Felled and Lure Trap Trees with Uncut Branches Are Only Weakly Attractive to the Double-Spined Bark Beetle, Ips duplicatus

Bark beetles are the most important forest pests in the Northern Hemisphere. The range of Ips duplicatus, an invasive bark beetle in central Europe, has been steadily expanding, and it is now responsible for a high proportion of the spruce wood infested by bark beetles. Apart from searching for and eliminating infested trees, there is no effective control method. The aim of this study was to determine whether trap trees with a pheromone evaporator can be used to capture I. duplicatus. Felled trap trees with branches and with pheromone lures (ID Ecolure®) were infested by I. duplicatus, at a median density of 1 nuptial chambers per 0.1 m2 (median); similar trees without lures and lying at a distance of 1, 5, or 10 m from the lure trees were rarely infested by I. duplicatus. The entire surface of the lure trees could capture <400 beetles per tree. The results indicate that lure trap trees (felled and with branches attached) could only be used in a limited number of situations; one such situation would involve forests that suffered wind damage and contained very high numbers of I. duplicatus.


Introduction
The double-spined bark beetle Ips duplicatus (Sahlberg, 1836) (Coleoptera, Curculionidae, Scolytinae) is native to Fennoscandia, Siberia, and East Asia [1]. During the 20th century, I. duplicatus spread south to Central Europe because of the increased transport of wood that had not been debarked [2][3][4]. The distribution of I. duplicatus has continued to steadily expand [5], and in places where it has been established for a long time, its population densities have increased. Since the 1990s in Poland, the Czech Republic, Slovakia, and Romania, I. duplicatus has infested high volumes of spruce wood [6][7][8][9][10].
The most effective way to control bark beetles that attack spruce trees is to remove infested trees from the forest before the new generation of adult beetles emerges [11]. Supporting control measures include the deployment of trap trees. Such trees, which are used to capture the spruce bark beetle Ips typographus (Linnaeus, 1758) in many countries [12], are mostly felled trees whose branches have been detached and placed over the trunk [13]. I. duplicatus, however, does not infest fallen trees or trap trees whose branches have been detached, and therefore such trees cannot be used for its efficient detection and control [6,14,15]. Because the biology and ecology of I. duplicatus is poorly understood, the reason why beetles do not attack lying trees is not known.
In some regions, branches have only recently been left attached to trap trees for the control of I. typographus [16]. Crown branches prevent the trunk from falling to the ground and cause substantial water loss from the trunk because the needles attached to the crown branches transpire water for some time. The branch-covered trunk is much more attractive to I. typographus than the trunk with attached branches [17][18][19][20][21]. Advantages of leaving branches attached are (1) the trees can be felled later; and (2) I. typographus fly to trunks with attached branches earlier than to trunks with detached branches [22,23]. Stem occupancy by I. typographus differs in trap trees with and without attached branches. I. typographus prefers to infest the crown of trap trees with attached branches but also prefers to infest the trunk of trap trees without attached branches [21].
In some localities, trap trees with attached branches can at least partially attract I. duplicatus [24]. It is possible that a trap tree with branches may attract I. duplicatus due to the intense release of volatiles from needles and branches; I. duplicatus beetles use both natural pheromones and specific host volatiles as olfactory stimuli to locate breeding material [9]. The addition of artificial pheromone lures, of which there are several types that differ in efficiency [25], could increase the infestation of trap trees by the first males.
As previously observed [26], the addition of a pheromone evaporator should increase the attractiveness of trap trees with branches, and the combination of trap trees with branches plus pheromone lures could be useful for I. duplicatus control. This would be appreciated by forest managers because they would save money by not cutting the branches; the trunks with branches attached would be processed by harvesters after I. duplicatus infestation but before the next generation emerged. This would be especially advantageous during I. duplicatus outbreaks when an enormous number of trap trees are deployed, e.g., >5 million per year in the Czech Republic [27][28][29].
The aim of this study was to assess the level of I. duplicatus infestation in felled trap trees with branches left attached and with and without pheromone lures.
Ips duplicatus adults emerge from the duff of the forest floor on warm spring days and fly to host trees that are stressed. The males are responsible for host tree selection. The females are attracted by male pheromones, which contain ipsdienol and E-myrcenol [33,34]. The pheromones attract both sexes. Several females join each male in his nuptial chamber [35].
Each female bores a maternal gallery and lays an average of 60 eggs in niches on both sides of the gallery. After approximately 1 to 2 weeks, the larvae hatch and begin to bore galleries that are approximately 5 cm long. The larval stage usually lasts 2 to 4 weeks, depending on temperatures. Pupation takes about 1 week, and the beetles require additional feeding time for maturation. Maturation feeding requires about 2 weeks, after which the beetles are ready to produce a new brood. The development of one generation usually takes 6 to 8 weeks in Central Europe. The beetles can establish sister generations after the establishment of the main generations. Ips duplicatus produces two or three generations per year, and the main peaks of beetle emergence in central Europe occur in April/May, July, and August/September [14,35,36].

