Densities of Agrilus auroguttatus and Other Borers in California and Arizona Oaks

We investigated within-tree population density of a new invasive species in southern California, the goldspotted oak borer, Agrilus auroguttatus Schaeffer (Coleoptera: Buprestidae), with respect to host species and the community of other borers present. We measured emergence hole densities of A. auroguttatus and other borers on the lower stem (bole) of naïve oaks at 18 sites in southern California and on co-evolved oaks at seven sites in southeastern Arizona. We sampled recently dead oaks in an effort to quantify the community of primary and secondary borers associated with mortality—species that were likely to interact with A. auroguttatus. Red oaks (Section Lobatae) produced greater densities of A. auroguttatus than white oaks (Section Quercus). On red oaks, A. auroguttatus significantly outnumbered native borers in California (mean ± SE of 9.6 ± 0.7 versus 4.5 ± 0.6 emergence holes per 0.09 m2 of bark surface), yet this was not the case in Arizona (0.9 ± 0.2 versus 1.1 ± 0.2 emergence holes per 0.09 m2). In California, a species that is taxonomically intermediate between red and white oaks, Quercus chrysolepis (Section Protobalanus), exhibited similar A. auroguttatus emergence densities compared with a co-occurring red oak, Q. kelloggii. As an invasive species in California, A. auroguttatus may affect the community of native borers (mainly Buprestidae and Cerambycidae) that feed on the lower boles of oaks, although it remains unclear whether its impact will be positive or negative.


Introduction
The goldspotted oak borer, Agrilus auroguttatus Schaeffer (Coleoptera: Buprestidae), is a new pest of oaks in southern California (CA) [1]. It was first collected in CA in 2004 [2]. Collection records throughout the 20th century indicate that A. auroguttatus is native to mountain ranges in southeastern Arizona (AZ) [3]. Current genetic evidence is consistent with AZ populations as the probable source for CA populations [4]. Prior to its discovery in CA, little was known about this beetle's biology and ecology. Knowledge of the effect of host species [5,6] and interactions with the community of other borers present within its native and introduced ranges are important for understanding the invasion biology of A. auroguttatus.
In their new ecosystems, invasive species populations can reach damaging levels because new habitats may lack important natural enemies (e.g., enemy release, [7]), or co-evolved resistance mechanisms among new host plants (e.g., biotic resistance, [8]). Consequently, invasive insects can have direct effects on populations of these naï ve host plants and indirect effects on native populations of specialist herbivores that may share hosts with the invader [9][10][11]. Comparing the biology and ecology of invasive species in their new habitats with those in their native ranges can further understanding of invasion biology and may help to identify possible management techniques for invasive pests.
Direct comparisons of A. auroguttatus population densities on species from the red (Section Lobatae) and white oak (Section Quercus) groups in CA or AZ have not been made among individual trees of similar physiological status. Such comparisons may confirm whether red oaks are more preferable or suitable hosts than white oaks for A. auroguttatus. Thus far, extensive mortality (approx. 25,000 dead trees) in southern CA has occurred primarily among two species in the red oak group: CA black oak, Quercus kelloggii Newb., and coast live oak, Q. agrifolia Né e [12]. Known host species of A. auroguttatus in AZ (also in Section Lobatae) include Emory oak, Q. emoryi Torrey, and silverleaf oak, Q. hypoleucoides A. Camus [3]. In CA, A. auroguttatus also colonizes the taxonomically intermediate species canyon live oak, Q. chrysolepis Liebm. (Section Protobalanus) and Engelmann oak, Q. engelmannii Greene, from the white oak group, although the latter is colonized at very low densities [13,14]. There have been no observations of A. auroguttatus on Arizona white oak, Q. arizonica L. (Section Quercus), which co-occurs with its red oak hosts in AZ [3,13].
Agrilus auroguttatus is the primary biotic agent causing tree mortality in southern CA oak forests [13], but its influence on native borer populations in the main stem of oaks remains largely unknown. Comparisons of A. auroguttatus population densities to densities of other borers may be a first step towards determining borer community interactions in oak ecosystems. A suite of native oak insects feed in subcortical tissues within the main stem (bole) of southern CA oaks [15][16][17]. Records of oak-feeding insects in southeastern AZ are more limited [16,18].
Our objective was to investigate within-tree densities of A. auroguttatus in its introduced and native ranges with respect to host species and the community of other borers present. We compared the density of emergence holes from A. auroguttatus and other borers on different species of recently dead oaks among several sites in CA and AZ. Our study focused on the community of borers in the lower bole because A. auroguttatus occurs most frequently in this portion of the stem [19]. We hypothesized that: (1) species in the red oak group would be better hosts for A. auroguttatus, and thus produce higher emergence densities than co-occurring white oak species; and (2) A. auroguttatus would outnumber other borers in its introduced range (CA), but not in its native range (AZ).

