Attraction of the cerambycid beetles Tetropium gabrieli, T. castaneum and T. fuscum to pheromones and host tree volatiles

Tetropium gabrieli is native to the European Alps where it breeds in European larch and may be an important pest during drought periods by killing trees. It has spread to larch plantations in several European countries including Sweden. In trapping studies conducted in Sweden, we tested whether T. gabrieli was attracted to E‐fuscumol, E,Z‐fuscumol, E,Z‐fuscumol acetate and the host volatile combination of α‐pinene plus ethanol. We also compared its response with those of the native congeners, T. fuscum and T. castaneum. All three Tetropium species were caught in significantly higher numbers in traps baited with E‐fuscumol plus host volatiles than in traps baited with host volatiles alone or in unbaited traps. There was no difference in catch of any Tetropium species between traps baited with E‐fuscumol plus host volatiles versus traps baited with E,Z‐fuscumol plus host volatiles, but the latter treatment caught more T. fuscum and T. castaneum than did E,Z‐fuscumol acetate plus host volatiles. The addition of E‐fuscumol to host volatile‐baited traps significantly reduced mean catch of T. gabrieli, significantly increased mean catch of T. castaneum, and had no significant effect on catch of T. fuscum. Mean catch of all three species was significantly greater in traps baited with E‐fuscumol alone than in traps baited with host volatiles alone. All three species were caught from May to August. The median date of captures was about 2 weeks earlier for T. fuscum compared with the other two species.


Introduction
The cerambycid beetle Tetropium gabrieli Weise (subfamily Spondylidinae) is native to the Alps in central Europe where it breeds in European larch, Larix decidua Mill. The species also reproduces in other larch species non-indigenous to Europe and is considered to be a secondary pest, breeding mainly in newly felled trees or weakened standing trees. However, in some regions in Europe T. gabrieli is classified as an important pest species during hot and dry summers (Gorius, 1955;Evans et al., 2004). Thus, the economic importance of this species may Correspondence: Martin Schroeder. Tel.: +46767924510; e-mail: martin.schroeder@slu.se increase in the future as a result of climate change increasing the frequency of prolonged droughts during summers.
The planting of larch outside the natural distribution of European larch has enabled T. gabrieli to expand its distribution and establish in many countries in Europe (Sama & Löbl, 2010). In Sweden, it has been reported several times from imported larch wood (Lundberg, 1986(Lundberg, , 1988Lundberg & Pettersson, 1997) and in 2007 T. gabrieli was detected infesting larch trees in a forest adjacent to a port in the province of Blekinge in southeastern Sweden (Ericsson, 2010). Surveys conducted from 2009 to 2011 using spring-felled larch trees demonstrated the beetle was present in several locations in southeastern Sweden but absent in more northern or western locations (Lindelöw et al., 2015). Although this survey method worked, it is practical only in larch stands, requires permission to cut trees, is quite laborious, and requires either splitting of logs to remove larvae or incubation of logs to collect emerging adults for reliable species identification. A more cost-efficient method would be to capture adult T. gabrieli in attractant-baited traps. In recent years, a trap lure was developed for the congener T. fuscum (F.), a species native to Europe but which has established in Eastern Canada where it has caused considerable tree mortality among native spruce species (Silk et al., 2007;Sweeney et al., 2004Sweeney et al., , 2006Sweeney et al., , 2010. Traps are baited with three lures: racemic (E)-6, 10-dimethyl-5,9-undecadien-2-ol (fuscumol), ethanol, and a blend of monoterpenes that approximates those sampled from bark of stressed spruce, i.e., racemic α-pinene, (−)-β-pinene, (+) 3-carene, (+) limonene and α-terpinolene. Male T. fuscum emit S-(E)-fuscumol but the synthetic racemic mixture of R-and S-(E)-fuscumol is as attractive as the pure S, and the combination of spruce host volatiles and racemic fuscumol synergizes attraction of both sexes of T. fuscum . In addition to T. fuscum, T. cinnamopterum Kirby (native to North America) males also emit S-(E)-fuscumol and males and females of T. cinnamopterum and T. castaneum (L.) (native to Europe) are attracted to the combination of fuscumol and host volatiles (Silk et al., 2007;Lemay et al., 2010;Sweeney et al., 2010). Thus, there is a high probability that additional Tetropium species may also be attracted to E-fuscumol and thus we wanted to test this for T. gabrieli.
