees in a specific physiological condition. Previous studies indirectly suggested X. germanus could distinguish among even slight differences in tree quality [9,21,22]. Furthermore, the extent to which X. germanus and X. crassiusculus utilize healthy naïve trees in non-native habitats, especially in the absence of weakened ones, is also poorly understood and unsubstantiated. Attacks have been suspected to occur on apparently-healthy trees [10,21,46–48], but such trees may actually have been inapparently-stressed at the time of attack . Comparing colonization success on flood-stressed and non-flooded trees could thereby provide crucial insight into host utilization within non-native habitats and aid in characterizing factors contributing to the invasiveness of X. germanus and X. crassiusculus.
The overall goal of this study was to assess factors involved with X. germanus and X. crassiusculus targeting living trees in non-native habitats, particularly the role of host physiological condition on preference behavior and colonization success. To characterize host preference specificity, we tested if tree species intolerant of flooding would be preferentially attacked under free choice conditions over species more tolerant of flooding. Ethanol in stem tissues of non-flooded and flood-stressed trees was also analyzed to determine if elevated concentrations were associated with the more intolerant species. To elucidate limitations imposed by tree condition on colonization success, we also compared the capability of X. germanus and X. crassiusculus to colonize flood-stressed and non-flooded trees. We hypothesized that X. germanus and X. crassiusculus can distinguish among trees of varying physiological conditions and preferentially attack species intolerant of flood stress, but exhibit non-preference and poor colonization of non-flooded trees. We also hypothesized interspecific variability in ethanol stem concentrations would be related to intolerance of flood stress.
Materials and Methods
Preference among Trees Varying in their Tolerance of Flood Stress
Free-choice experiments were conducted in Ohio and Virginia to characterize the role of host physiological condition on preference of X. germanus and X. crassiusculus and to determine if tree species intolerant of flooding would be preferentially attacked over moderately tolerant or tolerant species. Flowering dogwood C. florida (≈20 mm stem diam. at 15.24 cm above soil), Japanese snowbell Styrax japonicus S. et Z. (≈23 mm), and eastern redbud Cercis canadensis L. (≈25 mm) were selected as representatives of flood-stress intolerant species [49–51]; flowering cherry Prunus serrulata Lindl. (≈28 mm), American elm Ulmus americana L. (≈19 mm), and river birch Betula nigra L. (≈21 mm), were selected as moderately tolerant species [49,52,53]; and swamp white oak Quercus bicolor Willd. (≈19 mm) and silver maple Acer saccharinum L. (≈19 mm) were selected as tolerant species for deployment in Ohio [49,51]. Cornus florida and C. canadensis were selected as intolerant species; P. serrulata and M. virginiana represented moderately tolerant species; and Q. bicolor and A. saccharinum represented tolerant species for deployment in Virginia. Trees were 2−3 years old and maintained in 19-L pots.
Flooding conditions were imposed using a pot-in-pot system according to Ranger et al. , whereby a 26-L pot was first lined with a plastic waste bag of 3 mil (0.076 mm) thickness. A 19-L pot containing a single tree was then placed within the plastic lined pot. Flood stress was initiated by irrigating the media within the internal pot until standing water pooled around the base of the trees. Trees were checked daily to ensure standing water was maintained in the selected pots throughout the duration of experiments. Excess plastic liner surrounding the edge of the flooded pot was draped around the internal circumference of the pot and wrapped around the base of the stem to prevent beetles from landing in the standing water.
Non-flooded and flooded specimens of A. saccharinum, B. nigra, C. canadensis, C. florida, P. serrulata, Q. bicolor, S. japonicus, and U. americana were deployed in eight randomized complete blocks within a woodlot in Ohio (40°45’41.97”N 81°51’20.89”W). Permission was provided by The Ohio State University to use the described field sites. There was approximately 3 m between trees within each block and 10 m between adjacent blocks. Flood stress was initiated on 29 April 2013 and maintained until 14 May 2013. Similarly, non-flooded and flooded specimens of A. saccharinum, C. canadensis, C. florida, M. virginiana, P. serrulata, and Q. bicolor and were deployed in six randomized complete blocks along the edge of a woodlot in Virginia (37°17’17.64”N 76°39’0.07”W). Permission was provided by the Department of Parks and Recreation, York County, Virginia. Flood stress was initiated on 8 April 2013 and maintained until 23 April 2013 in Virginia. Trees were inspected for ambrosia beetle attacks one day after initiating flood stress and then every two to three days until the experiment was terminated.
Trees deployed in Ohio that sustained at least one ambrosia beetle attack were cut at the roots under field conditions on the last day of the experiment and subsequently transferred to a walk-in refrigerator held at 5°C. Trees were then dissected to recover foundress ambrosia beetles from the tunnels and galleries, and specimens were preserved in 95% ethanol for identification. Attacked trees deployed in Virginia were also cut at the base on the last day of the experiment and stem sections were held in plastic containers at room temperature until beetles were excavated for identification. No protected species were sampled during the course of this study.
Association of Ethanol with Trees Varying in their Tolerance of Flood Stress
Ethanol was quantified within the vascular tissue of the flood-stressed and non-flooded trees deployed in Ohio and Virginia as part of the experiment evaluating ambrosia beetle preference for trees varying in their tolerance of flooding. Based on Ranger et al. , three days after initiating flooding an Osborne arch punch (C.S. Osborne & Co., Harrison, New Jersey) was used to take four tissue core samples (1 mm depth, 5 mm diam.) at 10 cm above the base from flooded and non-flooded specimens of A. saccharinum, B. nigra, C. canadensis, C. florida, P. serrulata, Q. bicolor, S. japonicus, and U. americana deployed in Ohio. Similarly, tissue samples were also taken three days after initiating flooding from flood-stressed and non-flooded specimens of A. saccharinum, C. canadensis, C. florida, P. serrulata, Q. bicolor, and M. virginiana deployed in Virginia. The superficial tissue core samples included the outer bark, phloem, and vascular cambium, but not the sapwood and heartwood. Tissue core samples were placed in 2 mL Eppendorf tubes immediately after sampling, which were then covered among dry ice until transported back to the laboratory. Tissue samples were stored at -80°C until analysis. Tissues samples taken in Virginia were packed with dry ice and shipped overnight to the USDA-ARS-HIRL for analysis.
