Sympatric species are expected to minimize competition by partitioning resources, especially when these are limited. Herbivores inhabiting the High Arctic in winter are a prime example of a situation where food availability is anticipated to be low, and thus reduced diet overlap is expected. We present here the first assessment of diet overlap of high arctic lemmings during winter based on DNA metabarcoding of feces. In contrast to previous analyses based on microhistology, we found that the diets of both collared (Dicrostonyx groenlandicus) and brown lemmings (Lemmus trimucronatus) on Bylot Island were dominated by Salix while mosses, which were significantly consumed only by the brown lemming, were a relatively minor food item. The most abundant plant taxon, Cassiope tetragona, which alone composes more than 50% of the available plant biomass, was not detected in feces and can thus be considered to be non-food. Most plant taxa that were identified as food items were consumed in proportion to their availability and none were clearly selected for. The resulting high diet overlap, together with a lack of habitat segregation, indicates a high potential for resource competition between the two lemming species. However, Salix is abundant in the winter habitats of lemmings on Bylot Island and the non-Salix portion of the diets differed between the two species. Also, lemming grazing impact on vegetation during winter in the study area is negligible. Hence, it seems likely that the high potential for resource competition predicted between these two species did not translate into actual competition. This illustrates that even in environments with low primary productivity food resources do not necessarily generate strong competition among herbivores.
Citation: Soininen EM, Gauthier G, Bilodeau F, Berteaux D, Gielly L, Taberlet P, et al. (2015) Highly Overlapping Winter Diet in Two Sympatric Lemming Species Revealed by DNA Metabarcoding. PLoS ONE 10(1): e0115335. https://doi.org/10.1371/journal.pone.0115335
Academic Editor: Ulrich Melcher, Oklahoma State University, UNITED STATES
Received: August 25, 2014; Accepted: November 21, 2014; Published: January 30, 2015
Copyright: © 2015 Soininen et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited
Data Availability: Bryophyte reference library and lemming diet data are deposited at Dryad digital repository (datadryad.org/), doi: 10.5061/dryad.4rr39. All other relevant data are available as Supporting Information files.
Funding: This work was financed by Oskar Huttunen foundation (PhD scholarship to EMS), the Research Council of Norway (Leiv Eiriksson scholarship to EMS and the BarFrost project FRIBIO 191627/V40 to CB), the University of Tromsø (travel grant for EMS), the Natural Sciences and Engineering Research Council of Canada, the Canadian federal government’s International Polar Year program (project MD-021) and the Northern Student Training Program, both administered by Indian and Northern Affairs Canada, the Canadian Network of Centres of Excellence ArcticNet and the Polar Continental Shelf Program. SPYGEN provided support in the form of a salary for author EB, but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific role of this author is articulated in the ‘author contributions’ section.
Competing interests: Ludovic Gielly and Pierre Taberlet are co-inventors of a patent concerning g-h primers and the subsequent use of the P6 loop of the chloroplast trnL (UAA) intron for plant identification using degraded template DNA. The patent has the following numbers: CA 2581347 (Canada Patent), 2006/040448 (PCT Patent), EP1797201 (EPO Patent), 20090081646 and 20110143354 (both United States Patent Application). These patents, titled “Universal primers and their use for detecting and identifying plant materials in complex mixtures”, only restrict commercial applications and have no impact on the use of this locus by academic researchers. Eva Bellemain is employed by SPYGEN. This does not alter the authors’ adherence to the PLOS ONE policies on sharing data and materials
Closely related species living in sympatry are expected to reduce overlap in resource or habitat use to minimize competition [1–4]. Among small mammals, several species typically coexist, and thus the potential for competition is high. This should be especially true when resources are limited such as in low-productivity environments. In desert rodents, strong competitive interactions have indeed been found [5, 6]. However, for tundra rodents, a group living in another low-productivity environment, evidence for competitive interactions has been more equivocal [2, 7–10].
Lemmings and voles are dominant herbivores in the circumpolar tundra [11–13]. These small mammals are year-round residents and thus must survive on local primary production throughout the year. Lemmings are well-known for their regular, large-amplitude population cycles [12, 14, 15]. Every three to five years, populations reach very high densities, which can sometimes severely impact their food resources [16–18]. The impact of lemming grazing may be especially pronounced under the snow during the long Arctic winter because individuals tend to concentrate in restricted areas, such as those with deep snow (i.e. snow beds, [19–21]), and because no new plant growth occurs during this period. However, resource use by lemmings during winter remains poorly known due to the formidable challenges involved in studying them beneath the snow under the harsh Arctic conditions. Yet, events occurring during the winter, such as reproduction, may play a key role in lemming population dynamics [22–26].
