Introduction
The western spruce budworm (WSBW), Choristoneura occidentalis (Freeman) (Lepidoptera: Tortricidae), has become a forest pest species of concern in Alberta, Canada, due to its apparently recent presence within the province and its rapid population increase since initial detection. Defoliation of stands dominated by Douglas-fir, Pseudotsuga menziesii subspecies glauca (Beissner) Murray (Pinaceae), in the Porcupine Hills of southwestern Alberta was first documented in 2005 and the defoliator responsible was provisionally identified as WSBW (Sustainable Resource Development 2005). By 2008, over 20 000 ha of forest in and around the Porcupine Hills had moderate levels of defoliation (Sustainable Resource Development 2009). Populations of WSBW have also become established in neighbouring low elevation regions of Alberta where white spruce, Picea glauca (Moench) Voss (Pinaceae), and a recognised hybrid spruce, Picea englemannii Parry ex. Engelmann (Pinaceae) × P. glauca, are dominant conifers (Strong and Hills Reference Strong and Hills2006). Strong and Hills (Reference Strong and Hills2006) referred to this hybrid as Picea albertiana Brown subspecies ogilviei Strong and Hills. However, it is unknown whether WSBW develops differently on its alternate hosts in this area, or whether hybrid spruce facilitates WSBW movement between Douglas-fir and non-Douglas-fir dominant forests.
The range of WSBW is largely congruent across western North America with that of Douglas-fir, its primary host (Stehr Reference Stehr1967; Harvey Reference Harvey1985). In Alberta, the Porcupine Hills have the most extensive Douglas-fir stands in the province and are the northeastern limit of the historical range of WSBW (Harvey Reference Harvey1985; Shepherd et al. Reference Shepherd, Gray and Harvey1995). However, southwestern Alberta is a putative hybrid zone between several budworm species (Shepherd et al. Reference Shepherd, Gray and Harvey1995; Lumley and Sperling Reference Lumley and Sperling2011a), which puts the species identity of the outbreaking populations into question.
At least eight species comprise the spruce budworm (SBW), Choristoneura fumiferana (Clemens) (Lepidoptera: Tortricidae) species complex (Harvey Reference Harvey1985; Volney and Fleming Reference Volney and Fleming2007; Lumley and Sperling Reference Lumley and Sperling2010), and their ranges overlap in transitional regions between major forest types characterised by their coniferous hosts (Stehr Reference Stehr1967; Volney and Fleming Reference Volney and Fleming2007). In Alberta, the range of SBW extends across the northern boreal forest, where white spruce is dominant, and southward along the foothills where it contacts WSBW (Lumley and Sperling Reference Lumley and Sperling2011a). Isolated populations of these two species also occur in the Cypress Hills, on the boundary between Alberta and Saskatchewan (Lumley and Sperling Reference Lumley and Sperling2011b). A third species, the two-year cycle budworm, Choristoneura biennis Freeman (Lepidoptera: Tortricidae), feeds on Engelmann spruce and subalpine fir, Abies lasiocarpa (Hooker) Nuttall (Pinaceae). It contacts the ranges of both SBW and WSBW (Stehr Reference Stehr1967), but is separated from them by the occurrence of its primary hosts at higher elevations and its obligatory second diapause during the fourth larval instar (Harvey Reference Harvey1967; Nealis Reference Nealis2005). At least two pine-feeding species, Choristoneura pinus Freeman and Choristoneura lambertiana (Busck) (Lepidoptera: Tortricidae) also occur in parts of central and southern Alberta.
Since SBW species display significant phenological synchrony with their hosts (Volney and Cerezke Reference Volney and Cerezke1992), which is reinforced by selection on extensive variation in the timing of post-diapause larval emergence of overwintered second instar larvae (Volney et al. Reference Volney, Waters, Akers and Liebhold1983; Thomson et al. Reference Thomson, Shepherd, Harris and Silversides1984; Shepherd Reference Shepherd1985), developmental differences can be expected to result from any establishment on alternative hosts with different phenological schedules. This may be more pronounced in areas with hybrid populations, where selection on such traits can be elevated relative to pure populations (Volney et al. Reference Volney, Waters, Akers and Liebhold1983). In our study, we compare the timing of the development on Douglas-fir and hybrid spruce by Choristoneura populations responsible for the recent outbreak in southwestern Alberta, and we use a combination of molecular techniques to determine their species identity.