Solid phase microextraction-gas chromatography-mass spectrometry (SPME-GC-MS) was used to analyze ethanol content [9,43]. In short, phloem and vascular cambium tissues from the four core samples were placed in 2 mL glass vials and suspended in a water bath at 100°C for 30 min. Vials were removed from the water bath and a SPME fiber was exposed to the headspace within the vial for 5 min. The fiber was retracted after sampling and the syringe was immediately capped with a sealed section of PTFE microbore tubing (0.568 mm inner diam. × 1.07 mm outer diam., Cole-Parmer, Vernon Hills, Illinois) to prevent contamination of the fiber. A coating of carboxen-polydimethylsiloxane (CAR/PDMS; 75 μm coating; Sigma-Aldrich, St. Louis, Missouri) was used, which is ideal for gases and low molecular weight compounds (MW 30−225) .
Immediately after sampling, syringes were inserted into the injection port of a GC (Varian CP-3800; Varian Inc., Palo Alto, California) equipped with a Merlin Microseal septum system (Sigma-Aldrich) and thermally desorbed for 2 min at 250°C. The injection port was operated in splitless mode from 0–2 min and then split at a ratio of 1:20 for the remainder of the run. A capillary nonpolar DB-5 column (0.25 μm × 30 m × 0.25 μm; cross-linked/surface bonded 5% phenyl, 95% methylpolysiloxane; Agilent J&W, Santa Clara, California) was used for analysis according to the following program: 40–60°C at 3°C/min and 60–230°C at 20°C/min. A Varian 2200 MS detector was operated in electron impact mode with a scan range of 14−415 m/z. System control was accomplished with Star Chromatography Workstation software (Varian Star Toolbar, version 6.8). Fibers were conditioned before each analysis by exposure within the injection port for 20 min at 250 oC.
The external standard method [9,43] was used for determining concentrations of ethanol associated with the tissue core samples. Serial dilutions of ethanol ranging from 100 g/L to 0.0001 g/L were made in water and aliquots containing known amounts of ethanol were applied to filter paper discs (5 mm diam.) sealed in glass 2 mL autosampler vials for sampling purposes. Peak areas associated with the ethanol standards were measured using the Star Chromatography Workstation software. A standard concentration curve was then developed and used to determine concentrations of ethanol associated with the tissue samples.
Preference for Attacking Tree Parts
To document species preferences for attacking certain tree parts, we compared the propensity of ambrosia beetles to attack above-ground aerial parts of flood-stressed and non-flooded C. florida. Trees were flood-stressed as previously described and permission was provided by OSU to deploy trees in a woodlot in Ohio (40°45’41.97”N 81°51’20.89”W). Flooded and non-flooded trees were arranged 1 m apart in pairs with 5 m between adjacent pairs. Six pairs of replicated trees were deployed and trees were held under field conditions for 15 days from 7 July 2014 to 22 July 2014. Attacked trees were then excised at the base and transferred to a walk-in refrigerator held at 5°C. Tunnels and galleries within trees were then dissected to recover foundress ambrosia beetles, which were quantified and identified to species. Presence of the white ambrosial form (i.e. conidia and sprout cells) of the symbiotic fungus  and ambrosia beetle eggs within the galleries were also recorded.
Colonization Success on Flood-Stressed and Non-Flooded Trees
No-choice experiments were conducted in Ohio using X. germanus and in Virginia using X. crassiusculus to compare colonization success on flood-stressed and non-flooded C. florida. Bottle traps described by Ranger et al.  were used for obtaining field-collected adults, but instead of propylene glycol, a moistened paper towel rolled into a tube was placed in the bottom collection vessel of the trap to maintain ambrosia beetle specimens during 24 hrs under field conditions. Ambrosia beetles were then returned to the laboratory and transferred to parafilm-sealed petri dishes containing moistened filter paper and stored for 24 to 48 hrs at 3.3°C.
Flooding of container grown trees was established using the previously described pot-in-pot system on 19 May 2014 in a greenhouse on the campus of the OARDC and on 12 May 2014 on the campus of the HRAREC. Three days after initiating flooding, an individual adult female X. germanus field-collected in Ohio or X. crassiusculus field-collected in Virginia were placed inside of a chamber made of polytetrafluoroethylene (PTFE) tubing that was cut longitudinally (2.5 cm × 1 cm × 0.9 cm; l × w × h) and sealed at both ends with Molded Thermogreen LB-2 Septa (Supelco, Bellefonte, PA) cut into a semi-circle (Fig 1). Cables ties were used to snugly secure the chambers against the stem in parallel, thereby confining an individual beetle to 2.5 cm2 of bark tissue. Smooth bark of the C. florida trees allowed for close contact between the tissue and chambers, thereby effectively confining the beetle specimens. Ten chambers per tree were placed on six flood-stressed and six non-flooded C. florida trees starting from the base and extending linearly about 60 cm up the main stem with about 1.5−2 cm between cages. Generally, one to two unresponsive or injured beetles per tree were removed and replaced during the first day. Chambers confining foundress ambrosia beetles were held in place for 25 days, after which the stems were cut at the base and temporarily stored at 5°C until further analysis. As an indication of tunneling activity, chambers were carefully removed from the stems 1‒2 d later and ejected sawdust within each cage was weighed. Stem sections associated with each chamber as part of the experiment conducted in Ohio were also dissected to determine if the foundress was still alive and assess the presence/absence of eggs, larvae, pupae, and fungal growth within each tunnel/gallery. Flooded trees were drained 15 days into the experiment to avoid tree death and then watered accordingly for the remaining 10 days.
Cumulative ambrosia beetle attacks on trees varying in their tolerance of flooding were first examined using a repeated measures ANOVA to test for between-subject effects . When a significant between-subject treatment × time effect was detected (P <0.05), the number of attacks per treatment at a given time point were analyzed by two-way ANOVA using tree species and flood level as main effects and tree species × flood level as an interaction . Differences among treatments were determined using least-squares means (α = 0.05). Two-way ANOVA and least-squares means were also used to compare ethanol concentrations associated with the flood-stressed and non-flooded tree species (main effects = tree species and flood level; interaction = tree species × flood level). Regression analysis was used to test for a correlation between day 3 ethanol concentrations and cumulative attacks per tree associated with flood-stressed and non-flooded trees. Two-way ANOVA and least-squares means were used to compare cumulative attacks and ambrosia beetle specimens excavated per tree from stems and branches of flood-stressed and non-flooded trees (main effects = tree part and flood level; interaction = tree part × flood level). An unpaired t-test (α = 0.05) was used to compare ejected sawdust, and the percentage of galleries per tree with living foundresses, eggs, larvae, pupae, and fungal growth between flood-stressed and non-flooded trees. All data were square root transformed prior to analysis, but untransformed data are presented.