Study Sites: Southern California
Eighteen study sites were selected in the current zone of A. auroguttatus infestation within the Cleveland National Forest and one site was selected outside the currently infested zone (Warner Springs) in southern CA (Table 1; infestation zone as of 2010 in [3]). Sites were distributed approximately evenly throughout the public forested lands of the Descanso Ranger District to evaluate variation in borer populations throughout the National Forest. Ten sites were Q. agrifolia-dominant oak woodlands (900-1500 m in elevation) and the remaining nine sites were mixed Q. kelloggii-pine forests (mostly Pinus jeffreyi Grev. & Balf.; 1500-1800 m in elevation). We searched for, but could not find, sites where both of these red oak species occurred and A. auroguttatus was present. Thus, statistical comparisons of borer density between these two host species were not possible because the effect of host species on A. auroguttatus density was confounded with the effects of location and elevation. Comparisons of A. auroguttatus emergence holes could only be made between host species pairs at two sites (Roberts' Ranch and Thing Valley), where two red oak/non-red oak species pairs occurred (Table 1).
Climate in the Descanso Ranger District is Mediterranean with hot summers (mean August and January maximum temperatures: 30 °C and 14 °C , respectively [20]) [21]. The majority of precipitation occurs in the winter (total annual precipitation is 53 cm [20]). Soils are rocky, well-drained sandy loams above decomposed granitic bedrock [21].

Study Sites: Southeastern Arizona
Seven study sites were selected within the Coronado National Forest in southeastern AZ where oak mortality was identified during previous surveys [13]. Sites with Q. emoryi were oak woodlands (1400-1700 m elevation) and sites with Q. hypoleuocides and Q. arizonica were mixed oak-pine forests that also contained Juniperus deppeana, Pinus spp., Platanus racemosa Nutt., and Acer spp.
(1500-1750 m elevation) ( Table 1). It was difficult to find sites in southeastern AZ with both co-occurring oak species and with evidence of A. auroguttatus infestation. Only one study site, Cochise Stronghold Campground, contained all three oak species. Borer densities were also compared between Q. arizonica and Q. hypoleucoides at Carr and Miller Canyons. At all other sites, borer emergence holes were counted on only one oak species.  [22]. Borer emergence holes were counted on three dead Q. hypoleucoides before the fire and seven after the fire at Miller Canyon as well as four Q. emoryi before the fire and six after the fire at Pinery Canyon. Density of emergence holes did not differ significantly between dead trees sampled before vs. after the fire at either site (pooled variance t-tests, A. auroguttatus: t = 1. 23  Climate in the Coronado National Forest in southeastern Arizona is semi-arid. Total annual precipitation varies throughout these mountain ranges from 34-58 cm, and mean August and January maximum temperatures are 32 °C and 16 °C , respectively [20]. Soils are rocky and classified mostly as Aridisols, although some are Mollisols or Entisols [23].