Study Area
The trap trees were deployed in the hills (400-600 m asl) of the eastern Czech Republic in 2016-2018 (Table 1). In this area, I. duplicatus numbers have been increasing for a long time [8].
The I. duplicatus population had reached an outbreak level in the studied area. In 2016 and 2017, the volume of spruce wood infested by I. typographus, I. duplicatus, and other less numerous species ranged from 0.5 to 1.0 m 3 per ha of spruce forests; in 2018, this value exceeded 10 m 3 . The population of I. duplicatus has been outbreaking in the studied area for long time. In 2005, the numbers of beetles of I. duplicatus captured in Theysohn ® pheromone traps lured with ID Ecolure ® had already exceeded 1000 beetles per trap, i.e., these had already attained an outbreak population density [8]. The numbers reached almost 2000 in 2016 and 5000 beetles in 2018. In 2017, captures exceeded 5000 beetles in forest districts [27][28][29].

Sampling
Five study plots were designated at each of the three localities. Study plots at each locality were separated by ≥100 m. Each study plot contained one group of four felled spruce trees with branches; the trees were located in front of forest stand walls that were 60-70 years old and that faced south or west. In each plot and perpendicular to the direction of the forest stand wall, trees with a similar diameter and with a height of 20-24 m were felled either in a north/south or an east/west direction. The mean height of the bottom of the crown was 15.5 ± 1.5 m (±SD). The mean diameter at breast height was 28.2 ± 4.9 cm (±SD). The surface area of the trunks was ca. 9 m 2 per trap tree.
For each study plot, a central tree (the "lure" tree) was provided with a pheromone bait to attract I. duplicatus. The bait was ID Ecolure ® (produced by Fytofarm, s.r.o., and containing ipsdienol 1.6%, myrcenol, spruce resin and ethanol), which is the most effective lure [25]. The bait was placed on the first whorl of green branches of the crown on 25 April of each of the 3 years, which was immediately before the beginning of the flight activity of I. duplicatus. The three other trap trees were placed at 1, 5, and 10 m away from the tree with the lure. The sides alternated, so that at two study plots, the trap trees without lures were located north or west of the trees with lures, and at three study plots, the traps trees were located to the south or east of the trees with lures.
As a consequence, there were four types of trap trees: trap trees with lures and trap trees without lures that were 1, 5, or 10 m away from trap trees with lures.
Infestation of trap trees by I. duplicatus was assessed in mid-June, when the first pupae were detected under the bark. The number of nuptial chambers and the number of maternal galleries of all bark beetle species were recorded on debarked, consecutive bark strips (10 cm × 1 m) on the upper side of the trunk. The consecutive bark strips cover the entire bole of the tree. Strips that were 10 cm-wide were used because a 10 cm-wide strip represented the entire surface of the trunk on the thinnest part of the trees.

Data Analyses
Statistical analysis was performed with MATLAB R2021a (The MathWorks Inc., Natick, MA, USA). Numbers of nuptial chambers per 0.1 m 2 and maternal galleries per 1 dm 2 were used as response variables, and trap tree type, year, position of strip on trap tree (height), and plot (plot) were used as the explanatory variables. There were four types of trap trees: trap trees with lures and trap trees without lures that were 1, 5, or 10 m away from trap trees with lures. Because the response variable followed a Poisson distribution, a general linear regression model in the form log (numbers of nuptial chambers)~1 + plot + type + height (Poisson distribution of response variable) was used to quantify the effect of the predictors on the number of nuptial chambers for each bark beetle species. To quantify the effect of the predictors on the number of maternal galleries, the form of the generalized linear model was 1/(maternal gallery)~1 + plot + type + height, corresponding to the Gamma distribution of the response variable. For both models, link functions were chosen as the default link functions according to the distributions of the response variable. When calculating the effect of year on numbers of nuptial chambers and maternal galleries, year was used instead of plot. Although the position on trap tree trunks (height) was used as a predictor, the preferences of individual bark beetle species for positions on tree trunks are known [37]; as a result, we did not present the effects of this predictor.