Preference among Trees Varying in their Tolerance of Flood Stress
Within 3 days of imposing flood stress, ambrosia beetles began rapidly attacking tree species intolerant of flooding, C. florida and S. japonicus, when given free choice among flooded and non-flooded trees of varying tolerance (Fig 2A). A significant between-subject treatment × time effect was detected in cumulative ambrosia beetle attacks for trees deployed in Ohio (F = 22.23; df = 15,112; P <0.0001) (Fig 2A). By 5 days, a significant interaction effect was detected between tree species × flooding level (F = 12.27; df = 7; P <0.0001), along with significant main effects associated with tree species (F = 11.54; df = 7; P <0.0001) and flooding level (F = 42.18; df = 1; P <0.0001). Specifically, S. japonicus and C. florida sustained significantly more attacks 5 days after initiating flooding than all other flooded and non-flooded species, while C. canadensis received more attacks than all the remaining treatments. By 15 days, C. florida and S. japonicus sustained the highest number of cumulative attacks followed by C. canadensis (species × flooding interaction: F = 15.07; df = 7; P <0.0001; tree species effect: F = 14.59; df = 7; P <0.0001; flooding effect: F = 71.12; df = 1; P <0.0001). No attacks occurred on flood-stressed B. nigra and Q. bicolor or any of the non-flooded tree species throughout the duration of the experiment.
Data for non-flooded trees are not presented in the figure. Different letters indicate significant differences among treatments at day 15 by two-way ANOVA and least-squares means (P value provided for tree species × flooding interaction effect; see Results section for remaining output; n = 8 and n = 6 trees per treatment for Ohio and Virginia, respectively).
Non-native X. germanus was the predominant species excavated from galleries created in C. florida and S. japonicus deployed in Ohio, but specimens of the non-native species Ambrosiodmus rubricollis (Eichhoff) and native Monarthrum mali (Fitch) were also recovered (Table 1). Xylosandrus germanus (n = 102 specimens) represented 98.0% of the total ambrosia beetles recovered from C. florida, followed by 1.0% for A. rubricollis (n = 1) and 1.0% for M. mali (n = 1). Only X. germanus was recovered from flood-stressed S. japonicus (n = 112), C. canadensis (n = 35), P. serrulata (n = 8), A. saccharinum (n = 2), and U. americana (n = 1) (Table 1).
A significant between-subject treatment × time effect was also detected for flood-stressed and non-flooded trees deployed in Virginia (F = 158.34; df = 11, 60; P <0.0001) (Fig 2B). Attacks began to occur on flood-stressed C. florida three days after initiating flooding. By 7 days, a significant interaction effect was detected between tree species × flooding level (F = 43.8; df = 5; P <0.0001), along with significant effects associated with the main factors of tree species (F = 42.8; df = 5; P <0.0001) and flooding level (F = 48.6; df = 1; P <0.0001). Cornus florida sustained more attacks 7 days after initiating flooding than all other flooded and non-flooded species.
By 15 days after initiating flooding, more cumulative attacks occurred on flood-stressed C. florida than all other flooded and non-flooded trees (species × flooding: F = 107.1; df = 5; P <0.0001; tree species effect: F = 106.3; df = 5; P <0.0001; flooding effect: F = 162.2; df = 1; P <0.0001). Considerably fewer attacks occurred on flood-stressed C. canadensis and P. serrulata, which were higher than all the remaining non-flooded and flooded species treatments except flood-stressed M. virginiana. No difference in attacks were detected among non-flooded A. saccharinum, C. canadensis, C. florida, M. virginiana, P. serrulata, and Q. bicolor, along with flood-stressed A. saccharinum, M. virginiana, and Q. bicolor. No attacks occurred on flood-stressed A. saccharinum and Q. bicolor, or non-flooded C. florida, M. virginiana, and P. serrulata. However, one non-flooded C. canadensis sustained one attack.
Only non-native ambrosia beetles attacked experimental trees deployed in Virginia. Xylosandrus germanus was the predominant species excavated from galleries created in C. florida deployed in Virginia, but specimens of non-native X. crassiusculus and non-native Xylosandrus compactus (Eichhoff) were also recovered (Table 2). Xylosandrus germanus (n = 241 specimens) represented 70.3% of the total ambrosia beetles recovered from C. florida, followed by 28.6% for X. crassiusculus (n = 98) and 1.2% for X. compactus (n = 4).
Association of Ethanol with Trees Varying in their Tolerance of Flood Stress
At 3 days after initiating flooding, differences were detected in ethanol concentrations associated with trees varying in their tolerance of flooding. A significant interaction effect on ethanol concentrations was detected between tree species and flooding for trees deployed in Ohio (F = 10.29; df = 7; P <0.0001) (Fig 3A). Significant main effects were also associated with tree species (F = 9.85; df = 7; P <0.0001) and flood stress (F = 59.41; df = 1; P <0.0001). Higher levels of ethanol were detected in tissues collected from flood-stressed C. florida, S. japonicus, and C. canadensis compared to all the remaining treatments (Fig 3A). Ethanol levels associated with A. saccharinum were higher than the remaining treatments, while no differences were detected among flooded and non-flooded Q. bicolor, B. nigra, U. americana, and P. serrulata, and non-flooded A. saccharinum, C. canadensis, C. florida, and S. japonicus. A significant positive correlation (r2 = 0.52) was detected between day 3 ethanol concentrations and cumulative attacks on the flood-stressed and non-flooded trees (F = 129.03; df = 1, 126; P <0.0001) (Fig 4A).
Ethanol concentrations were analyzed 3 days after initiating flood stress. Different letters indicate significant differences among treatments by two-way ANOVA and least-squares means (P value provided for tree species × flooding interaction effect; see Results section for remaining output; n = 8 and n = 6 trees per treatment for Ohio and Virginia, respectively).
See Fig 3A and 3B for tree species tested and Results section for additional statistical output (n = 128 and n = 72 trees deployed in Ohio and Virginia, respectively).
A significant interaction was also detected between tree species and flooding level on ethanol concentrations associated with trees deployed in Virginia (F = 12.30; df = 5; P <0.0001), along with significant main effects for tree species (F = 9.94; df = 5; P <0.0001) and flooding level (F = 32.92; df = 1; P <0.0001) (Fig 3B). Higher levels of ethanol were detected in flood-stressed C. florida tissues compared to all the remaining flooded and non-flooded treatments (Fig 3B). Ethanol levels associated with flood-stressed P. serrulata were considerably lower than C. florida, but higher than all the remaining treatments except for C. canadensis. A significant positive correlation (r2 = 0.76) was detected between day 3 ethanol concentrations and cumulative attacks on the flood-stressed and non-flooded trees (F = 222.61; df = 1, 70; P <0.0001) (Fig 4B).