Although the main focus of the study was T. gabrieli, we also expected catches of T. fuscum and T. castaneum which both are common all over Sweden (Ehnström & Holmer, 2007). The following questions were addressed by the study: (i) Is T. gabrieli attracted to fuscumol or fuscumol acetate? (ii) Does addition of the host tree volatiles α-pinene and ethanol increase attraction of T. gabrieli compared with pheromone alone? (iii) Does the phenology of adult flight differ among the three species? (iv) How long is the emergence period of T. gabrieli? (v) What parasitoids species are associated with T. gabrieli? The host volatile α-pinene was chosen because it is one of the major monoterpenes in larch (Holm & Hiltunen, 1997;Ruuskanen et al., 2007;Kajos et al., 2013) while ethanol is released from stressed or decayed trees. Phenology of flight, emergence period and parasitoids are of interest for understanding the invasion biology of the species.

Study area
The study was conducted in a 4.3 × 3.3 km (1440 ha) rectangular area in south-eastern Sweden (56 ∘ 19 ′ 13.8"N 14 ∘ 49 ′ 32.2 ′′ E, province of Blekinge) in 2017 in which T. gabrieli had been detected in earlier surveys using cut larch trees (Lindelöw et al., 2015). Managed forest stands comprised 986 ha (72.6%) of the study area, including 60.5 ha (4.2% of study area) of hybrid larch stands, Larix x eurolepis A. Henry, (13 stands including one fresh clear-cut) and 925.5 ha (68.4% of study area) of Norway spruce, Picea abies (L.) H. Karst., and Scots pine, Pinus sylvestris L., dominated stands. The remainder of the study area included pastures and swamps (20.8%) and lakes (6.6%). In a 1.2 km buffer zone (2420 ha) around the study area, there was only one larch stand (4.6 ha, 0.2% of buffer area). The proportion of larch in the study area is much higher than the average for southern Sweden which is 0.1% of the total growing stock including all tree species (Anonymous, 2016).

Lures
Racemic E-fuscumol, racemic E,Z-fuscumol, and racemic E,Z-fuscumol acetate (99% purity) lures were purchased from Sylvar Technologies Inc. (Fredericton, New Brunswick, Canada). The lures were natural red rubber septa (sleeve stoppers), 5 × 11 mm (OD) (Wheaton, Millville, NJ) loaded with 50 mg of fuscumol or fuscumol acetate; release rates were 0.5-2.0 mg/day at 20 ∘ C (measured indoors by weight loss). The host volatiles ethanol (95%) and (−)-α-pinene (98%, optical purity ee: ≥81%, GLC, Aldrich) were released from separate, sealed 50 ml PVC-containers filled with 30 ml of compound, with holes bored through the sides of the container just below the lid. Ethanol dispensers had four 0.5 mm-holes and (−)-α-pinene dispensers had eight 2 mm-holes. Release rates (determined by weight loss in the laboratory at 20 ∘ C) were 140 mg/day for ethanol and 940 mg/day for (−)-α-pinene. We also measured host volatile release rates in the field as follows. At each trap check, the dispensers were refilled if more than half of the compound had been released and the amount that was added to each dispenser was recorded. By this method, we estimated average release rates in the field of 272 and 1605 mg/day for ethanol and (−)-α-pinene, respectively. However, this is an underestimate of the mean outdoor release rate of (−)-α-pinene because about half of the (−)-α-pinene dispensers were found empty during trap checks, i.e., all 30 ml dispensers had volatilized since the previous trap check 1-3 weeks earlier. This means that some traps were releasing no (−)-α-pinene for at least a portion of the experimental trapping period.