Evaluation of Borer Density
Cumulative borer emergence was estimated on the lower boles (≤1.52 m) of the first 10, or as many as we could find if fewer than 10, recently dead (died within the last 1-3 y, borer emergence holes were still clearly defined, and at least half of the bark on the lower bole remained) oaks encountered at all CA and AZ sites. A total of 91 Q. kelloggii (nine sites); 99 Q. agrifolia (10 sites); 10 Q. chrysolepis (one site); 7 Q. engelmannii (one site); 40 Q. emoryi (four sites); 33 Q. hypoleucoides (four sites); and 23 Q. arizonica (three sites) were sampled for the entire study. We did not sample living trees because our goal was to obtain a final measure of borer density from trees that were not likely to produce additional generations of borers. While some borer species may feed on dead and decaying wood, it was assumed that including only standing, recently-dead trees would limit the sample to those species that were primary or secondary mortality agents feeding on dying vascular tissue in weakened or stressed hosts. These species were most likely to interact with A. auroguttatus (as opposed to species that feed on dead or decaying wood).
All emergence holes were counted within a 0.09 m 2 (1 ft 2 ) sampling window (cardboard frame) at three arbitrary locations on the lower bole (≤1.52 m) of each oak tree, where A. auroguttatus emergence holes are dispersed randomly and occur most frequently [19]. Sampling efforts were focused on the lower bole because the goal was to estimate densities of A. auroguttatus where it was most likely to be present in relation to other borers.
Emergence (and by assumption, development) by A. auroguttatus or other borers was determined according to emergence hole size and shape [14]. Agrilus auroguttatus emergence holes were D-shaped and about 4 mm wide [14,24]. The size and shape of these holes are uniquely diagnostic for this species among the community of subcortical insects on California oaks. Emergence holes from flatheaded borers (Buprestidae) other than A. auroguttatus (Chrysobothris spp.) were oblong or crescent-shaped and 5-13 mm wide; roundheaded borer (Cerambycidae) (Rosalia funebris (Motschulsky), Xylotrechus nauticus (Mannerheim) or Neoclytus conjunctus (LeConte)) emergence holes were round or oblong and 6-10 mm wide; false powderpost beetles (Bostrichidae) (Scobicia declivis (LeConte)) entrance and emergence holes were perfectly round and 4 mm in diameter (the same size as A. auroguttatus only perfectly round); bark and ambrosia beetles (Curculionidae: Scolytinae) (Pseudopityophthorus spp. or Monarthrum spp.) entrance and emergence holes were also round, but ≤2 mm in diameter; clearwing borer (Lepidoptera: Sessiidae) (Synanthedon resplendens (Edwards)) emergence holes were round and 5-6 mm in diameter and found exclusively in bark cracks near deteriorated bark and phloem; and carpenterworm (Lepidoptera: Cossidae) (Prionoxystus robiniae Peck) emergence holes were oval often with jagged edges and 10-20 mm wide [15][16][17][18]25,26]. Emergence holes were categorized as -unknown‖ if they could not be identified to insect family or if bark had deteriorated at all since the insect emerged to the extent that the hole's original shape was somewhat unclear. There was overlap in the shape and size of emergence holes for Chrysobothris spp. and Cerambycidae (particularly X. nauticus), so counts were grouped from both taxa (bup + cer) for analyses.

Data Analyses
All statistical analyses were conducted within the R statistical environment, version 2.15.0 [27]. We used pooled variance t-tests or analyses of variance (ANOVA) to test for the effect of oak species on density of A. auroguttatus emergence holes. We used ANOVA to test for the effect of borer group on density of emergence holes followed by Tukey's HSD ad-hoc tests, applied separately to each oak species at CA sites where two oak species co-occurred.
We applied Pearson's correlations to search for a relationship between density of emergence holes of A. auroguttatus and all other borer groups combined (on red oak species only), performed for each oak species (all sites combined), and separately by site. We used mixed effect linear models with restricted maximum likelihood estimation to compare number of A. auroguttatus emergence holes against that of all other borer groups combined, tested separately by host species for red oaks. Borer type was a fixed effect and site was a random effect.
For all linear models, model checks included the Shapiro-Wilk normality test [28], standardized residual plots, and normal probability plots. Data were transformed to meet model assumptions (adding 0.5 to original values and then applying the Box-Cox method [29] to determine appropriate transformations, used for all response data except for t-tests in Figure 1A,B). If data could not be transformed to meet parametric assumptions (for effect of borer group on emergence holes for Q. kelloggii at Thing Valley and Q. agrifolia at Warner Springs, Figure 2A,E, respectively), we applied ANOVA to ranks. We present results from ANOVA on ranks (Figure 2A,E) as means with standard errors for visual comparisons with corresponding parametric models of different oak species. Statistical significance was set at p < 0.05, all means in graphs are untransformed values, and errors are one standard error from the mean.