Preference for Attacking Tree Parts
Stems of flood-stressed C. florida deployed under field conditions in Ohio were preferentially attacked over branches of flood-stressed trees, and stems and branches of non-flooded trees (Fig 5A–5D). A mean of 199.8 and 35.7 attacks occurred on stems and branches of flood-stressed trees, while 1.5 and 0.0 attacks occurred on stems and branches of non-flooded trees, respectively (Fig 5A). A significant interaction was detected between the influence of tree part and incidence of flood stress on cumulative attacks (F = 25.18; df = 1; P <0.0001), along with significant main effects associated with tree part (F = 34.08; df = 1; P <0.0001) and flood stress (F = 145.70; df = 1; P <0.0001) (Fig 5A). Attacks were not observed to occur on any tree parts other than stems and branches.
Different letters within a biological parameter indicate significant differences by two-way ANOVA and least-squares means (P value provided for tree part × flooding interaction effect; see Results section for remaining output; n = 6 flood-stressed and non-flooded trees).
Stems on two of the six non-flooded C. florida trees sustained one and eight cumulative attacks per tree; no attacks occurred on the remaining non-flooded trees. The two non-flooded trees attacked by ambrosia beetles were paired with the two most heavily attacked flood-stressed trees that sustained 338 and 337 attacks per tree, respectively, compared to the remaining flood-stressed trees with 173, 260, 120, and 185 attacks per tree. Only superficial tunneling occurred on the two non-flooded trees and no ambrosia beetles were excavated from the tunnels (Fig 5B–5D). In contrast, a mean (±SE) of 23.0 ± 5.2 percent of attacks on the flood-stressed trees resulted in galleries, 73.3 ± 6.4 percent of which contained the white ambrosial form (i.e. conidia and sprout cells) of the symbiotic fungus and 14.0 ± 3.9 percent of which contained symbiotic fungus and ambrosia beetle eggs, respectively.
Only non-native ambrosia beetles were excavated from the flood-stressed trees (Fig 5A–5D). Xylosandrus germanus (n = 676) represented 85.1% of the total specimens excavated from the stems, followed by 14.2% for Anisandrus maiche Stark (n = 113) and 0.6% for X. crassiusculus (n = 5). Xylosandrus germanus (n = 56) was also the dominant species recovered from C. florida branches and represented 73.7% of the total excavated specimens followed by 26.3% for A. maiche (n = 20).
More X. germanus, A. maiche, and X. crassiusculus were excavated from stems of flood-stressed trees compared to branches on flood-stressed trees and stems and branches on non-flooded trees (Fig 5B–5D). A significant interaction was detected between the influence of tree part and flood stress on the number of X. germanus excavated per tree (F = 162.10; df = 1; P <0.0001), along with significant main effects for tree part (F = 52.72; df = 1; P <0.0001) and impact of flood stress (F = 54.24; df = 1; P < 0.0001) (Fig 5B). A significant interaction (F = 9.91; df = 1; P <0.0001), tree part main effect (F = 9.15; df = 1; P = 0.007), and flood stress main effect (F = 40.66; df = 1; P < 0.0001) were detected for A. maiche (Fig 5C). A significant interaction (F = 25.01; df = 1; P <0.0001), tree part main effect (F = 25.01; df = 1; P <0.0001), and flood stress main effect (F = 25.01; df = 1; P <0.0001) were also detected for X. crassiusculus (Fig 5D).
Colonization Success on Flood-Stressed and Non-Flooded Trees
When confined under no-choice conditions, more sawdust resulting from burrowing/tunneling activities accumulated in chambers confining adult female X. germanus (t = 4.74; df = 10; P = 0.0008) and X. crassiusculus (t = 7.91; df = 10; P <0.0001) to flood-stressed C. florida compared to non-flooded trees (Fig 6A and 6B). Living foundress X. germanus were detected within tunnels/galleries created in flood-stressed C. florida, but no living specimens were associated with non-flooded trees (t = 11.38; df = 10; P <0.0001) (Fig 6C). Living foundress X. crassiusculus were also detected within tunnels/galleries created in flood-stressed C. florida, but no living specimens were associated with non-flooded trees (t = 1.58; df = 10; P = 0.07) (Fig 6D). Eggs of X. germanus were detected in galleries created in flood-stressed trees, but no eggs were associated with beetles confined to non-flooded trees (t = 4.65; df = 10; P = 0.001) (Fig 6E). No X. crassiusculus eggs were associated with flood-stressed or non-flooded trees (Fig 6F). Larvae of X. germanus (t = 4.17; df = 10; P = 0.002) and X. crassiusculus (t = 4.17; df = 10; P = 0.002) were recovered from galleries within flood-stressed trees, but no larvae were associated with non-flooded trees (Fig 6G and 6H). Pupae of X. germanus were recovered from flood-stressed trees, but not from non-flooded trees (t = 0.11; df = 10; P = 0.11) (Fig 6I). No X. crassiusculus pupae were associated with flood-stressed or non-flooded trees (Fig 6J).
Mean (±SE) ejected sawdust per chamber (A-B) and percentage of galleries per tree with living foundresses (C-D), eggs (E-F), larvae (G-H), and pupae (I-J). Tunnels/galleries were excavated 25 days after confining individual foundress beetles. Differences within a species and biological parameter were compared using an unpaired t-test (see Results section for full statistical analysis output; 10 chambers containing 1 adult female were placed on n = 6 flood-stressed and non-flooded trees; NA = no analysis).
Ambrosia beetles, including X. germanus and X. crassiusculus, have traditionally been viewed as secondary colonizers of dying or dead trees, or those already under attack by more aggressive tree killing bark beetles [4,6]. Attacks on living trees, even within non-native habitats, were limited to injured parts of trees . Yet, a variety of ambrosia beetles established in non-native habitats have reportedly begun attacking living trees [4,6,27]. Results from our current study demonstrate that while living trees were indeed attacked by non-native X. germanus and X. crassiusculus, beetles efficiently distinguished among varying host qualities and rapidly targeted species intolerant of flood stress. A few attacks occurred on non-flooded trees under free choice field conditions, but the tunnels were superficial, abandoned, and no galleries were formed. Subsequent experiments under no-choice conditions demonstrated that both X. germanus and X. crassiusculus readily colonized flood-stressed trees, as indicated by the production of eggs, larvae, and pupae, but were unwilling or unable to colonize non-flooded trees. A positive correlation was also demonstrated between ethanol stem concentrations and ambrosia beetle attacks. We therefore propose the basis for the shift by X. germanus and X. crassiusculus from attacking dying/dead trees to living ones involves early colonization of living but weakened trees.