It is common that two species of lemmings live in sympatry in the Arctic. Typically, when this occurs, one species belongs to the genus Lemmus and the other to the genus Dicrostonyx. These two genera tend to have different ecologies, including different habitat and food preferences. Lemmus generally prefers wetter habitats whereas Dicrostonyx prefers drier habitats [27–29]. However, habitat segregation may be less clear in winter. In Arctic Canada, both genera tend to concentrate in deep snow patches that, on the treeless tundra, are often limited to areas where topography is conducive to high snow accumulation such as the leeward side of slopes [20, 21]. Generally, Lemmus feed primarily on monocots and, to a lesser extent, mosses, whereas Dicrostonyx feed mostly on dicots [8, 19, 30–32]. Although broad dietary patterns have been generally consistent among study sites and show relatively little diet overlap between these two genera, some variability has been found [8, 32]. Differences in diet among localities have been mainly attributed to variations in forage availability, suggesting some flexibility in resource use of lemmings ([8, 32], see also ). However, whether dietary overlap between coexisting species increases during winter in areas where both species concentrate in snow beds such as in Arctic Canada remains unknown.
Previous studies that examined diets in lemmings were based on microhistological identification of plant fragments in stomach contents or fecal pellets (hereafter called traditional methods). Recently, a new method, DNA metabarcoding, has become available to study animals’ diets (reviewed in ). This method is based on amplifying and high-throughput sequencing a standardized DNA region from feces or stomach contents, and subsequently identifying and counting the taxa composing the diet by comparing the obtained sequences to a taxonomic reference library [34–36]. It has lately been used successfully to study diets of various herbivores such as brown bear (Ursus arctos) , golden marmot (Marmota caudata) , chamois (Rupicapra rupicapra) , European bisons (Bison bonasus)  and small rodents [40, 41]. Compared to traditional methods, DNA metabarcoding generally provides finer taxonomic resolution, has the potential to identify more taxa, and can allow the analysis of a large number of samples without observer biases [35, 40].
We here present the first analysis of collared (Dicrostonyx groenlandicus) and brown (Lemmus trimucronatus) lemmings’ winter diets based on DNA metabarcoding of fecal pellet contents. Considering that these two species use similar habitats during winter in the Canadian High Arctic , we hypothesized that they should minimize interspecific competition by showing little interspecific diet overlap. Our hypothesis appeared strong because the two species are thought to have different dietary preferences . We also examined how plant availability affected diet and quantified food selection and diet diversity of each species. We finally discuss the consequences of our findings for species interactions in Arctic food webs.
The study site is located in the Qarlikturvik glacial valley (73°08’N, 80°00’W) of Bylot Island, Sirmilik National Park, Nunavut Territory, Canada. The study area (40 km2) consists of tundra polygons, thaw lakes and ponds forming wetlands at the bottom of the valley and is surrounded by mesic tundra on higher ground and nearby slopes and hills. Mesic tundra accounts for most of the landscape whereas wetlands cover about 23% of the study area and dry, xeric tundra with a sparse vegetation cover occupies <5% of the area and is limited to the tops of hills and ridges . Wetlands have extensive grass/sedge meadows dominated by mosses and graminoids (Dupontia fisheri, Eriophorum scheuchzeri and Carex aquatilis; ). The mesic tundra is dominated by prostrate shrubs (Salix arctica, S. herbacea, Cassiope tetragona, Dryas integrifolia; erect shrubs are very scarce), with a sparse cover of forbs (Saxifraga spp., Potentilla spp., Ranunculus spp., Pedicularis spp.), graminoids (Arctagrostis latifolia, Alopecurus alpinus, Poa spp., Luzula spp.), mosses and lichens. Small, intermittent streams running through upland areas often create gullies but their floristic composition is generally similar to the surrounding mesic tundra . Plant names follow the nomenclature and taxonomy of the Annotated Checklist of the Panarctic Flora  for vascular plants and bryophyte floras of Arctic Canada [46–48] for bryophytes (mosses and liverworths). Nomenclature of the bryophyte reference library follows tropicos.org .
