Testing trapping protocols for detecting the Citrus Longhorn Beetle, Anoplophora chinensis (Coleoptera: Cerambycidae)

Citrus Longhorn Beetle (CLB), Anoplophora chinensis (Coleoptera: Cerambycidae) is a highly polyphagous species native to eastern and south‐eastern Asia. In 2000, an outbreak of A. chinensis was detected in Lombardy Region (Italy). In 2017, an extensive trapping experiment was conducted at three infested sites in order to evaluate the effectiveness of three variables: trap model (Econex soft cross‐vane trap, Witasek cross‐vane trap and Witasek multi‐funnel trap), trap position (on a wooden pole in an open space or in the canopy of a host tree) and type of lure (ChemTica, Synergy and Glabriwit pheromones). Each combination of variables was replicated five times at each site, giving a total of 270 traps. At the end of the study, 162 adults had been caught, with catches gradually increasing during the month of June to peak in early July. The two cross‐vane traps outperformed the multi‐funnel traps and the Econex traps captured more females than the Witasek traps, probably due to the structure of the collecting funnel. The three lures had similar catch performance, although the best combination was the Econex trap with Synergy blend, due to the remarkable variability in catches observed with Glabriwit blend. Finally, traps set in the tree canopy outperformed traps set on wooden poles. In conclusion, the best protocol was the use of Econex cross‐vane traps baited with Synergy blend and deployed on the canopy of the host trees.

latitude (Haack et al., 2010;Van Der Gaag et al., 2010). Consecutive generations of beetles may affect the same plant for several years.
Anoplophora chinensis is a highly polyphagous species reported from a wide range of broadleaf host-trees belonging to 26 families (Lingafelter & Hoebeke, 2002), but occasionally recorded also from conifers of the genera Cryptomeria and Pinus (Hoppe et al., 2019).
In addition to control measures carried out according to EU legislation, Lombardy Region implemented a specific survey of the regional territory based on the use of traps baited with lures attractive to CLB. Nevertheless, very little information is available about the best trap model and lure for CLB detection. Similarly, no data are known concerning the best trap position-that is, height from ground, type of trap support and distance from possible host treesfor CLB interception. Moreover, there is no commercially available pheromone blend specific for CLB, only those applied against the Asian Longhorn Beetle (ALB), Anoplophora glabripennis, which, apparently, shares one of its male-produced volatile pheromones 4-(n-heptyloxy) butanal (Hansen et al., 2015).
Improving trapping protocols available for the interception of invasive species is the first step for their detection and one of the most crucial factors facilitating their eradication. In this context, this study aims at evaluating the effectiveness of various trap models, different lure blends and trap positions for trapping CLB adults. Based on a large field experiment conducted in northern Italy, this study is focused on the identification of the best protocol (i.e., trap model, lure, trap position and their combinations) for CLB detection.

| Experimental sites and periods
The study was conducted on three CLB populations occurring in northern Italy (Figure 1), used as replicates: a. 'Milano' site, including the infestation in Milan city and in South Milan Agricultural Park; b. 'Altomilanese' site, including the infestation in Nerviano (Milan province) and neighbouring municipalities; c. 'Brescia' site, including the infestation in the municipality of Gussago (Brescia province).
All the experimental sites were within CLB infested areas ensuring the presence of the species and, therefore, suitable to evaluate the effectiveness of the tested protocols. Table 1 provides information about the infestation rates (i.e., number of monitored trees, number of infested trees, number of exit holes and percentage of infested trees) in the three sites.
For the same reasons, the study was conducted during the period of maximum CLB flight activity of the emerging beetles which in northern Italy is generally concentrated in June and July. Traps were, therefore, set up and baited with lures between 31 May and 7 June 2017, and withdrawn between 31 July and 3 August 2017, covering 2 months. All the traps were checked and emptied every second week, that is, four times at mid and the end of June and July, recording the number of CLB males and females collected in each date and trap.
The specimens were sexed on the basis of morphological characters (Lingafelter & Hoebeke, 2002) and the mating status of females was not determined. Lures were replaced after 4 weeks (in the middle of monitoring) to maintain the recommended emission rate.