Materials and methods
Study sites and sampling design
Choristoneura larvae, pupae, and adults were sampled at five locations in the Crowsnest Pass region of southwestern Alberta (Fig. 1). Two 45 cm branch tips were pruned from the mid-crown of a single Douglas-fir or hybrid spruce tree at each location, with sampling occurring between 20 May and 8 August 2009 at 8–10-day intervals. Species identity of the trees was determined using seed cone morphology (Farrar Reference Farrar1995; Strong and Hills Reference Strong and Hills2006). All larval instars were collected into 95% ethanol and were delimited by measuring head capsule width to the nearest 10 μm and developing instar separation criteria using the program HCAP (Logan et al. Reference Logan, Bentz, Vandygriff and Turner1998). Pupae were collected live and reared to eclosion. Average observed instar (AI obs = ∑pii where pi is the proportion of WSBW in each instar i [Lysyk Reference Lysyk1988]) was calculated for each site and tree species and sampling date. Adult flight period was monitored between 12 July and 30 September 2009 by deploying two Multi-Pher I insect traps (Services Bio-Contrôle, Ste-Foy, Québec, Canada) with eastern SBW flex lures (Contech, Victoria, British Columbia, Canada) and Vaportape II™ (Hercon, Emigsville, Pennsylvania, United States of America) at each site. Cumulative proportion of male moth trap catch throughout the flight period was arcsine square-root transformed to approach normality. The effects of tree species and site on the relationship between AI obs and day of year, and transformed cumulative proportion of male moths and day of year, were tested using general linear models testing the homogeneity of slopes, e.g., AI obs = day of year + site + day of year × site in R 2.15.1 (R Development Core Team 2012).
Fig. 1 Southwestern Alberta, Canada sample sites showing larval hosts (hybrid spruce, Douglas-fir), and elevation. Photograph is from Google Maps (http://maps.google.com), with the addition of major highways and location of the study area within Alberta.
Budworm species identity
To verify the species identity of budworms at each site, we used two types of molecular markers: mitochondrial (mt) DNA and microsatellites. We sequenced a 470 bp region of the cytochrome oxidase I (coI) mt gene in 41 adults (eight to nine specimens per location) using the methods of Sperling and Hickey (Reference Sperling and Hickey1994) and Lumley and Sperling (Reference Lumley and Sperling2010). Four adults collected at pheromone traps and four to five reared from larvae were used. Their coI sequences were then related to 169 previously published coI haplotypes (Sperling and Hickey Reference Sperling and Hickey1994; Roe and Sperling Reference Roe and Sperling2007; Lumley and Sperling Reference Lumley and Sperling2010, Reference Lumley and Sperling2011a, Reference Lumley and Sperling2011b), using maximum parsimony and likelihood as in Lumley and Sperling (Reference Lumley and Sperling2011a).
For microsatellites, we genotyped each of the 41 individuals at eight previously published loci (Lumley et al. Reference Lumley, Davis and Sperling2009) and analysed the data as in Lumley and Sperling (Reference Lumley and Sperling2011a, Reference Lumley and Sperling2011b) using structure v2.3.2 (Pritchard et al. Reference Pritchard, Stephens and Donnelly2000). Data for the 1135 specimens of Lumley and Sperling (Reference Lumley and Sperling2011a), encompassing all known North American SBW species, was used to train the admixture model with locations. Genotypes for the 41 new southwestern Alberta specimens were then used to infer assignment probabilities for each of six previously defined populations, corresponding to C. fumiferana, Western A, Western B, C. lambertiana, C. pinus, and Choristoneura retiniana genetic clusters (Lumley and Sperling Reference Lumley and Sperling2011a).
Results
A total of 698 larvae, 149 pupae, and 7189 male moths were collected over 132 days (18 sampling times; Table 1). Larval abundance was highest at Beaver Mines and lowest at Porcupine South, Alberta with 239 and 104 larvae/pupae collected in total from each site, respectively (Table 1). The average observed instar (AI obs) at each sampling time and site is shown in Figure 2 along with the cumulative proportion of male moths throughout the flight period at each location. The relationship between AI obs and ordinal date did not differ significantly among sites (F 4,32 = 0.2856; P = 0.8851) or hosts (F 1,38 = 1.223; P = 0.2758).
Table 1 Sample size of larval/pupal and adult (male) Choristoneura collections by ordinal date.
Note: Parentheses denote adult collections that occurred concurrently with larval collections.