Results
Numbers of I. duplicatus nuptial chambers on individual bark strips ranged from 0 to 9 per 0.1 m −2 with a median of 1.0. Numbers of I. typographus and P. chalcographus nuptial chambers ranged from 0 to 10 per 0.1 m −2 (Figure 1). The maximum number of maternal galleries per dm −2 of bark was 4, 5, and 6 for I. typographus, I. duplicatus, and P. chalcographus, respectively ( Figure 2).
Numbers of I. typographus and P. chalcographus nuptial chambers, and maternal galleries of I. duplicatus did not significantly differ on lure trees vs. trees without lures that were 1, 5, or 10 m away from lure trees (Tables 2 and 3). For I. duplicatus, in contrast, the numbers of nuptial chambers were highest on the lure trees and significantly decreased with distance from the lure tree; as low numbers were detected on trees without lures that were 1 m away from lure trees, but none was detected on trees without lures that were 5 or 10 m away from lure trees ( Table 2). types of trap trees: trap trees with lures and trap trees without lures that were 1, 5, or 10 m away from trap trees with lures. Because the response variable followed a Poisson distribution, a general linear regression model in the form log (numbers of nuptial chambers) ~ 1 + plot + type + height (Poisson distribution of response variable) was used to quantify the effect of the predictors on the number of nuptial chambers for each bark beetle species. To quantify the effect of the predictors on the number of maternal galleries, the form of the generalized linear model was 1/(maternal gallery) ~ 1 + plot + type + height, corresponding to the Gamma distribution of the response variable. For both models, link functions were chosen as the default link functions according to the distributions of the response variable. When calculating the effect of year on numbers of nuptial chambers and maternal galleries, year was used instead of plot. Although the position on trap tree trunks (height) was used as a predictor, the preferences of individual bark beetle species for positions on tree trunks are known [37]; as a result, we did not present the effects of this predictor.

Results
Numbers of I. duplicatus nuptial chambers on individual bark strips ranged from 0 to 9 per 0.1 m −2 with a median of 1.0. Numbers of I. typographus and P. chalcographus nuptial chambers ranged from 0 to 10 per 0.1 m −2 (Figure 1). The maximum number of maternal galleries per dm −2 of bark was 4, 5, and 6 for I. typographus, I. duplicatus, and P. chalcographus, respectively ( Figure 2).
Numbers of I. typographus and P. chalcographus nuptial chambers, and maternal galleries of I. duplicatus did not significantly differ on lure trees vs. trees without lures that were 1, 5, or 10 m away from lure trees (Tables 2 and 3). For I. duplicatus, in contrast, the numbers of nuptial chambers were highest on the lure trees and significantly decreased with distance from the lure tree; as low numbers were detected on trees without lures that were 1 m away from lure trees, but none was detected on trees without lures that were 5 or 10 m away from lure trees ( Table 2).    Table 2. Relationships between the number of nuptial chambers and plot, type of trees (distance from lure tree), and the height of the position of strip on the trap tree as indicated by the general linear model log (numbers of nuptial chambers) = 1 + plot + type + height. The coefficient estimates are accompanied with the standard error of estimate (SE) and the p-values.   Table 2. Relationships between the number of nuptial chambers and plot, type of trees (distance from lure tree), and the height of the position of strip on the trap tree as indicated by the general linear model log (numbers of nuptial chambers) = 1 + plot + type + height. The coefficient estimates are accompanied with the standard error of estimate (SE) and the p-values.