| Trap setting
At each of the three experimental sites, traps were set up in five dif-

| Experimental design
The study was based on an experimental design aimed at identifying the best combination of different trap models, lures and trap position in the territory in order to increase the probability of catching CLB adults.
In this regard, three different models of traps were tested

| Statistical analysis
Data of mean CLB captures per trap were subjected to analysis of variance (ANOVA) by the general linear model for randomized block designs (Zar, 1999) to test differences between trap model, lures, trap position and their interactions, using the STATISTICA for WINDOWS software. Homogeneity of variance was tested using Cochran's test, and when necessary, data were log-transformed [X = log (x + 1)] to obtain homogeneous variances. Where significant differences among variables occurred, Tukey's honestly significant difference (HSD) multiple comparison test was applied for mean separation (Zar, 1999). Differences at 0.05 level of confidence were considered significant. Means were calculated for all the same traps across all the sites and locations: in fact, sites and locations were considered replicates with homogeneous characteristics.

| Total captures and flight period
A total of 162 CLB adults were captured in June-July 2017, 84 males and 78 females, without significant differences between sexes (ANOVA, df = 1; 268, F = 2.13, p > 0.05). Catches gradually increased during the month of June to peak in early July and then gradually drop in the following month ( Figure 3). There were fewer catches in Brescia (the site with a lower level of infestation) than the two sites in Milan.

| Trap-lure interactions
Mean catches of CLB adults were significantly affected by the interactions between trap model and lure blend used to bait the trap (ANOVA, df = 4; 264, F = 1.29, p < 0.02). In particular, Econex crossvane traps baited with Synergy or Glabriwit pheromones and the Witasek cross-vane traps triggered with Synergy pheromone were the combinations providing significantly higher captures than all the others ( Figure 6). The lowest trapping values were recorded with Witasek cross-vane traps baited with the pheromone Glabriwit. In general, multi-funnel traps showed relatively low catches regardless of the lure tested.

| DISCUSS ION
The results of the CLB trapping protocols tested in this study indicate cross-vane traps as the best trap model to be used to increase the probability of catching CLB. This result is in agreement with previous reports for similar species: cross-vane traps are more effective than multi-funnel traps in catching many families of bark and woodboring beetles, including long-horn beetles (Allison & Redak, 2017).
Moreover, even in ALB monitoring and eradication protocols, mainly cross-vane traps were used (Eyre & Barbrook, 2021;Nehme et al., 2014). Although non-significant differences occurred between Among lures tested in the study, the Synergy ones provided the best results with high mean capture levels, although they did not differ statistically from the other two blends. Looking at the formulations of the blends, it can be seen that they are similar in their components, but differ significantly in kairomone quantities; in particular, the Synergy one has the highest quantity of them.  (Dodds, 2011;Sweeney et al., 2020). According to the results of our study, the protocol that provides the highest levels of CLB captures is therefore the installation of cross-vane traps in the crowns of CLB host trees. In this context, however, there is a risk of an over-spilling effect, that is, insects attracted by the pheromone to a host plant and then not caught by the trap but directly infesting the tree. In this respect, healthy plants used to install traps must be carefully surveyed to avoid initiating new local infestations.
In relation to the insect phenology recorded by this study, the best season to conduct a survey of CLB populations by pheromone traps in northern Italy-or to verify the presence of this species in a new territory-falls between mid-June and mid-July. The main flight activity of CLB adults occurs in this period and, therefore, the highest probability of insect interception.
In conclusion, the best protocol is the use of Synergy blend to bait Econex cross-vane traps, which have a higher, though not significant, catch rate than the Witasek model and allows easier winter storage due to the soft panels. Finally, the best position for traps is on host trees, although it is necessary to plan periodic checks in order to avoid the over-spilling effect.
Finally, although the mean catches per trap recorded in this study are in general particularly low, they are in line with ALB catches reported in other papers (Nehme et al., 2014). Moreover, the use of pheromone traps against CLB allows species detection and a spatial and temporal survey of its populations, giving crucial information to set up survey programme and to assess the effectiveness of control measures applied.

ACK N OWLED G EM ENTS
We thank Lombardy Region for partially funding the research, data providing and their help in field works. A special thanks to University

CO N FLI C T O F I NTE R E S T
We have no conflicts of interest to declare.

AUTH O R CO NTR I B UTI O N S
Bianchi A, Ciampitti M and Faccoli M conceived research. Bianchi A conducted experiment. Faccoli M analysed data and conducted statistical analyses. All authors, in equal parts, wrote the manuscript.
Ciampitti M secured funding. All authors read and approved the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data used for this work are available in a public repository (Research Data Unipd), identified with http://resea rchda ta.cab.unipd.it/id/ eprin t/540 (Marchioro et al., 2021).