Colonization success of a tree‐killing bark beetle: Geographic variation and mismatch with host preference

Abstract The preference–performance hypothesis (PPH) predicts that female insects maximize their fitness by ovipositing on hosts where their offspring perform the best. The preference–performance relationships in bark beetles are complex because before offspring development can occur in the phloem, adult bark beetles must first successfully invade host trees, and then construct galleries beneath the bark. Therefore, a positive correlation between host preference and successful colonization is necessary for the PPH in bark beetles to hold (i.e., the preference–colonization hypothesis in bark beetles). In this study, through field choice experiments, I investigated the successful colonization of the bark beetle, Polygraphus proximus, within four allopatrically distributed Abies species across a distinct biogeographic boundary in Japan. The results of this study showed that the biogeographic boundary did not limit the successful colonization by P. proximus. I observed that successful colonization was low in A. firma, despite it being an exotic species in the study sites and the most preferred at the study sites, indicating a mismatch between preference and colonization success. Additionally, I observed that A. sachalinensis had a high colonization success rate, even though it was the least preferred species at the study sites.

colonization is necessary for the PPH to hold true in bark beetles (i.e., the preference-colonization hypothesis in bark beetles).
Polygraphus proximus is a nonaggressive phloeophagous bark beetle that feeds on Far Eastern Abies species (Kerchev, 2014a;Koizumi, 1977;Nobuchi, 1966). The bark beetle is native to throughout Japan, northeastern China, Korea, and the southern Russian Far East (Kerchev, 2014a;Nobuchi, 1966), and it infests freshly cut logs and trees that have been weakened by fire, pathogens, typhoons, or defoliation during the endemic phase of their native range (Koizumi, 1977;Nobuchi, 1966). P. proximus has invaded European Russia and West Siberia for the past 10 years, causing mortality across large areas of fir forests (Baranchikov et al., 2010;Kerchev, 2014b). P. proximus also caused mass attacks and mortality events in Abies trees in its native range in Japan (Chiba et al., 2020;Takagi et al., 2018Takagi et al., , 2021Tokuda et al., 2008). They exhibit a bivoltine cycle in their native distribution (EFSA, 2020;EPPO, 2014). The emergence period and subsequent flight of the first generation of adults occur at the end of spring or at the beginning of summer (May-July), and the second generation emerges from the infested trees at the end of summer (August-September;EFSA, 2020;EPPO, 2014). The mating system of P. proximus is monogynous (Kerchev, 2014a;Köbayashi & Takagi, 2020). Although a male sex pheromone is suspected to play a role in mating, this role has not been confirmed yet (Kabe, 1959;Nobuchi, 1980). Five native Abies species are potential hosts of P. proximus in Japan ( Figure 1). The five Abies species in Japan are divided into three phylogenetic groups based on their genetic relatedness: (i) section Balsamea, which includes A. veitchii and A. sachalinensis, (ii) section Momi, which includes A. firma and A. homolepis, and (iii) section Amabilis, which includes A. mariesii (Suyama et al., 2000;Tsumura & Suyama, 1998). The five Abies species grow in allopatry, with their natural distribution limited in Japan by a distinct biogeographic boundary known as Blakiston's Line; A. sachalinensis is the only species native to Hokkaido, the northernmost island in Japan, and not to Honshu, the largest main island in Japan (Figure 1).
In contrast, the other four Abies species are native to Honshu and not Hokkaido. Thus, the Blakiston's Line has been recognized as a distinct north-south boundary between Hokkaido and Honshu, and the Tsugaru Strait acts as the geographic representation of Blakiston's Line and represents a physical barrier to the movement of many terrestrial species (Aizawa et al., 2007;Haba et al., 2008;Kubota et al., 2014;Suzuki et al., 2004;Tsuyama et al., 2014;Yokoyama & Goto, 2002).
A previous study determined the P. proximus host preference among four allopatrically distributed Abies species in Japan (i.e., A. firma, A. homolepis, A. sachalinensis, and A. veitchii), and demonstrated that the preference did not vary between Hokkaido and Honshu; P. proximus preferred A. veitchii and A. firma as its hosts, followed by A. homolepis, with A. sachalinensis then receiving the lowest preference on either side of the Tsugaru Strait (Takagi, 2022). However, the colonization success of P. proximus and the differences across biogeographic boundaries which separate host species have not been investigated.
Based on the host preference pattern of P. proximus and the positive correlation between host preference and successful colonization, I hypothesized that the rate of successful colonization among the Abies species would be similar to the host preference, and the geographic barrier separating the Abies species would not limit the successful colonization by P. proximus.
Therefore, this study aimed to determine the proportion of successful colonization of P. proximus across a biogeographic boundary in Japan, and whether a positive relationship was present between host preference, as shown in the previous study (Takagi, 2022), and colonization success.

