Introduction

The Subarctic is experiencing increasingly frequent warm weather; for example, western Hudson Bay air temperatures have increased by 2–3 °C over the past 50 years (Gagnon and Gough Reference Gagnon and Gough2005), and near Churchill, Manitoba, Canada the decade between 1997 and 2007 saw six of the 10 hottest years on record (and all of the top five) (Fernandez-Triana et al. Reference Fernandez-Triana, Smith, Boudreault, Goulet, Hebert and Smith2011). These temperature changes may be linked to faunal shifts observed over the past half century (Fernandez-Triana et al. Reference Fernandez-Triana, Smith, Boudreault, Goulet, Hebert and Smith2011).

In 1969, Robert E. Gregg (Reference Gregg1972) sampled ants (Hymenoptera: Formicidae) near the tree-line in northern Manitoba in the vicinity of the town of Churchill. He collected at three sites along a small north-south transect (~6–7 km) that featured sites of exposed tundra and bluffs immediately along Hudson Bay (near the mouth of the Churchill River), boggy tundra, and boggy boreal forest. The ant community that he uncovered was not diverse but was abundant in some cases near the southern end of his investigation. In total, Gregg found five ant species in the Churchill region: Camponotus herculeanus (Linnaeus) (Formicinae), Formica neorufibarbus algida Wheeler (Formicinae), Leptothorax canadensis kincaidi Pergande (Myrmicinae), Leptothorax canadensis Provancher (Myrmicinae), and Myrmica brevinoidis Emery (Myrmicinae). Francouer (Reference Francouer1983) re-examined Gregg’s specimens, now held at the Field Museum of Natural History in Chicago (Illinois, United States of America), and noted that the specimens originally identified as L. canadensis kincaidi were L. acervorum Fabricius, and M. brevinoidis were actually M. alaskensis Wheeler.

As climate changes across the world, the importance of comparing contemporary collections to any measure of “historical” material is critical – as it allows us to determine whether there have been shifts in species composition (Tingley and Beissinger Reference Tingley and Beissinger2009). Documenting the presence or absence of any biotic change is particularly important when collections occur on large ecological ecotones since the isotherms that determine these areas are predicted to move drastically in the coming decades and, further, since the effects of a changing climate are unlikely to be felt in the same manner by all taxa. For example, in Churchill, some groups have seen a substantial change (parasitoid Hymenoptera (Braconidae) – (Fernandez-Triana et al. Reference Fernandez-Triana, Smith, Boudreault, Goulet, Hebert and Smith2011)), while others were not significantly different across collection periods (Diptera – (Renaud et al. Reference Renaud, Savage and Roughley2012b)). In 2012, we collected ants from the same locations visited by Gregg in 1969 to determine if the ant community had changed in the intervening time.

Methods

As part of a field course in Arctic Ecology offered by the University of Guelph (Guelph, Ontario, Canada) through the Ontario Universities Program in Field Biology, the instructors and students (22 people total) used descriptions from Gregg (Reference Gregg1972) (including his Fig. 2) to revisit the exact sampling locations that Gregg had sampled (1 site/day between 9 and 11 August 1969) on 10 July 2012 (south to north: Goose Creek (bog), Welcome Sign (tundra), and Cape Merry (tundra)) (Fig. 1). Due to descriptions and aspects of the local environment that were likely to be invariant between collection times (e.g., intersections of railway and road, peninsula) we were confident in sampling the same sites that Gregg did in 1972. At each collection locality we actively searched for ants using standard myrmecological collection methods (actively turning woody debris and stones, breaking twigs, pulling apart moss and lichen hummocks, and sampled ants using feather weight forceps and aspirators) for three hours within an area bounded by ~100×100 m. In addition to the ant fauna, we recorded the non-grass vascular plants and documented the habitat at each site using high-resolution, panoramic photographs captured using a GigaPan (www.gigapan.org) system (Smith et al. Reference Smith, Adamowicz, Boyd, Britton-Foster, Cahill and Desnoyers2013). The resultant panoramic photographs are publically available (http://www.gigapan.com/galleries/9086/gigapans).

Fig. 1

Ant sampling localities in the Churchill region. Yellow triangles represent the Gregg, 1969 sampling locations with numbers corresponding to site names in Table 1. Grey circles are the ant collections made in 2005/2006, while black circles are the 2010 ant collections that were analysed in Stahlhut et al. (Reference Stahlhut, Fernández-Triana, Adamowicz, Buck, Goulet and Hebert2013).

Collected ants were separated into genera and then morphospecies. Ant species were identified using available taxonomic resources (Creighton Reference Creighton1950; Francoeur Reference Francoeur1973, Reference Francouer1983, Reference Francouer1997; Fisher and Cover Reference Fisher and Cover2007). From these, we selected representatives to be DNA barcoded (~10/morphospecies). These individuals were then point-mounted, photographed, and a single leg (the point-side middle leg) was removed during preparation to serve as template for the total genomic extraction. DNA extractions were prepared from this tissue using a glass-fibre extraction protocol (Ivanova et al. Reference Ivanova, DeWaard and Hebert2006), and the DNA was resuspended in 30 μL of dH₂O. We amplified the DNA barcode region of cytochrome c oxidase I (COI) (a 658-base pair region near the 5′ terminus of the COI gene) using the insect primers LepF1 (5′- ATTCAACCAATCATAAAGATATTGG -3′) and LepR1 (5′- TAAACTTCTGGATGTCCAAAAAATCA-3′) and uni-directionally sequenced the amplicons using LepR1. All laboratory information associated with the individual sequences can be retrieved from the Barcode of Life Data System (BOLD, www.barcodinglife.org (Ratnasingham and Hebert Reference Ratnasingham and Hebert2007)) using the Process ID (sequence accessions). All sequence data and detailed collection information is available on BOLD in the public dataset: Arctic Ecology Student Projects 2012 (dx.doi.org/10.5883/DS-AEDNA12) and in Table 1.