Field experiments
Three trapping experiments were conducted sequentially with the same traps, each with three treatments. In all three experiments, we used α-pinene and ethanol as host volatiles, released from the dispensers described above. The first experiment, conducted 8 May-13 June, compared catch of T. gabrieli in traps baited with E-fuscumol plus host volatiles, host volatiles alone, or left unbaited. The aim of this experiment was to test whether T. gabrieli was attracted by the combination of E-fuscumol and host tree volatiles as demonstrated for other Tetropium species. The second experiment, conducted 13 June-29 June, compared E-fuscumol, E,Z-fuscumol, and E,Z-fuscumol acetate, each paired with host volatiles. The objective here was to compare the response of T. gabrieli, as well as the native Tetropium species, to E,Z-fuscumol and E,Z-fuscumol acetate. The third experiment, conducted 29 June-12 September, compared catches in traps baited with E-fuscumol alone, E-fuscumol plus host volatiles, and host volatiles alone. Here, our objective was to determine whether the combination of fuscumol and host volatiles synergized attraction of T. gabrieli as has been observed with T. fuscum and T. castaneum . In addition, because traps baited with E-fuscumol plus host volatiles were included in all three experiments, this allowed us to compare the seasonal phenology of adults of the three Tetropium species.
Treatments were replicated 12 times in a randomized complete block design, with re-randomization between experiments. The 12 blocks were distributed among 10 sites, 9 of which consisted of larch stands (each with one block) and one which was a fresh larch clear-cut (with three blocks). The average size of the larch stands was 4.6 ha (SE = 0.7 ha) while the clear-cut was 11.8 ha. Minimum distance between sites was 300 m and maximum distance 3700 m. Traps were spaced 50 m apart within blocks and at least 10 m from the edge of the stands.
Beetles were caught in 12-unit multiple funnel traps (ECONEX MULTIFUNNEL-12®, Econex SL) hung from 1.8 m wooden poles with 50% solution of ethylene glycol in the collecting cups. The traps were emptied seven times (exp. 1: twice; exp. 2: twice; exp. 3: three times) with intervals from 1 to 3 weeks (except for the last one with almost 5 weeks).
Tetropium adults were determined to species based on Pfeffer (1995) and Ehnström and Holmer (2007). The beetles were sexed by inspection of the tip of the abdomen. As is common in many Cerambycidae, a portion of the sixth abdominal tergite is visible beyond the pygidium of males (Yanega, 1996), whereas the female's pygidium is flatter and broader and the tip of the ovipositor is often visible.
On May 8-9 2017, we felled a larch tree in each of five larch stands in which we had traps and on September 11-14, we inspected them for colonization by Tetropium by debarking parts of the trunk every 3 m. In addition, 2 m logs were cut from the three trees that were colonized by Tetropium and also from a colonized wind-felled tree. The logs were stored outside the laboratory in Uppsala. On March 13, two 60 cm stem sections were cut from each of the logs and ends were waxed. Each stem section was placed in a separate cage in a climate chamber (20.4 ∘ C, 18 h day-length) and checked each day (except weekends) for emerging insects for 78 days. On April 9, additional two 60 cm stem sections were cut from logs of the same four trees and handled in the same way as the first batch of stem sections. Mean diameter of the 16 stem sections was 21.3 cm (SE = 0.4). Emerging parasitoids were identified by Niklas Johansson at the Swedish Species Information Centre, Swedish University of Agricultural Sciences. Rate of parasitism was estimated for each tree by dividing the number of emerging parasitoids by the sum of emerging Tetropium beetles and parasitoids (assuming that each parasitoid killed one Tetropium larvae).
In addition, part of the colonized logs from three of the trees (40, 73 and 33 cm length stem sections respectively) were split on 15 March to count and collect Tetropium larvae. For each larva, the width of head-capsule was measured and it was noted if it occurred under the bark or in a gallery in the wood. The lengths of all galleries in wood, and of the plug blocking their entrances, were measured.
All nine larch stands used for trapping were inspected for occurrence of wind-felled larch trees which were checked for colonization by Tetropium by debarking. In addition to the cut and wind-felled trees, we also checked 30 stumps (30-40 cm in height) in the fresh clear-cut (cut winter 2016/2017) in early April 2018 for colonization by Tetropium.