Discussion
Only three species of bark beetles (Ips typographus, Ips duplicatus, and Pityogenes chalcographus) were found on the felled trap trees in the current study. This was not surprising because these three species are the most numerous bark beetles in the study area [13,38] and because these abundant species rapidly infest felled trees and thereby out-compete other species that might infest spruce. The infestation level of the trap trees varied from year to year, and this yearly variation was consistent with the abundance of the species in the wider area. For example, I. duplicatus was the most abundant species on our trap trees in 2017, which was the year when the largest number of I. duplicatus beetles were captured in pheromone traps that were deployed in the surrounding districts [27][28][29].
Among the four kinds of trap trees (lure and the trap trees without lures located at 1, 5, or 10 m distances from the lure trees), Ips duplicatus numbers were highest in the lure trap trees, and numbers among trap trees without lures declined with the distance from the lure trap trees. I. duplicatus attraction to the lure was probably also evident on the trap trees without lures located at a 1 m distance from the lure trees; researchers previously reported that the release of high concentrations of pheromones from one tree can cause I. duplicatus adults to also infest neighboring trees [31]. The failure of I. duplicatus to infest trap trees without lures located 5 or 10 m from lure trap trees confirms the well-known fact that I. duplicatus beetles attack standing trees only [6,15].
It is very likely that host volatiles function as olfactory stimuli that attract I. duplicatus. This is supported by the tendency of I. duplicatus beetles, regardless of sex, to aggregate in areas where large amounts of fresh material are available [9] and to be attracted to α-pinene [51].
Because the height of the dispersal flight of I. duplicatus generally corresponds to the height of the branches of the standing trees [52], this species is probably attracted by crown volatiles. Volatiles from spruce needles, knots, and wood extracts include characteristic compounds [53][54][55][56]. I. typographus, in contrast, mainly infests the subcrown parts of spruce trees because those parts emit the highest concentrations of monoterpenes [49,57]. Felled trees with branches attached can be especially attractive because the dying branches and needles may release abundant volatiles. The low emission of volatiles by trap trees whose branches have been removed would explain why these kinds of trap trees tend not to be infested by I. duplicatus, even when they were lured by a pheromone evaporator (Holuša, Galko observ.).
The release of volatiles varies greatly among trees and depends on temperature, light, and humidity [44,58,59]. This could help explain why >1.5 I. duplicatus nuptial chambers per dm −2 were observed on several lure trees in a previous study [26].
Ips duplicatus probably prefers standing trees for another reason. The essential pheromones are ipsdienol (Id) and E-myrcenol (EM) for I. duplicatus but are 2-methyl-3-buten-3-01 (MB) and cis-verbenol (cV) for I. typographus [60]. In trees at an advanced stage of attack, after female egg deposition, I. typographus males produce ipsdienol (R-(−)) and ipsenol (R-(+)) in their hindguts [61,62]. Ipsdienol is attractive to females [63,64] at low doses but repels females at higher doses [65]. Because the enantiomer ratio of ipsdienol is much higher in I. typographus than in I. duplicatus, I. duplicatus is unlikely to respond to the enantiomer ratio of ipsdienol in I. typographus. In a tree colonization model, the response of the two competing species to their respective pheromones shows a good separation during the mass attack, but a small initial cross-attraction [60]. It could benefit I. duplicatus to receive a cue indicating that the tree has been successfully attacked or weakened by I. typographus [60]. We therefore conclude that I. duplicatus is attracted to standing trees whose defenses have already been overcome by I. typographus. The two species apparently reduce their interspecific competition by infesting different parts of weakened trees; I. duplicatus prefers the crown, and I. typographus prefers the subcrown. On the other hand, we must admit that very abundant populations of I. duplicatus can occupy entire spruces, at least those weakened by drought. This happens especially in the case of smaller trees unsuitable for I. typographus (Holuša, observ.).

Conclusions
We found that felled spruces with branches and with pheromone lures (ID Ecolure ® ) were only weakly infested with I. duplicatus. We suspect that this low level of infestation may have been due to the absence of both needle volatiles and an aggregation pheromone, two scents that were probably supplied by standing ring-barked lured trap trees. In addition, lures in our experiments were placed at a practical but low height of 3 m. Our observations suggest that I. duplicatus is not attracted to lures placed at low heights (Holuša unpubl.). In the current study, based on the number of nuptial chambers (one per 0.1 m 2 ) and sex ratio (1 male: 2.5 females) on the whole surface of the studied trees, we captured fewer than 400 beetles per whole lure tree. Although the number of I. typographus beetles is sufficient to overcome a weakened spruce tree [66], it is far lower than the number of I. duplicatus captured per monitoring trap previously reported in the study region [27][28][29]. This shows that pheromone traps are more efficient than trap trees for capturing I. duplicatus beetles. Unfortunately, captured beetles in pheromone traps must be removed regularly, because the smell from the accumulated dead beetles significantly reduces trapping efficacy [67].
Felled and lure trap trees with branches could be effectively used in only a limited number of situations; for example, they could be used in areas with substantial wind damage and with very high levels of I. duplicatus. If forest managers do not manage to process all of the trees in a wind-damaged area, we recommend the placing of pheromone traps at intervals of approximately 5-10 m to attract and capture at least part of the I. duplicatus population.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Acknowledgments:
The authors thank the foresters from the LesyČeské Republiky, s.p. (state enterprise) for cooperation in the research.

Conflicts of Interest:
The authors declare no conflict of interest.
Appendix A Table A1. Relationships between number of nuptial chambers and year, type of plot (distance from lure tree), and tree height of position of strip on trap tree as indicated by the general linear model log (number of nuptial chambers) = 1 + year + type + height. The coefficient estimates are accompanied with the standard error of estimate (SE) and the p-values.

Species
Year