Fig. 2 Developmental progression of Choristoneura populations at five sites in southwestern Alberta, Canada. Left side shows observed average instar for each site and the regression line for all sites combined, with two to six for larval instars and seven for pupae. Right side shows arcsine square-root transformed cumulative proportion of male moth abundance per collection date at each site with the regression line for all sites combined. The right axis is presented back-transformed. Horizontal axis shows ordinal date.
Male moth abundance was highest at Coleman (2553 moths) and lowest at Porcupine South (870 moths; Table 1). However, when abundance is represented as the cumulative proportion of moths relative to the total caught per site to account for potential differences in population size (Fig. 2), there are no significant differences in the relationship between male moth flight phenology and ordinal date between sites (F 4,35 = 0.3981; P = 0.8086) or hosts (F 1,41 = 0.3533; P = 0.5555).
Species identity
Maximum parsimony and likelihood analyses of new mtDNA sequences combined with previously published haplotypes gave trees with very similar topologies and the same major lineages as in previous studies (Sperling and Hickey Reference Sperling and Hickey1994; Lumley and Sperling Reference Lumley and Sperling2010, Reference Lumley and Sperling2011a, Reference Lumley and Sperling2011b). Thirty-nine o-lineage and two oβ-lineage coI haplotypes were found among the 41 specimens from southwestern Alberta (Table 2). Three new haplotypes (o94, o95, and o96; GenBank accession numbers KC172836:KC172838) differed by at least 1 bp from known haplotypes. No f-lineage haplotypes, characteristic of C. fumiferana, were sampled at the five locations.
Table 2 Choristoneura collection site, method of collection (adult reared from larva or pheromone trapped), and combined genetic profile (microsatellite population/mtDNA co1 haplotype (sample size)).
Note: Previously named haplotypes are as in Sperling and Hickey (Reference Sperling and Hickey1994) and Lumley and Sperling (Reference Lumley and Sperling2010, Reference Lumley and Sperling2011a, Reference Lumley and Sperling2011b).
*New mtDNA co1 haplotypes, Genbank accession numbers KC172836:KC172838.
mtDNA, mitochondrial DNA; coI, cytochrome oxidase I.
Microsatellite analyses were also congruent with previous work (Lumley and Sperling Reference Lumley and Sperling2011a, Reference Lumley and Sperling2011b). Each of the 41 specimens from southwestern Alberta was assigned to the previously known genetic clusters (Table 2). Most were assigned to Population 2 (Western A, n = 26) or Population 3 (Western B, n = 12) genetic clusters, with approximately equal numbers of reared and pheromone-trapped specimens included in each cluster. The three remaining individuals were assigned to Population 1 (C. fumiferana, n = 3), with two from Coleman (one reared on Douglas-fir and one pheromone-trapped) and one from Maycroft (pheromone-trapped). However, each of these three individuals had o-lineage mtDNA, and assignment probabilities to the Population 1 microsatellite cluster were not high (44%, 61%, and 72%). Although the Western A and B genetic clusters defined by Lumley and Sperling (Reference Lumley and Sperling2011a) contained species in addition to C. occidentalis that are potentially present in southwestern Alberta, none of our larval specimens exhibited the second diapause characteristic of C. biennis, nor were any larvae taken from pines or adults from the pheromone lures normally associated with C. lambertiana (Harvey Reference Harvey1985).
Discussion
DNA evidence as well as life history traits (lack of second diapause and type of pheromone attraction), indicate that the Choristoneura populations responsible for severe defoliation of coniferous forests in southwestern Alberta are best identified as WSBW. Although the presence of this species in Alberta is not new (Harvey Reference Harvey1985; Shepherd et al. Reference Shepherd, Gray and Harvey1995), it does represent the first major outbreak of WSBW in the province (Sustainable Resource Development 2007). Aerial surveys conducted concurrently with the work described in this study estimated more than 30 000 ha of forest with moderate to severe levels of defoliation in the Porcupine Hills and surrounding areas (Sustainable Resource Development 2010). This outbreak declined precipitously to just 2900 ha in 2010 (Sustainable Resource Development 2011), but larval populations remain high in restricted areas on Douglas-fir and hybrid spruce.
Our study is also the first report of WSBW feeding freely as larvae on an unconventional conifer host, hybrid spruce, within the province. However, feeding on this alternative host to Douglas-fir had no apparent impact on rate of development. We failed to find any significant effect of site or host on the relationship between AI obs and ordinal date. Second instar larvae were first observed in late May (day 140). Late instar larvae and pupae were found up to the date of the last larval collection at the beginning of August (day 220). Similarly, the cumulative proportion of male moths caught throughout the sampling period did not differ between sites or hosts. Male moths were first observed on 21 July (day 201), but may have been present as early as 12 July (day 193) when the pheromone traps were first deployed and pupae were first observed. Peak densities of male moths were observed between 17 and 26 August (days 229–238) and moths were still present in low densities at four of the five sites at the end of September (day 273).