Statistical analyses
For each species, catches were totalled for each treatment and block over the entire trapping period for each experiment and analysed using generalized linear models (SAS PROC GLIM-MIX) with treatments fixed and blocks random. Blocks in which all three lure treatments had zero catch were deleted from analysis. The models were run with Gaussian, Poisson, and negative binomial distributions; results are reported for the distribution that best fit the data, according to the lowest value of the corrected Aikaike's Information Criterion (AIC c ). Post-hoc comparisons of least square means were done with Tukey-Kramer multiple comparisons which control the experiment-wise error rate (P ≤ 0.05). We also analysed the raw data and data transformed by log (y + 1) using standard ANOVA (SAS PROC GLM), testing residuals for normality using Shapiro-Wilks test, and compared results with those from the generalized linear mixed models. When results from the GLM and GLIMMIX analyses differed greatly, e.g., zero values for F in the generalized mixed model and significant F values in the GLM for the same data, we assumed the data were not suitable for analysis by the mixed models. This occurred in two cases (Experiment 3), possibly because the host volatile treatment caught zero T. gabrieli and T. fuscum. In these cases, count data were converted to ranks and analysed by Friedman's nonparametric ANOVA for randomized blocks, followed by the Ryan-Einot-Gabriel-Welsh multiple range test. For each Tetropium species, chi-square contingency tables (or Fischer's exact test in cases when expected values were below 5) were used to test whether sex ratio of trapped beetles was affected by lure treatment. Only treatments with at least four caught individuals were included in the sex ratio tests.

Results
Three Tetropium species were caught in the trapping experiments: the native T. fuscum (154 specimens) and T. castaneum (2230), and the exotic T. gabrieli (57). All three species were caught in all 12 blocks. Total catch per block (all experiments and treatments pooled per species) for T. fuscum was positively correlated with that of T. castaneum (r = 0.81, N = 12, P = 0.001) while there was no correlation between catches of T. gabrieli and For each species, blocks in which all three lure treatments had zero catches were deleted from analysis and n gives the number of remaining blocks. Data analysed by generalized linear models using distribution with best fit as measured by corrected Aikaike's information criterion. Means followed by different letters differed significantly, Tukey-Kramer multiple comparison test, P ≤ 0.05. Sex ratio only given for treatments with ≥4 caught individuals.
Experiment 1. Catches of T. gabrieli, T. fuscum and T. castaneum were 28, 55 and 62 times higher in traps baited with E-fuscumol plus host volatiles than in unbaited traps and these differences were significant (Table 1). All three were also caught in significantly higher numbers in traps baited with E-fuscumol plus host volatiles than in traps baited with only host volatiles (Table 1). Traps baited with host volatiles alone caught 2-5 times more beetles than unbaited traps but the difference was significant only for T. castaneum ( Experiment 3. Contrary to our expectations, adding α-pinene plus ethanol to traps baited with E-fuscumol significantly reduced mean catch of T. gabrieli; in fact mean catch in traps baited with E-fuscumol plus host volatiles was less than half of that in traps baited with E-fuscumol alone (Table 3). In contrast, traps baited with E-fuscumol plus host volatiles caught significantly more T. castaneum (5 times as many) than traps baited with E-fuscumol alone, which, in turn, caught significantly more T. castaneum (12 times as many) than traps baited with host volatiles alone (Table 3). Catches of T. fuscum were low. Mean catches in traps baited with E-fuscumol and E-fuscumol plus host volatiles did not differ, but both were significantly greater than For each species, blocks in which all three lure treatments had zero catches were deleted from analysis and n gives the number of remaining blocks. Data analysed by generalized linear models using distribution with best fit as measured by corrected Aikaike's information criterion. Means followed by different letters differed significantly, Tukey-Kramer multiple comparison test, P ≤ 0.05. Sex ratio only given for treatments with ≥4 caught individuals. For each species, blocks in which all three lure treatments had zero catches were deleted from analysis and n gives the number of remaining blocks. Data for T. castaneum analysed by generalized linear models using distribution with best fit as measured by corrected Aikaike's information criterion; means followed by different letters differed significantly, Tukey-Kramer multiple comparison test, P ≤ 0.05. Data for T. gabrieli and T. fuscum analysed by Friedman's test (ANOVA on ranks) followed by Ryan-Einot-Gabriel-welsh multiple range test (P ≤ 0.05). Sex ratio only given for treatments with ≥4 caught individuals. the zero catch in traps baited with host volatiles alone (Table 3). The addition of host volatiles to E-fuscumol did not affect the sex ratio for any of the three species (T. gabrieli, P = 0.33, chi-square test, T. fuscum, P = 1, Fischer's exact test, T. castaneum, P = 0.25, chi-square test) (Table 3).