| Study sites
To investigate the successful colonization of P. proximus on either side sachalinensis is the only Abies species native to the UTHF site. In the UTHF, A. veitchii was introduced from central Honshu and A. sachalinensis from Hokkaido in the 1960s. The annual rainfall and average annual temperature at both sites were similar (SRS: 1220.5 mm and F I G U R E 1 Natural distribution of five Abies species in Japan and the locations of the study sites. The natural distributions are based on Tsumura and Suyama (1998). Two tree species can be sympatrically distributed near their boundary areas.
proximus on A. veitchii were reported at both the SRS and UTHF sites (Takagi et al., 2018(Takagi et al., , 2021, suggesting that the population densities of P. proximus were high at both sites.

| Field experiment
To investigate the differences in P. proximus colonization success among the four Abies species across the geographic boundary, test  Table 1). Each tree was cut into 10 logs (i.e., 50-60 logs for each species) with a mean length of 1 m. When obtaining the logs, the associations between the logs and some trees were lost, making it difficult to determine which logs came from which trees. To address this issue, I conducted randomization for the logs in order to minimize the probability of selecting logs from the same tree. At the SRS and UTHF sites, five logs of each species (diameter 8-20 cm) randomly selected were placed vertically on the ground before the swarming period of first-generation beetles (Table 1). From late June 2015 (after the offspring larvae started to make galleries) to early July 2015 (before the offspring adults emerged from the logs), the bark was peeled from the logs. The number of successful and failed galleries was recorded. Galleries with living offspring were identified as successful galleries (Figure 2a). In contrast, galleries without oviposition, or with oviposition but larval galleries and adults buried in resin were identified as failed galleries ( Figure 2b).

| Statistical analyses
A generalized linear model (GLM) with a quasi-binomial distribution and logit link was used to determine the effects of host tree species, study sites, and their interactions on the proportion of successfully constructed beetle galleries. p-values were calculated using F-test.
I conducted Tukey's honestly significant difference (HSD) post-hoc tests to determine differences among the four species and to determine significant effects by tree species. Statistical significance was set at p < .05.

Study site Species
Collection date F I G U R E 2 Successful galleries of Polygraphus proximus in Abies veitchii barks (a) and an unsuccessful gallery in Abies firma bark (b).

| RE SULTS
All Abies species were attacked by P. proximus adults at both study sites, and successfully colonized galleries were observed in all Abies species on both sides of the Tsugaru Strait ( Figure 3). The proportion of successfully colonized galleries differed significantly among the four Abies species (F 3, 32 = 143.8, p < .001; Figure 3). The proportion of successfully colonized galleries was the lowest in A. firma (Figures 2 and 3) and the highest in A. veitchii (Figure 3). The proportion of successfully colonized galleries in A. sachalinensis and A.
homolepis logs was significantly higher than that in A. firma logs but lower than that in A. veitchii logs (Figure 3). No significant difference in the proportion of successfully colonized galleries between the study sites (F 1, 32 = 1.50, p = .23), or no significant interaction between sites and species occurred (F 3, 32 = 2.10, p = .12).