Q. hypoleucoides or Q. arizonica at Cochise Stronghold. At Carr and Miller Canyons in AZ,
A. auroguttatus density did not differ significantly between sites (F = 0.25; df = 1, 36; p = 0.624), but differed significantly by oak species (F = 21.66; df = 1, 36; p < 0.001; Figure 1D). A. auroguttatus density on Q. hypoluecoides was very low at these two sites, but was significantly greater than on Q. arizonica where no holes were found.  pairs (A,B) and at AZ sites with more than one co-occurring host species (C,D). Significant differences between species are denoted by *. Sample sizes are in parentheses following species names. Oaks in Section Lobatae are red symbols; Section Protobalanus are gray symbols; and Section Quercus are white symbols.

Figure 2.
Mean densities of emergence holes on the lower boles of dead oaks by borer group, tested separately by host species at CA sites containing host species pairs (A-D) and at Warner Springs, a site outside the current zone of A. auroguttatus infestation in CA (E). Letters above bars indicate significant differences according to Tukey's HSD; sample sizes are the same as corresponding species and sites for Figure 1; and n = 10 trees (E). Oaks in Section Lobatae are red circles; Section Protobalanus are gray circles; and Section Quercus are white circles. X-axis categories are as follows: gsob = A. auroguttatus; bup + cer = other Buprestidae combined with Cerambycidae; bos = Bostrichidae; cur = Curculionidae (Scolytinae); sess = Sessiidae; and uk = unknown.
At sites in CA where pairs of oak species co-occurred, A. auroguttatus density was significantly greater than the density of other borer groups on the lower boles of Q. kelloggii (F = 27.47; df = 4, 45; p < 0.001; Figure 2A) and Q. agrifolia within the current zone of infestation (F = 26.84; df = 5, 54; p < 0.001; Figure 2C). Bup + cer density was also relatively high on the lower bole of CA oaks. Bup + cer and A. auroguttatus densities were greater than density of other borer groups on Q. chrysolepis (F = 30.51; df = 4, 45; p < 0.001; Figure 2B). Bup + cer density was greater than densities of all other borers on Q. agrifolia outside the current zone of A. auroguttatus infestation (F = 36.65; df = 4, 45; p < 0.001; Figure 2E). Emergence hole density was very low (mean of 1.0 per 0.09 m 2 of bark, among all borer groups), and did not vary significantly by borer group on Q. engelmannii (F = 1.76; df = 3, 24; p = 0.182; Figure 2D).
On the lower bole, A. auroguttatus emergence was greater than that of all other borer species combined from red oak species across several sites in CA (F = 72.49; df = 1, 170; p < 0.001 and F = 70.46; df = 1, 172; p < 0.001; Figure 3A for Q. agrifolia and Q. kelloggii, respectively). Agrilus auroguttatus densities were not significantly different from those of all other borer species combined from red oak species in AZ (F = 2.85; df = 1, 82; p = 0.095 and F = 1.71; df = 1, 61; p = 0.196; Figure 3B for Q. emoryi and Q. hypoleucoides, respectively). In general, emergence hole densities of A. auroguttatus were not significantly correlated with those of all other borer species combined among red oaks or at different sites. There were, however, three exceptions-emergence hole densities of A. auroguttatus and other borers were significantly positively correlated on Q. hypoleucoides among all sites (r p = 0.521; p = 0.003; not shown), on Q. kelloggii at Horse Heaven in CA (r p = 0.849; p = 0.002; not shown) and on Q. hypoleucoides at Miller Canyon in AZ (r p = 0.995; p < 0.001; not shown).