Hulcr and Dunn  suggested a component of exotic ambrosia beetles attacking living trees may involve an ‘olfactory mismatch’ whereby beetles perceive volatiles emitted from live trees as representative of emissions from dead trees. Our current study indicates the stress-induced emission of ethanol from living trees plays an important role in the selection of living but weakened trees. An asymptomatic outward appearance of living but weakened trees in the early onset of physiological stress can be anthropogenically perceived to indicate a living and healthy tree being targeted for attack . However, physiologically stressed trees can visually appear healthy, but emit ethanol and other stress-related compounds that function as important host-location cues for opportunistic ambrosia beetles [9,43]. Ethanol was detected within stem tissues of flood-stressed trees by 3 days after initiating flooding as part of our current study. Beetles also began attacking flood-stressed C. florida and S. japonicus within 3 days of initiating flooding as part of our current study. Similarly, X. germanus began landing on flood-stressed C. florida within 1 day of initiating flooding . These results indicate the rapid production and emission of stress-induced volatiles, particularly ethanol, from otherwise apparently-healthy trees are advantageously used by X. germanus and X. crassiusculus to locate vulnerable hosts. Rapidly locating resources in a weakened physiological state that are often spatially and temporally unpredictable is an important adaptive trait for insects that primarily or exclusively target stressed hosts . Thus, selection pressures may be favoring individuals that rapidly locate and successfully colonize living but recently weakened trees.
Semiochemical-mediated attacks by aggressive bark beetles on an individual tree are demonstrated to result in beetles ‘spilling over’ and attacking neighboring healthy trees [58,59]. Such spill-over of X. germanus and X. crassiusculus from heavily infested trees onto neighboring healthy trees may occur under natural settings, and colonization may then be facilitated by infecting the hosts with symbiotic and auxiliary fungi. Spill-over could explain the clustering of attacked trees documented to occur in diverse landscapes , but localized site conditions may have predisposed clusters of trees to attack. However, numerous free-choice experiments have not detected attacks by X. germanus, X. crassiusculus, and other ambrosia beetles on untreated trees adjacent to flood-stressed trees or those baited, irrigated, or injected with ethanol [9,28,29,31–33,60]. A few attacks on neighboring non-flooded (and apparently-healthy) trees were documented as part of our current study, but the tunnels were superficial and abandoned. If X. germanus and X. crassiusculus do attempt to colonize healthy trees adjacent to stressed ones, our current study suggests they are poor colonizers of such trees. Since oviposition by various Xyleborine species does not occur prior to establishment of the symbiotic fungus , host factors disrupting fungal establishment could have important implications for the role of host physiological condition on tree utilization by non-native species.
Our current study used flood-stress as a technique for characterizing attacks on living trees by X. germanus and X. crassiusculus, particularly the role of host physiological condition on preference behavior and colonization success. Other physiological stressors also likely play a role in beetles targeting living but weakened trees in natural and managed non-native habitats. For instance, spring frost events following a record-setting mild winter preceded attacks on trees intolerant of late frost injury, namely, Acer palmatum Thunb., C. canadensis, Liridodendron tulipifera L., S. japonicus, and Zelkova serrata (Thunb.) . Similarly, forested stands of beech trees, Fagus sylvatica L., regarded as apparently-healthy at the time of attack by X. germanus and other ambrosia beetles were subjected to extreme frost events in the year prior to initiation of wide spread attacks [11,62]. Subsequent studies  demonstrated experimentally frost injured tissues on F. sylvatica trees were more attractive than non-injured trees to X. germanus and other non-native ambrosia beetles. Ambrosia beetle attacks were also observed on the frost injured tissues.
Generally, researchers rarely have the opportunity to assess the physiological condition of a host prior to discovering the initiation of attack by wood-boring insects . As a result, many species initially considered fully capable of colonizing trees ranging from healthy to decaying were subsequently found to restrict their attacks to hosts in a specific condition. Ploetz et al.  discussed examples of ambrosia beetle attacks and associated diseases in orchards and ornamental nurseries, and pointed out that in none of the provided cases was the pre-existing health of the trees known at the time of attack. However, the pronounced incidences of extensive attacks and associated diseases within the orchards/nurseries compared to natural settings suggests the affected trees were predisposed to attack . Similarly, field observations involving orchard-grown black walnut trees, Juglans nigra L., considered apparently-healthy at the time of attack were serendipitously determined to have exhibited slower growth rates in the year before attack compared to non-attacked trees . Dieback, basal sprouts, and Fusarium cankers were also associated with trees not attacked by X. germanus , thereby indicating trees within the orchard may have been in a weakened state prior to beetles attacking.
Physiological stressors that initially predispose living but weakened trees to attack by ambrosia beetles could also predispose the trees to infection by the symbiotic and auxiliary fungi after attacks are initiated [6,64]. Trees could conceivably become even more attractive due to the emission of ethanol and other key volatiles from cankerous tissue and wilting foliage following infection. However, the role of symbiotic and auxiliary fungi on host interactions and colonization success of X. germanus, X. crassiusculus, and numerous other non-native ambrosia beetles is not well characterized . Ambrosiella hartigii Batra is reportedly the symbiotic fungus of X. germanus, but Cerarocystis ulmi (Buisman), Fusarium lateritium Nees, Fusarium solani, and Fusarium oxysporum Schlechter ex Fries have been isolated from their galleries and cankers [65–69]. Ambrosiella roeperi sp. nov. has recently been identified as the mycangial symbiont of X. crassiusculus .