The use of hybrid spruce as a larval host by WSBW is not unexpected, since the patchy landscape of the cordilleran region of western North America may predispose WSBW populations to tolerate variation in host-tree species (Volney Reference Volney1985). In fact, WSBW are known to feed on firs, Douglas-fir and spruces (Fellin and Dewey Reference Fellin and Dewey1982; Sanders Reference Sanders1991), in addition to one other white spruce × Engelmann spruce hybrid, which has long been identified as a host for budworms located in the hybrid zone between white spruce and Engelmann spruce (Stehr Reference Stehr1967). However, oviposition preferences by gravid females suggest that firs and Douglas-fir are the favoured hosts (Leyva et al. Reference Leyva, Clancy and Price2000, Reference Leyva, Clancy and Price2003).
The relatively labile feeding behavior of WSBW raises the possibility that hybrid stands may facilitate budworm movement between stands of more favourable hosts. The “hybrid-bridge” hypothesis, for instance, proposes that hybrid host trees may act as physiological intermediates between parental plant species and allow sustained host-shifts by herbivores (Floate and Whitman Reference Floate and Whitham1993). A current example is the expansion of the mountain pine beetle, Dendroctonus ponderosae Hopkins (Coleoptera: Curculionidae), onto Jack pine, Pinus banksiana Lambert (Pinaceae), in Alberta via hybrids between Jack pine and the closely related lodgepole pine, Pinus contorta Douglas (Pinaceae) (Cullingham et al. Reference Cullingham, Cooke, Dang, Davis, Cooke and Coltman2011). Although there is no evidence that WSBW is experiencing a similar host shift in southwest Alberta, it is noteworthy that its populations are surviving on hybrid spruces that not only bridge the region between boreal and montane forests along the eastern slopes of the Rocky Mountains, but also range far into Canada's interior (Strong and Hills Reference Strong and Hills2006).
We also found evidence for hybridisation among WSBW and SBW in the populations surveyed. Although all 41 individuals analysed for microsatellite characteristics had some degree of admixture between the six predefined genetic clusters, three individuals were not assigned to genetic clusters associated with WSBW. Instead these three individuals were assigned to the genetic cluster associated with SBW. All individuals also had o- or oβ-lineage haplotypes, indicating that hybridisation between these two species would have resulted from matings between male SBW and female WSBW. These results are highly concordant with recent reports of genetic introgression among budworm species whose ranges overlap in Alberta (Lumley and Sperling Reference Lumley and Sperling2011a, Reference Lumley and Sperling2011b). They are also supported by mating experiments that suggest introgression between WSBW and SBW may be most likely to occur via a male SBW, since egg weights from that combination are greater than the reciprocal cross (Harvey Reference Harvey1997). Previous studies have shown that hybrid populations can rapidly adapt to divergent developmental patterns of different host species (Volney et al. Reference Volney, Waters, Akers and Liebhold1983; Volney Reference Volney1999); however, the survival rates of hybrids have not been examined for southwestern Alberta.
In summary, WSBW is successfully establishing itself at the edge of its range in southwestern Alberta. Populations on Douglas-fir and hybrid spruce in the Porcupine Hills and surrounding areas do not differ developmentally, at least at the scale of this study. Fine-scaled phenology models and within-site comparisons would shed light on the developmental consequences of different host associations, and should eventually benefit the management of these populations. Moreover, investigation of the genetic structure of the hybrid zone between WSBW and SBW, and any correlations with hybridisation between white spruce and Engelmann spruce, may provide useful insight into the evolutionary and ecological characteristics of the SBW species complex.
Acknowledgements
The authors thank Dan Lux of Alberta Sustainable Resource Development for facilitating a contract to complete this study; R. Hewison and G. Bloom (Alberta Sustainable Resource Development) for operational support; J. Dombroskie, S. Brunet, and B. Proshek for assistance with sampling; J. Logan for making HCAP available; B. Mori and C. Whitehouse for statistical assistance; and J. Dupuis and other Sperling lab members, and two anonymous reviewers for helpful criticisms and suggestions on earlier drafts. This work was supported by a Natural Sciences and Engineering Research Council of Canada Discovery Grant to F. Sperling and a Queen Elizabeth II graduate scholarship to B. Brunet.