Discussion
The scarcity of sites in southern CA and southeastern AZ oak forests with co-occurring oak species where A. auroguttatus was also present prevented some direct comparisons among oak species.
Although few direct comparisons were made, this study documents patterns of emergence from natural populations of A. auroguttatus that suggest red oaks are either more preferable or more suitable hosts than white oaks (i.e., Q. agrifolia versus Q. engelmannii and Q. hypoleucoides versus Q. arizonica; Figure 1A,D). A red oak, Q. kelloggii, may be equally preferable/suitable when compared with taxonomically intermediate Q. chrysolepis ( Figure 1B). On native hosts at one site (Cochise Stronghold; Figure 1C), densities of A. auroguttatus on red oaks were so low that they were not significantly different from a white oak, Q. arizonica. At this site, neither Q. hypoleucoides nor Q. arizonica showed evidence of A. auroguttatus colonization. The results of this study are consistent with previous observations, which report more frequent evidence of A. auroguttatus colonization on red oaks than white oaks [3,13].
At the individual tree level, presence of A. auroguttatus in CA oak forests may result in lower densities of native buprestids and cerambycids on the lower boles of dying Q. agrifolia. At sites in CA with co-occurring oak species, A. auroguttatus was five times more abundant than all other borer groups on respective red oaks at each site (Q. agrifolia and Q. kelloggii; Figure 2A,C). Other buprestids and cerambycids were also relatively abundant on three of the four oak species examined, and more abundant than all other borer groups on Q. agrifolia in the absence of A. auroguttatus ( Figure 2E).
This study focused only on the lower portion of tree boles. Community interactions elsewhere in the bole, including species partitioning with height, were not measured and may be important. For example, in eastern North American oaks attacked by A. bilineatus, bark and ambrosia beetles were more prevalent in upper portions of boles [30]. In that study, buprestid and cerambycid densities remained relatively constant with bole height or were greatest on the lower bole and scolytine densities were greatest at ~12.5 m above ground [30]. It is possible that A. auroguttatus may compete with native Buprestidae and Cerambycidae on the lower boles of dying Q. agrifolia, although data collected at different heights from trees of varying physiological conditions and from more sites outside the currently infested zone would be necessary to evaluate this hypothesis.
Across several sites throughout its invaded range in CA, and on both red oaks, A. auroguttatus outnumbered all species of native borers combined, yet this was not the case across sites in AZ ( Figure 3). Across these sites, A. auroguttatus densities were similar to those of all other borer groups combined, which suggests that A. auroguttatus does not outnumber all other borer species in its native range as it does in its introduced range. Thus, although it is a primary mortality agent of red oaks in CA, it is more likely a secondary, though relatively frequent, mortality agent of red oaks in AZ oak forests.
The mean and range densities of A. auroguttatus that emerged from susceptible naï ve hosts varied somewhat from those reported in other studies of Agrilus species on their respective naï ve hosts ( Table 2). The range that we reported for A. auroguttatus on Q. agrifolia was influenced by one tree that contained 50 emergence holes per 0.09 m 2 , the next greatest density was only 29 emergence holes per 0.09 m 2 . Similar to our observations of A. auroguttatus on co-evolved hosts, emergence densities of other Agrilus species were much lower on co-evolved than naï ve hosts ( Table 2). Many factors undoubtedly varied among the other studies referenced in Table 2 and ours. However, in those studies, trees had died recently or were moribund, and emergence densities reported were likely a result of the cumulative number of Agrilus beetles produced by dying trees.

Conclusions
Sites with co-occurring oak species that also contained populations of A. auroguttatus were scarce, which made testing the role of host species in within-tree densities of A. auroguttatus difficult. Nonetheless, in a small number of white oaks surveyed in CA and AZ, few A. auroguttatus emergence holes were recorded, whereas at some locations, significantly higher emergence hole densities were recorded for various red oaks. As an invasive species in CA, A. auroguttatus may affect the community of native borers feeding in the lower boles of oaks, although it remains unclear whether its impact will be positive or negative. One possibility is that primary attack by A. auroguttatus could weaken trees and create more suitable hosts on the landscape for native borer populations. This does not seem to be the case, since we did not find much evidence of a positive correlation between emergence hole densities for A. auroguttatus and other borers among oak species or sites. Alternatively, by attacking apparently healthy trees and colonizing them for several successive generations, A. auroguttatus may out-compete native borers in dying trees. Both of these possibilities should be investigated further in CA and AZ oak forests.