In addition to preferentially attacking flood-stressed trees, results from our current study also documented a preference by X. germanus and X. crassiusculus for attacking stems over branches on flood-stressed C. florida. Our current study also appears to be the first recovery of A. maiche from a host tree in North America  and demonstrated stems over branches on living but weakened trees were preferentially attacked. Similarly, more than 90% of ambrosia beetle attacks occurred on the main stem of C. florida trees growing at an ornamental nursery , and the majority of attacks by X. germanus and X. crassiusculus occurred less than 30 cm above the ground on stems of chestnut Castanea mollissima Blume . Beetle preferences for certain tree parts may be due to variations in anatomy and physiology between stems and branches [73,74] or represent a strategy to minimize interspecific competition among ambrosia beetles. Attacks on stems could have a more deleterious impact on whole-tree physiology than attacks on the branches due to disruption of the transpiration stream within xylem vessels by symbiotic or auxiliary fungal colonization and/or host hypersensitive responses to infection.
Although not always consistent , higher amounts of ethanol have been detected from flood stress intolerant tree species compared to tolerant species [44,45]. Cornus florida is generally recognized as intolerant of flooding [49,51], and flood-stressed C. florida deployed in Ohio and Virginia were associated with the highest concentrations of ethanol and cumulative ambrosia beetles as part of their respective experiments. However, flood-stressed C. canadensis deployed in Ohio and P. serrulata deployed in Virginia were associated with relatively high ethanol concentrations, but few ambrosia beetle attacks. This discrepancy may be attributed to interspecific variability in metabolic responses [37,75] and suggests vascular tissue concentrations of ethanol may not always reflect emission rates from the epidermis. Monitoring the emission of ethanol and other stress-induced volatiles from the epidermis using SPME-GC-MS  may be an improved approach for characterizing the role of host-derived volatiles during host selection by opportunistic ambrosia beetles. Differences in tissue processing/handling and sampling locations, and personnel may have resulted in intraspecific variability in ethanol concentrations, particularly for C. florida, C. canadensis, and P. serrulata. However, we limited our comparisons of ethanol concentrations among trees to only those deployed within a specific location for the purposes of our current study.
Overall, our current study indicates X. germanus and X. crassiusculus established in non-native habitats are early colonizers of living but weakened trees. This strategy may be contributing to their invasion success and proliferation. Healthy hosts may be opportunistically attacked by X. germanus and X. crassiusculus, but with limited success at colonization. Our current study also indicates the non-native A. maiche preferentially attacks living but weakened trees. Thus, other non-native ambrosia beetles that use stress-induced ethanol for locating vulnerable trees, and exhibit a capability for attacking a wide range of tree genera, may use a similar strategy.
We thank Jenny Barnett (USDA-ARS), Betsy Anderson (USDA-ARS), Leslie Morris (USDA-ARS), Hélène Doughty (VT-HRAREC), Elizabeth Barekzi (VT-HRAREC), and Elizabeth Gliem (VT-HRAREC) for technical assistance. We also thank anonymous reviewers for useful comments. Mention of proprietary products or companies does not imply any endorsement or preferential treatment by the USDA-Agricultural Research Service.
Conceived and designed the experiments: CMR PBS SDF JC MER. Performed the experiments: CMR PBS. Analyzed the data: CMR PBS SDF JC MER. Contributed reagents/materials/analysis tools: CMR PBS MER. Wrote the paper: CMR PBS SDF JC MER.
- 1. Haack RA. Exotic bark- and wood-boring Coleoptera in the United States: recent establishments and interceptions. Can J Forest Res. 2006;36: 269–288.
- 2. Rabaglia RJ, Dole SA, Cognato AI. Review of American Xyleborina (Coleoptera: Curculionidae: Scolytinae) occurring north of Mexico, with an illustrated key. Ann Entomol Soc America. 2006;99: 1034–1056.
- 3. Rabaglia R, Duerr D, Acciavatti R, Ragenovich I. Early detection and rapid response for non-native bark and ambrosia beetles. US Department of Agriculture Forest Service, Forest Health Protection. 2008.
- 4. Hulcr J, Dunn RR. The sudden emergence of pathogenicity in insect–fungus symbioses threatens naive forest ecosystems. Proc Royal Soc B: Biol Sci. 2011. https://doi.org/10.1098/rspb.2011.1130
- 5. Haack RA, Rabaglia RJ. Exotic bark and ambrosia beetles in the USA: potential and current invaders. In: Peña JE, editor. Potential Invasive Pests of Agricultural Crops. United Kingdom: CABI; 2013. pp. 48–74.
- 6. Ploetz RC, Hulcr J, Wingfield MJ, De Beer ZW. Destructive tree diseases associated with ambrosia and bark beetles: black swan events in tree pathology? Plant Disease. 2013;97: 856–872.
- 7. Brockerhoff EG, Bain J, Kimberley M, Knížek M. Interception frequency of exotic bark and ambrosia beetles (Coleoptera: Scolytinae) and relationship with establishment in New Zealand and worldwide. Can J For Res. 2006;36: 289–298.
- 8. Jordal B, Beaver RA, Kirkendall LR. Breaking taboos in the tropics: incest promotes colonization by wood-boring beetles. Global Ecol Biogeogr. 2001;10: 345–357.
- 9. Ranger CM, Tobin PC, Reding ME. Ubiquitous volatile compound facilitates efficient host location by a non-native ambrosia beetle. Biol Inv. 2015;17: 675–686.
- 10. Haack RA, Rabaglia RJ, Peña JE. Exotic bark and ambrosia beetles in the USA: potential and current invaders. In Peña J, editor. Potential invasive pests of agricultural crops. CAB International, Wallingford; 2013. pp. 48–74.
- 11. Grégoire JC, Piel F, De Proft M., Gilbert M. Spatial distribution of ambrosia-beetle catches: a possibly useful knowledge to improve mass-trapping. Int Pest Manag Rev. 2001;6: 237–242.
- 12. Henin JM, Versteirt V. Abundance and distribution of Xylosandrus germanus (Blandford 1894) (Coleoptera, Scolytidae) in Belgium: new observations and an attempt to outline its range. J Pest Sci. 2004;77: 57–63.
- 13. Bouget C, Noblecourt T. Short‐term development of ambrosia and bark beetle assemblages following a windstorm in French broadleaved temperate forests. J Appl Entomol. 2005;129: 300–310.
- 14. Wulf A, Pehl L. Forest health and forest protection 2004 in the Federal Republic of Germany-summary and preview 2005. Forst und Holz. 2005;60: 145–149.
- 15. CABI/EPPO [Internet]. Xylosandrus germanus (black timber bark beetle). Wallingford, UK: CABI. c2008 - [updated 2015 Jan 20; cited 2015 Mar 11]. Available from: http://www.cabi.org/isc/datasheet/57237
- 16. Galko J, Nikolov C, Kimoto T, Kunca A, Gubka A, Vakula J, et al. Attraction of ambrosia beetles to ethanol baited traps in a Slovakian oak forest. Biologia. 2014;69: 1376–1383.