The most important herbivores present on the island are the two lemming species, which are present throughout the year, and the greater snow goose (Chen caerulescens atlantica), which is present only in summer. No other herbivorous small mammal is present and large mammalian herbivores are absent. Arctic hares (Lepus arcticus) and Rock ptarmigans (Lagopus mutus) are present in very small numbers. During winter, the two lemming species share the same habitats and concentrate in mesic tundra, especially in the small gullies along streams where snow accumulates [20, 21]. Based on characteristics of vegetation and topography, we recognize three habitats for wintering lemmings: mesic tundra, stream gullies and wetlands.
Collection of lemming pellets
We sampled lemming winter nests across the study area shortly after snowmelt in 2011, following a winter of high lemming density . In each of the three habitats, 20 transects, each 500 m long, were distributed randomly. All winter nests (n = 347) found along transects were collected and their habitat and position recorded. More details of the winter nest sampling are given in . Additional winter nests were collected from a systematic search of three grids (7 to 11 ha each) used for summer live-trapping of lemmings (n = 327 nests) and from winter nest boxes (n = 10 nests) described by . In this study, we used a subset of 74 of those winter nests; 55 from transects, 9 from trapping grids and 10 from nest boxes.
Lemming species using winter nests were identified based on the size, shape and color of feces found in nests [20, 52]. This visual identification was based on a sample of 15 pellets from each winter nest and was confirmed by DNA analysis (see below). Some nests (12 out of 74) had been used by both species, and from two of these nests we sampled pellets of both species. The final sample size was thus 76 samples (n = 22 collared lemmings and 54 brown lemmings). Pellets were dried in a filter bag placed in silica gel.
Lemming species identification using genetic methods
We took a random sample of three to five pellets from each 15-pellet sample used for visual species identification. To verify the accuracy of the genetic identification, we analyzed 6 muscle samples of each lemming species using the same methods. These samples were provided by concurrent studies on stable isotopes [3, 53]. We extracted DNA of these samples, i.e. both pellets and muscle, using methods described in  and in Supporting information (S1 Text). The same DNA extracts were then used both for lemming species identification and for diet analysis.
For both lemming species, we first downloaded from GenBank all the available mitochondrial Cytochrome Oxydase I (COI) sequences of the standard barcode for animals . After aligning the retrieved sequences, we calculated a consensus sequence for each species. We then identified two locations where the lemming species differ by two consecutive nucleotides, with 61 nucleotides in between these locations. We designed a pair of primers specific for each species, by locating the 3′-end of each forward and reverse primers on these two different consecutive nucleotides (S1 Table). Each primer for brown lemmings was tagged with an additional 10 base pairs poly-A on the 5′-end to differentiate the two lemming genera by amplicon size (104 bp for collared and 125 bp for brown).
The DNA extracts were amplified in a 40 µl volume reaction containing 2.5 mM MgCl2, 0.2 mM of each dNTP, 0.25 µM of each primer and 0.8 U AmpliTaq Gold DNA Polymerase (Life). After 10 min at 95°C, PCR reactions were performed for 35 cycles of 30 sec in 95°C, 30 sec in 58°C and 30 sec in 72°C. Amplicons were checked and amplicon sizes estimated on the QIAxcel System (QIAGEN). See S1 Fig. for an example of a capillary electrophoresis output from the program.
Lemming diet analysis using DNA metabarcoding
We analysed the vascular plant and bryophyte content of the DNA extracts of lemming pellets using DNA metabarcoding. The method is based on first amplifying a targeted plastid DNA region (trnL (UAA) intron) using universal primer for plants, and thereafter high-throughput DNA sequencing [40, 55]. We used two complementary primer pairs, g-h and c-h [55, 56]. The g-h primer pair gives precise taxonomic results for small rodent diets  but is biased towards seed plants. To assess also the abundance of bryophytes in lemming diets, we used primer pair c-h, which is universal for all plant taxa. Details of the DNA analysis are given in  and in Supporting information (S1 Text).