| DISCUSS ION
This study demonstrated the colonization success of P. proximus within four Abies species across a distinct biogeographic boundary separating the Abies species in Japan (i.e., the Tsugaru Strait). In the unsuccessful galleries, P. proximus was buried in the resin, although the defense mechanisms are still unclear from this study. Resin characteristics, such as resin flow, chemistry, viscosity, and crystallization, may affect colonization behavior and success. The results showed that the order of colonization success among the Abies species was not significantly different on either side of the Tsugaru Strait. These findings are consistent with those of Takagi (2022) reporting that the Tsugaru Strait does not represent a geographic boundary of host preferences for P. proximus, indicating that host specialization has not occurred across the Tsugaru Strait.
According to the PPH and a positive correlation between host preference and successful colonization (i.e., the preferencecolonization hypothesis) in bark beetles, the rates of successful col-  (Tokuda et al., 2008). Based on the distribution of the Abies species (Figure 1), neither of the P. proximus populations observed in the present study had previously encountered A. firma. Thus, P.
proximus populations in the present study likely had little evolutionary experience with A. firma and a poor estimation of its suitability, resulting in a preference-colonization mismatch (Jaenike, 1978).
This study also showed that the successful colonization rate in A. sachalinensis was high (71.2%-100%). A. sachalinensis is the only native Abies species at the UTHF site. Nevertheless, it was the least preferred host among the four Abies species (Takagi, 2022). This weak preference-colonization relationship in A. sachalinensis appears to be another mismatch between host preference and colonization success. The larval performance in A. sachalinensis may shed light on this mismatch.
Taxonomic proximity among host tree species may influence host preference and the larval performance of bark beetles (Bertheau, Salle, Rossi, et al., 2009;). As a result of feeding on plants that are phylogenetically distant from their preferred hosts, insects generally lose fitness (Bertheau et al., 2010).
However, no positive correlation between host taxonomic proximity and host preference in P. proximus was observed in a previous study (Takagi, 2022). In contrast, this study showed that the relationship between taxonomic proximity and colonization success is complex.
The successful colonization rate in A. sachalinensis, which belongs to the section Balsamea, was not significantly different from that in A.

F I G U R E 3
The proportion of successfully (colored) and unsuccessfully (light gray) colonized galleries of Polygraphus proximus in four Abies species (A. sachalinensis, A. veitchii, A. homolepis, and A. firma) at the UTHF (left column of each species: The University of Tokyo Hokkaido Forest), and SRS (right column of each species: Sugadaira Research Station) study sites. Means with the same letter are not significantly different, p < .05, Tukey Honestly Significant Difference tests. Numerals above each column are the total number of entry holes on each tree species at each study site according to Takagi (2022).
homolepis, which belongs to the section Momi. Conversely, the successful colonization rate of A. veitchii, which belongs to the Balsamea section, was the highest among the four Abies species, whereas that of A. firma, which belongs to the Momi section, was the lowest. These results suggest that there is no positive correlation between host taxonomic proximity and colonization success in P. proximus.
Most preference-performance studies have been carried out on lepidopteran and dipteran species, resulting in skewed information that conforms to the theory (Gripenberg et al., 2010). A previous study reported a lack of correlation between preference and performance in monophagous insects compared with that in oligophagous and polyphagous species, probably because of the small number of studies on this group of insects (Gripenberg et al., 2010). A few studies have reported positive preference-performance relationships in monophagous bark beetles (Bertheau, Salle, Rossi, et al., 2009;Eidson et al., 2018 While this study demonstrated the mismatches between host preference and colonization success in bark beetles, it is worth noting that freshly cut logs were used in this study. Although P. proximus in the unsuccessful galleries were found buried in resin, logs, while containing some residual resin, are largely undefended substrates, incapable for example of induced defenses. The volatilization rates of defense compounds may differ between logs and standing trees.
Further studies are necessary to determine the preferences and colonization success using standing trees.

FU N D I N G I N FO R M ATI O N
This work was supported by JSPS KAKENHI, Grant number 19K15874.

CO N FLI C T O F I NTE R E S T S TATE M E NT
The author declares no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
All files related to this study are available from the DRYAD at https:// doi.org/10.5061/dryad.vq83b k3wh (Takagi, 2023).