- 17. Pennacchio F, Roversi PF, Francardi V, Gatti E. Xylosandrus crassiusculus (Motschulsky) a bark beetle new to Europe (Coleoptera Scolytidae). Redia. 2003;86: 77–80.
- 18. CABI/EPPO [Internet]. Xylosandrus crassiusculus (Asian ambrosia beetle). Wallingford, UK: CABI. c2009 - [updated 2015 Jan 20; cited 2015 Mar 11]. Available from: http://www.cabi.org/isc/datasheet/57235
- 19. Weber BC, McPherson JE. World list of host plants of Xylosandrus germanus (Blandford) (Coleoptera: Scolytidae). Coleop Bull. 1983;37: 114–134.
- 20. Schedl KE. Scolytidae und Platypodidae Afrikas. II. Revista de Entomologia Moçambique. 1962;5: 1–594.
- 21. Weber BC, McPherson JE. Notes: Attack on black walnut trees by the ambrosia beetle Xylosandrus germanus (Coleoptera: Scolytidae). For Sci. 1984;30: 864–870.
- 22. Weber BC, McPherson JE. Relation between attack by Xylosandrus germanus (Coleoptera: Scolytidae) and disease symptoms in black walnut. Can Entomol. 1985;117: 1275–1277.
- 23. Atkinson TH, Foltz JL, Wilkinson RC, Mizell RF. Granulate ambrosia beetle, Xylosandrus crassiusculus (Motschulsky) (Insecta: Coleoptera: Curculionidae: Scolytinae). Florida Department of Agriculture and Consumer Services, Division of Plant Industry, Entomology Circular. 1988;310. Available from: http://edistt.ifas.ufl.edu/pdffiles/IN/IN28800.pdf
- 24. Sreedharan K, Balakrishnan MM, Samuel SD, Bhat PK. A note on the association of wood boring beetles and a fungus with the death of silver oak trees on coffee plantations. J Coffee Res. 1991;21: 145–148.
- 25. Oliver JB, Mannion CM. Ambrosia beetle (Coleoptera: Scolytidae) species attacking chestnut and captured in ethanol-baited traps in middle Tennessee. Environ Entomol. 2001;30: 909–918.
- 26. Keshavareddy G, Verghese A, Kumar HM. Spatial distribution of the damage of shot hole borer (Coleoptera: Scolytidae) on grapes. Current Biotica. 2008;2: 18–31.
- 27. Kühnholz S, Borden JH, Uzunovic A. Secondary ambrosia beetles in apparently healthy trees: Adaptations, potential causes and suggested research. Int Pest Manag Rev. 2001;6: 209–219.
- 28. Ranger CM, Reding ME, Persad AB, Herms DA. Ability of stress-related volatiles to attract and induce attacks by Xylosandrus germanus and other ambrosia beetles (Coleoptera: Curculionidae, Scolytinae). Agric For Entomol. 2010;12: 177–185.
- 29. Ranger CM, Reding ME, Gandhi KJ, Oliver JB, Schultz PB, Cañas L, et al. Species dependent influence of (−)-α-pinene on attraction of ambrosia beetles (Coleoptera: Curculionidae: Scolytinae) to ethanol-baited traps in nursery agroecosystems. J Econ Entomol. 2011;104: 574–579. pmid:21510207
- 30. Ranger CM, Gorzlancyk AM, Addesso KM, Oliver JB, Reding ME, Schultz PB, et al. Conophthorin enhances the electroantennogram and field behavioural response of Xylosandrus germanus (Coleoptera: Curculionidae) to ethanol. Agric For Entomol. 2014;16: 327–334.
- 31. Ranger CM, Reding ME, Schultz PB, Oliver JB. Ambrosia beetle (Coleoptera: Curculionidae) responses to volatile emissions associated with ethanol-injected Magnolia virginiana L. Environ Entomol. 2012;41: 636–647. pmid:22732622
- 32. Ranger CM, Tobin PC, Reding ME, Bray AM, Oliver JB, Schultz PB, et al. Interruption of semiochemical-based attraction of ambrosia beetles to ethanol-baited traps and ethanol-injected trap trees by verbenone. Environ Entomol. 2013;42: 539–547. pmid:23726063
- 33. Ranger CM, Reding ME, Oliver JB, Schultz PB, Moyseenko JJ, Youssef N. Comparative efficacy of botanical formulations for managing ambrosia beetles and their corresponding mass spectral characterization. J Econ Entomol. 2011;104: 1665–1674. pmid:22066197
- 34. Borden JH, Brownlee RG, Silverstein RM. Sex pheromone of Trypodendron lineatum (Coleoptera: Scolytidae): production, bio-assay, and partial isolation. Can Entomol. 1968;100: 629–636.
- 35. Renwick JAA, Vite JP, Billings RF. Aggregation pheromones in the ambrosia beetle Platypus flavicornis. Naturwissenschaften. 1977;64: 226–226.
- 36. Kimmerer TW, Kozlowski TT. Ethylene, ethane, acetaldehyde, and ethanol production by plants under stress. Plant Physiol. 1982;69: 840–847. pmid:16662306
- 37. Kimmerer TW, MacDonald RC. Acetaldehyde and ethanol biosynthesis in leaves of plants. Plant Physiol. 1987;84: 1204–1209. pmid:16665585
- 38. MacDonald RC, Kimmerer TW. Ethanol in the stems of trees. Physiol Plant. 1991;82: 582–588.
- 39. Joseph G, Kelsey RG. Ethanol synthesis and water relations of flooded Pseudotsuga menziesii (Mirb.) Franco (Douglas-fir) seedlings under controlled conditions. Int J Plant Sci. 1997;158: 844–850.
- 40. Kelsey RG. Chemical indicators of stress in trees: their ecological significance and implication for forestry in Eastern Oregon and Washington. Northwest Sci. 2001; 75: 70–76.
- 41. Drew MC. Oxygen deficiency and root metabolism: injury and acclimation under hypoxia and anoxia. Ann Rev Plant Biol. 1997;48: 223–250.
- 42. Tadage M, Dupuis I, Kuhlemeier C. Ethanolic fermentation: new functions for an old pathway. Trends Plant Sci. 1999;4: 320–325. pmid:10431222
- 43. Ranger CM, Reding ME, Schultz PB, Oliver JB. Influence of flood-stress on ambrosia beetle host selection and implications for their management in a changing climate. Agric For Entomol. 2013;15: 56–64.