Seasonal phenology of Tetropium spp. All three species were caught from May to August (Fig. 1). Most of the flight activity occurred in May and June. The median date of captures was about 2 weeks earlier for T. fuscum compared with the other two species. For both T. fuscum and T. castaneum (catches of T. gabrieli were too low for evaluation) a somewhat larger part of the flight activity of females than that of males occurred late in the season: 90% of the cumulative catches occurred about 10 days later for females than for males (data not shown). Three of the five larch trees cut in the spring 2017 were colonized by Tetropium. About 75% of the tree stems were colonized. No other bark-and wood-boring beetle species were found on the stems. Wind-felled larch trees were found in two of the nine stands: six trees colonized in 2016 in one stand and one tree colonized in 2017 in another stand (same stand in which a cut tree was not colonized). There were no Tetropium emergence holes in the autumn in the colonized trees cut in the spring the same year. None of the 30 larch stumps inspected on the clear-cut were colonized by Tetropium spp.
Tetropium gabrieli was the only bark-and wood-boring insect species emerging from the 16 stem sections cut from four naturally infested cut and wind-felled larch trees except for four individuals of Urocerus gigas (L.) (Hymenoptera, Siricidae). A total of 782 individuals of T. gabrieli emerged: 276 (50% males), 212 (52% males), 142 (44% males) and 152 (50% males) individuals respectively from tree 1, 2, 3 and 4. The average production of T. gabrieli adults was 122 per m 2 bark (SE = 19, N = 4) and 2309 per m 3 wood (SE = 347) individuals. The emergence pattern was similar among trees: starting 14-17 days after stem sections were put in the climate chamber, 50% emergence after 19-21 days and 90% emergence after 25-32 days (Fig. 2). There was almost no difference in emergence pattern between sexes, 50% emergence of males occurred less than 1 day earlier than that for females (data not shown).
A total of 97 Tetropium larvae were found during the splitting of stem sections from three colonized trees. The Tetropium larvae were not determined to species but because only T. gabrieli adults emerged from the stem sections it is highly unlikely that the larvae were another species. All larvae were alive. About 34% of the larvae were found in galleries in the wood while the remaining 66% were found under the bark. Head-capsule width ranged from 1.3 to 3.4 mm with the majority (92%) >1.9 mm (Fig. 3). Head-capsule width was significantly larger for larvae in galleries in the wood than for larvae under bark: mean = 2.84 mm (SE = 0.06) and 2.53 mm (SE = 0.06) respectively (P < 0.001, t-test). Average length of galleries in the wood was 45 mm (SE = 2) and the entrance of all galleries were blocked with a plug (mean length = 7.4 mm, SE = 0.7 mm).

Discussion
We present evidence from field-trapping experiments that T. gabrieli is attracted to racemic E-fuscumol. This is the fourth Tetropium species demonstrated to be attracted by E-fuscumol. Tetropium castaneum, T. fuscum and T. cinnamopterum, all with spruces as the main host tree species, are significantly attracted to the combination of E-fuscumol and host tree volatiles, and males of the latter two species emit S-fuscumol (Silk et al., 2007;Sweeney et al., 2010). These results suggest that T. gabrieli males may also emit fuscumol, but this must be confirmed by volatile collection and chemical analysis.
The effects of the host volatiles, α-pinene and ethanol, on trap catches varied among species. The addition of (−)-α-pinene plus ethanol to either unbaited traps or traps baited with E-fuscumol significantly increased catches of T. castaneum. The combination of E-fuscumol and host volatiles appeared to synergize attraction of T. castaneum, similar to what Sweeney et al. (2010) observed using a different blend of host volatiles that included α-pinene, ethanol, and four other monoterpenes. In contrast, neither T. gabrieli nor T. fuscum were attracted to (−)-α-pinene plus ethanol by themselves, and adding them to traps baited with E-fuscumol reduced catch of T. gabrieli and had no effect on catch of T. fuscum. These results suggest that T. gabrieli and T. fuscum respond to additional host-specific volatiles or volatile blends when searching for hosts. This was previously demonstrated for T. fuscum by Sweeney et al. (2004) who showed that trap catch of T. fuscum was significantly lower in traps baited with ethanol plus racemic α-pinene than in traps baited with ethanol plus a blend of five monoterpenes that approximated volatiles emitted from the boles of stressed spruce trees. Although there is variation among and within Larix species in the relative amounts of terpenes in leaves and twigs, common major components include (−)-and (+)-α-pinene, (−)-β-pinene, (+)-3-carene, and myrcene (Von Rudloff, 1987;Holm & Hiltunen, 1997;Ruuskanen et al., 2007;Kajos et al., 2013). Future research should compare the attraction of T. gabrieli to a synthetic blend of major monoterpenes emitted from larch trees (i.e., a larch blend) plus ethanol versus the (−)-α-pinene plus ethanol used in this study. These host volatile treatments should be also be tested with and without E-fuscumol and compared to E-fuscumol alone to determine whether the host volatiles synergize attraction of T. gabrieli to fuscumol.