Table 1

Two incidence matrices comparing Gregg’s 1969 collections and our July 2012 resampling from the same locations in and around Churchill, Manitoba.

Note: Matrices are presented separately since, without barcoding specimens from the historic collections, we are uncertain which BIN-divergent record to align with which morphologically determined historic collection.

We assessed the degree to which our sampling effort approached the estimated site maximum for alpha diversity (based on the techniques described above) using the incidence matrices in Table 1 and a sample-based rarefaction completed using EstimateS and 1000 randomisations (v9.1; Colwell and Elsensohn Reference Colwell and Elsensohn2014).

Results

When we revisited the Churchill sites sampled by Gregg in the summer of 1969 we located nests of all species in all collection sites. Using morphological and genetic methods, we found evidence for seven species of ants (Table 1 and Fig. 2). Two of these species were members of species complexes known to be widely distributed across the Nearctic (F. neorufibarbis and Leptothorax muscorum Nylander). In each case we found evidence for multiple DNA barcode clusters, suggesting the presence of multiple species under the same name.

Fig. 2

Accumulation curve for ant collections made in Churchill in 2012 at sites originally collected by Gregg in 1969. Dashed lines represent the 95% confidence interval, and grey bars represent the standard deviation of the expected number of species (S(est)) based on 1000 simulations in EstimateS (v9.1).

Most of the seven species were found in all three sites. Exceptions were Myrmica alaskensis and F. neorufibarbis (BOLD:AA1470) (found only in the most southerly bog site) and two of the Leptothorax species (L. muscorum complex and L. acervorum) (each found only at one or both of the two more northerly tundra-like sites) (Table 1).

Sampling the broader Churchill area (over the years 2005–2010 yielded three additional species compared to our geographically restricted Gregg resampling (Fig. 3). As in the species set recovered in the 2012 Gregg resampling, the new species from across the wider Churchill area were also DNA barcode divergent and not apparently morphologically different, and likely further members of the F. neorufibarbis and L. muscorum complexes as well as one further species of Myrmica Latreille.

Fig. 3

Accumulation curve for all ant collections made in Churchill (2005–2012). Dashed lines represent the 95% confidence interval, and grey bars represent the standard deviation of S(est) based on 1000 simulations in EstimateS (v9.1).

Although sample sizes varied among species (n ranged from 1 to 303), Wolbachia infection was detected in all species, with incidence rates varying between 16% and 100% (Table 2).

Table 2

Incidence of Wolbachia in Churchill ants determined using the wsp polymerase chain reaction assay.

Note: All species tested were found to contain the endosymbiont.

Discussion

In 1969 Gregg documented the occurrence of five species of ants in the vicinity of Churchill, Manitoba – the same ant species that would be expected to be found near the tree-line in Canada (Francouer Reference Francouer1983). Using the integration of traditional methods based on morphology and DNA barcoding, we report seven species present at the same sites Gregg sampled 40 years earlier. We conclude that the apparent increase is more likely due to morphologically cryptic species (that are genetically divergent) rather than the arrival of species from the south. However, it should be noted that since we were unable to DNA barcode the Gregg collections we are unable to conclusively exclude the possibility that the “new” BINS recovered in 2012 were also present 43 years earlier. Gregg (Reference Gregg1972) found only a single worker of the carpenter ant, Camponotus herculeanus, in the exposed tundra near Cape Merry, but no nests. This differentiates our study as we located nests of all species in all collection sites.

The apparent increase we report in ant diversity at Churchill is more likely to be due to changes in systematics (Francouer Reference Francouer1983; Heinze Reference Heinze1993) and the resolution of species-level diversity afforded by DNA barcoding (Smith et al. Reference Smith, Rodriguez, Whitfield, Deans, Janzen and Hallwachs2008, Reference Smith, Fernandez-Triana, Roughley and Hebert2009) than to the intervening arrival of more southerly distributed species (Fernandez-Triana et al. Reference Fernandez-Triana, Smith, Boudreault, Goulet, Hebert and Smith2011). The two new species recorded in our 2012 collection, one Formica Linnaeus and one Leptothorax Mayr species, are members of species complexes that are difficult to identify morphologically. Furthermore, we did not detect any species that others have predicted could soon be added to the northern ant fauna due (in part) to increased human activity in the area (Francouer Reference Francouer1997) (e.g., Tapinoma sessile (Say)).

Conclusions

The ant fauna in the subarctic locality surrounding Churchill, Manitoba does not appear to have significantly changed since the area was surveyed in the late 1960s. The fauna is dominated by species known to be characteristic of the tree-line and tundra environment in Canada and is characterised by taxonomically difficult species complexes. As a locality with a long history of hosting scientific inquiry, there is great value in continuing to regularly survey the insect and arthropod fauna in and around the Churchill region, as the response of this diverse community is unlikely to be simple.