- 44. Joly CA, Crawford RMM. Variation in tolerance and metabolic responses to flooding in some tropical trees. J Exp Bot. 1982;33: 799–809.
- 45. Kennedy RA, Rumpho ME, Fox TC. Anaerobic metabolism in plants. Plant Physiol. 1992;100: 1–6. pmid:16652929
- 46. Heidenreich E. Oekologische Bedingungen für Primärbefall durch “Xylosandrus germanus.” Z. Angew. Entomol. 1964;54: 131–140.
- 47. Kovach J, Gorsuch CS. Survey of ambrosia beetle species infesting South Carolina peach orchards and a taxonomic key for the most common species. J. Agric. Entomol. 1985;2: 238–247.
- 48. Solomon JD. Guide to insect borers in North American broadleaf trees and shrubs. AH-706. Washington: Agriculture Handbook. 1995.
- 49. Broadfoot WM, Williston HL. Flooding effects on southern forests. J Forestry. 1973;71: 584–587.
- 50. Gilman EF, DG Watson. Styrax japonicus (Japanese snowbell) Fact Sheet ST-605. Environmental Horticulture Department, Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences, University of Florida. c1993 - [updated 2006 Dec; cited 2015 Mar 11]. Available from: http://edis.ifas.ufl.edu/st605
- 51. Day SD, Seiler JR, Persaud N. A comparison of root growth dynamics of silver maple and flowering dogwood in compacted soil at differing soil water contents. Tree Physiol. 2000;20: 257–263. pmid:12651462
- 52. Ranney TG. Differential tolerance of eleven Prunus taxa to root zone flooding. J Environ Hort. 1994;12: 138–138.
- 53. Hall RBW, Harcombe PA. Flooding alters apparent position of floodplain saplings on a light gradient. Ecol. 1998;79: 847–855.
- 54. Sigma-Aldrich. SPME fiber assembly Carboxen/Polydimethylsiloxane (CAR/PDMS). 2015. Available: http://www.sigmaaldrich.com/catalog/product/supelco/57344u?lang=en®ion=US
- 55. French JR, Roeper RA. 1972. Interactions of the ambrosia beetle, Xyleborus dispar (Coleoptera: Scolytidae), with its symbiotic fungus Ambrosiella hartigii (Fungi Imperfecti). Can. Entomol. 1972;104: 1635–1641.
- 56. SAS Institute. PROC User’s Manual. Version 6. North Carolina: SAS Institute; 2001.
- 57. Dunn JP, Kimmerer TW, Nordin GL. The role of host tree condition in attack of white oaks by the twolined chestnut borer, Agrilus bilineatus (Weber) (Coleoptera: Buprestidae). Oecologia. 1986;70: 596–600.
- 58. Borden JH. Semiochemicals and bark beetle populations: exploitation of natural phenomena by pest management strategies. Hol Ecol. 1989;12: 501–510.
- 59. Laidlaw WG, Prenzel BG, Reid ML, Fabris S, Wieser H. Comparison of the efficacy of pheromone-baited traps, pheromone-baited trees, and felled trees for the control of Dendroctonus pseudotsugae (Coleoptera: Scolytidae). Environ Entomol. 2003;32: 477–483.
- 60. Reding ME, Oliver J, Schultz P, Ranger CM, Youssef N. Ethanol injection of ornamental trees facilitates testing insecticide efficacy against ambrosia beetles (Coleoptera: Curculionidae: Scolytinae). J Econ Entomol. 2013;106: 289–298. pmid:23448043
- 61. Biedermann PH, Klepzig KD, Taborsky M. Fungus cultivation by ambrosia beetles: behavior and laboratory breeding success in three xyleborine species. Environ Entomol. 2009;38: 1096–1105. pmid:19689888
- 62. La Spina S, De Cannière C, Dekri A, Grégoire JC. Frost increases beech susceptibility to scolytine ambrosia beetles. Agric For Entomol. 2013;15: 157–167.
- 63. Hanks LM. Influence of the larval host plant on reproductive strategies of cerambycid beetles. Ann Rev Entomol. 1999;44: 483–505.
- 64. Schoeneweiss DF. Predisposition, stress, and plant disease. Ann Rev Phytopathol. 1975;13: 193–211.
- 65. Weber BC, McPherson JE. The ambrosia fungus of Xylosandrus germanus (Coleoptera: Scolytidae). Can Entomol. 1984;116: 281–283.
- 66. Dochinger LS, Seliskar CE. Fusarium canker found on yellow-poplar. J Forestry. 1962;60: 331–333.
- 67. Norris DM. The mutualistic fungi of Xyleborini beetles. In Batra LR, editor. Insect fungus symbiosis: nutrition, mutualism, and commensalism. New Jersey: Allanheld, Osmun; 1979. pp. 53–63.
- 68. Kessler K. Jr. An apparent symbiosis between Fusarium fungi and ambrosia beetles causes canker on black walnut stems. Plant Dis. Rep. 1974;58: 1044–1047.
- 69. Anderson RL, Hoffard WH. Fusarium canker ambrosia beetle complex on tulip poplar in Ohio. Plant Dis Rep. 1978;62: 751.
- 70. Harrington TC, McNew D, Mayers C, Fraedrich SW, Reed SE. Ambrosiella roeperi sp. nov. is the mycangial symbiont of the granulate ambrosia beetle, Xylosandrus crassiusculus. Mycologia. 2014;106: 835–845.
- 71. Rabaglia RJ, Vandenberg NJ, Acciavatti RE. First records of Anisandrus maiche Stark (Coleoptera: Curculionidae: Scolytinae) from North America. Zootaxa. 2009;2137: 23–28.
- 72. Reding M, Oliver J, Schultz P, Ranger C. Monitoring flight activity of ambrosia beetles in ornamental nurseries with ethanol-baited traps: influence of trap height on captures. J Environ Hort. 2010;28: 85–90.
- 73. Sprugel DG, Hinckley TM, Schaap W. The theory and practice of branch autonomy. Ann Rev Ecol Syst. 1991;22: 309–334.
- 74. Damesin C, Ceschia E, Le Goff N, Ottorini JM, Dufrêne E. Stem and branch respiration of beech: from tree measurements to estimations at the stand level. New Phytol. 2002;153: 159–172.
- 75. Tripepi RR, Mitchell CA. Metabolic response of river birch and European birch roots to hypoxia. Plant Physiol. 1984;76: 31–35. pmid:16663817