Traps baited with host volatiles plus E,Z-fuscumol acetate caught significantly fewer individuals of T. fuscum and T. castaneum than did traps baited with host volatiles and E,Z-fuscumol, suggesting E,Z-fuscumol acetate is not attractive to these species. Lack of attraction of T. fuscum to fuscumol acetate is consistent with males not producing this compound (Silk et al., 2007;Sweeney et al., 2010); however, volatiles emitted by males of T. castaneum and T. gabrieli have not yet been determined. Catches of T. gabrieli in traps baited with host volatiles plus E,Z-fuscumol acetate were very low but did not differ significantly from catches in traps baited with host volatiles plus E,Z-fuscumol or E-fuscumol. However, to confirm whether or not E,Z-fuscumol is attractive to these species, one would have to compare catches in traps baited with E,Z-fuscumol acetate versus host volatiles versus E,Z-fuscumol acetate plus host volatiles versus unbaited traps.
There was no difference in catches of any of the Tetropium species between traps baited with E,Z-fuscumol plus host volatiles versus traps baited with E-fuscumol plus host volatiles, suggesting that the presence of Z-fuscumol did not inhibit response to E-fuscumol. This is further reinforced by the fact that the release rate of the attractive E isomer from the E,Z fuscumol lures was only about half of that from the E-fuscumol lures, because both lures were loaded with the same volume of E,Z-fuscumol or E-fuscumol. Flaherty et al. (2019) observed that traps baited with E,Z-fuscumol, E,Z-fuscumol acetate and ethanol captured significantly more T. fuscum and T. cinnamopterum than traps baited with monochamol, ipsenol, α-pinene and ethanol, whereas the converse was true for T. castaneum, suggesting the optimal combination of host volatiles and pheromone differs among these species.
The total catch (all experiments combined) of T. gabrieli was considerably lower than that of the other two Tetropium species. This is not surprising because T. gabrieli is monophagous on larch (Gorius, 1955) and larch stands constitute only 0.1% (0.9 million m 3 ) of the total growing stock (i.e. all tree species) in southern Sweden compared with 48.5% (431 million m 3 ) for Norway spruce (Anonymous, 2016).
The positive correlation between T. castaneum and T. fuscum may be explained by the two species sharing the same host tree species Norway spruce. This result is in agreement with an earlier study conducted on these two species in Poland and on T. fuscum and T. cinnamopterum (also sharing the same host tree species) in Canada . Because the experimental stands only included larch, all trapped T. castaneum and T. fuscum must have dispersed from colonized spruces in the surrounding landscape which means that the correlation in catches was not caused by the presence of previously colonized breeding substrates in the immediate vicinity of the traps. In contrast, there was no correlation between catches of T. gabrieli and those of T. castaneum and T. fuscum, and this was expected because of differences in host tree species.
An earlier study conducted in eastern Canada demonstrated that the exotic T. fuscum had a somewhat earlier flight period than the native T. cinnamopterum (Rhainds et al., 2010). The median date of captures was >2 weeks earlier for T. fuscum. The present study demonstrates that also in its native range T. fuscum is an early occurring Tetropium species. In Sweden, the median date of captures was about 2 weeks earlier for T. fuscum compared with the native T. castaneum and the exotic T. gabrieli. Thus, T. fuscum should have an advantage compared with T. castaneum regarding colonization of the generally limited amounts of wind-felled or weakened standing spruce trees in the managed forests of Sweden.
The only Tetropium species emerging from the larch stem sections was T. gabrieli. This is in accordance with an earlier Swedish study, including stem sections from 17 trees in 6 stands, in which also only T. gabrieli emerged (Lindelöw et al., 2015). Most T. gabrieli emerged during a relatively short time period after the breeding material was brought in to the climate room. Emergence started after about 2 weeks and after 4 weeks 90% of the beetles had already emerged. This is similar to what have been demonstrated for T. fuscum in Canada (Rhainds et al., 2010). The synchronized emergence is in accordance with the fact that most of the flight activity occurred during May and June. Tetropium gabrieli most commonly overwinters as a mature pre-pupal larva, though a proportion of the population may overwinter as immature larvae (Crawshay, 1907;Schimitschek, 1929). This is in accordance with the variation in larval head-capsule width found in the present study for hibernating larvae of which the smaller ones may represent individuals emerging and flying later in the season.
The three species of parasitoids that emerged from the stem sections have all earlier been demonstrated to use T. gabrieli as a host (Kenis & Hilszczanski, 2004). They are also commonly associated with the native T. castaneum and T. fuscum. All three species have earlier been recorded from Sweden and are classified as native (www.dyntaxa.se). The rate of parasitism in the present study is low compared with earlier studies on Tetropium (see Kenis & Hilszczanski, 2004 for an overview). It is highly unlikely that this was a result of the methodology used because the Tetropium larvae in the cut larch trees were exposed to parasitoids during the whole summer before the stem sections were cut in September. The two most common parasitoids we collected, R. xoridiformis and H. dentator, are endoparasitic koinobionts, i.e., females oviposit in the early instar host larva which continues to develop to a mature larva or prepupa before the parasitoid consumes the host, emerges to spin a cocoon and pupates. Both R. xoridiformis and H. dentator overwinter as pupae and emerge as adults the following spring, shortly after emergence of Tetropium adults (Schimitschek, 1929). The median emergence of R. xoridiformis and H. dentator in our study was only 5 and 15 days, respectively, after median emergence of T. gabrieli, providing evidence that these species likely parasitize early instar larvae. The ectoparasite, Neoxorides collaris, develops externally on parasitized larvae and overwinters as a pupa between the bark and the wood (Kenis & Hilszczanski, 2004). Thus, exposure of fresh larch logs in May and removal of logs from the field in September would have provided ample opportunity for parasitism by these species.
Tetropium gabrieli seems to be able to establish, and continuously maintain populations, in forest landscapes with only a low proportion of larch stands (4.2% in the present study area) and with few wind-felled larch trees. In an earlier study, conducted during 2 years (2009 and 2010) in the same area as the present study, T. gabrieli colonized most cut larch trees (Lindelöw et al., 2015). In the present study, it colonized wind-felled trees in 2016 (assuming the Tetropium larval galleries belonged to T. gabrieli, one adult found under bark in one of the trees) and cut and wind-felled trees in 2017. Factors explaining the ability to maintain populations in landscapes with low proportions of host trees may be a high production of adults per tree, a synchronized emergence and flight activity, and a strong attraction to their pheromones ensuring a high probability of mate finding. Low levels of interspecific competition and rate of parasitism should have contributed to the high production. No other bark-and wood-boring insects, except a few U. gigas, utilized the stems of the larch trees in this study and only a few in the earlier study (Lindelöw et al., 2015).
The results demonstrate that E-fuscumol and E,Z-fuscumol can be used for early detection of T. gabrieli in areas where the species has not yet established. However, more research is needed to develop a synthetic blend of host volatiles that is more representative of those emitted by freshly felled larch trees than the α-pinene and ethanol used in this study, and to determine the enantiomer or enantiomers of fuscumol emitted by T. gabrieli. Combining E-fuscumol with an attractive 'larch blend' host volatile lure plus ethanol may result in even greater attraction of T. gabrieli, in much the same way as the combination of spruce blend, ethanol, and E-fuscumol synergizes attraction of T. fuscum. Trap catches would also likely be improved by coating trap surfaces with Fluon to reduce friction (Allison et al., 2011;Graham & Poland, 2012;Allison et al., 2016). Even though T. gabrieli is considered to be more of a secondary pest species in Europe, it may become more aggressive if encountering new host tree species, for example, different larch species in North America, as was the case with T. fuscum when it established in North America. Furthermore, the results of the present study demonstrate that the species is able to also establish in regions where its habitat is strongly fragmented and constitutes only a very small proportion of total forest cover which is a strong argument for including this species in early detection programs. In addition, the bait may also be useful for monitoring population levels in areas where it is already established and for documenting the spread in areas where it has recently established.