Sequence reads were analyzed using the OBITools software package (http://metabarcoding.org/obitools/doc/index.html). As reference for the primer pair g-h, we used a combined reference library of 815 arctic  and 835 north boreal  vascular plant species. Sequences with poor match with these reference libraries were compared with data retrieved from the EMBL Nucleotide Sequence Database (version 111, available at http://www.ebi.ac.uk/embl/). For the c-h primer pair, we used the same taxonomic reference library of arctic and boreal vascular plant species, supplemented with a new library of 455 arctic and boreal bryophyte species (see details below and a detailed list of taxa in S2 Table). For both primer pairs, the retrieved taxon lists were compared with the local flora of Bylot Island (Benoit Tremblay, in prep). Among all sequences identified at the family or genus level, 0.2% belonged to taxa absent from the site (e.g. Pinus, Picea, Betula). These were considered identification errors or contamination and removed from the dataset. At the species level, some identified species were absent from the general area but in most cases, a closely related species was known to be present but not included in our reference libraries. In those cases, we assigned the sequences to the species known to be present at the site. Details of sequence cleaning and annotation are described in  and in Supporting information (S1 Text).
Reference library of arctic and boreal bryophytes
Sampling for construction of the bryophyte taxonomic reference library was carried out in two museum collections (the Bryophyte herbaria at the Norwegian University of Science and Technology (TRH) and V. L. Komarov Botanical Institute (LE)), and the selected specimens were checked by taxonomic experts. For most species, we sampled two specimens originating from different parts of the species distribution area to cover possible intraspecific sequence variation. All DNA extracts are preserved in the DNA Bank of the Natural History Museum, University of Oslo, Norway.
Approximately 5 mm2 of dried leaf tissue was ground in 2.0 mL tubes with tungsten carbide beads for 2 min at 15 Hz in a mixer mill (MM301, Retsch). A Gene Mole extraction robot was used to extract the DNA using the MoleStripTM Plant DNA kit. Amplification and sequencing of the P6 loop of the trnL intron was performed using the c and d primers . PCRs (10 μL) contained 3μL 1:10 diluted DNA, 0.4 μM of each primer (), 1 mM dNTPs, 0.1% bovine serum albumin, 2.5 mM MgCl2, 1x PCR buffer and 0.4 UAmpliTaq DNA polymerase (Applied Biosystems). After 10 min at 95°C, PCR reactions were performed using 30 cycles of 30 sec in 95°C, 30 sec in 50°C and 2 min in 72°C, followed by a final extension of 5 min at 72°C. PCR products were sequenced in both directions on an ABI 3730 sequencer. Quality checking and cleaning of the library was performed by comparing all sequences to published sequences with NCBI/BLAST, and by carrying out phylogenetic analyses including sequences from closely related taxa to verify taxonomic identity. In some cases, new specimens were selected and sequenced after the first round of library cleaning. Formatting of the reference library, including annotations of the sequences, was carried out using the OBITools. The final library was formatted by in silico PCR on the obtained sequences (using the program ecoPCR; ), with the trnL c and trnL h primers (five mismatches allowed between primer and the target sequence).The library is deposited in the Dryad Digital Repository (http://datadryad.org/), doi: 10.5061/dryad.4rr39.
To assess plant availability, we used the data of  (data given in S3 Table). In their study, plant biomass was sampled at the end of the 2010 growing season (early August) in snowbeds (n = 16) where signs of lemming use had been found in previous years. Plant biomass data thus represent what was available for the lemmings at the onset of winter. Snowbeds were sampled in two different habitats; stream gullies (n = 8) and mesic tundra (n = 8). The snowbeds were separated from each other by at least 50 m. Within each snowbed, one quadrate (20 × 50 cm) was located randomly and plant availability was estimated. Vascular plant biomass was measured by clipping all aboveground vascular plant material at the ground level. Dead material was removed and the remaining live material was sorted by family, genus or species, dried to constant mass at 45°C, and weighted. Moss proportion cover (to the nearest 5%) was visually estimated for each genus or species; all aboveground live (i.e. green) material was clipped, dried to constant mass at 45°C and weighed. Total biomass was divided by the surface area of the sampled quadrates and multiplied by the proportion estimates for each taxon.
All necessary permits were obtained for the described field work on Bylot Island which is within the Sirmilik National Park (Parks Canada permit #SIR-2011–8213). No protected species were sampled.
The resulting datasets of lemming diets consisted of a count of sequences per plant taxon per pellet sample. We combined the information obtained from the two primer pairs as follows. We first calculated the proportion of each plant family for every sample (sample representing pellets of a given species from a given winter nest) based on the number of DNA sequences identified, for each primer pair separately. For the g-h primer pair, we discarded the few bryophyte sequences identified and retained only the vascular plants (including ferns) for those calculations. We then weighted (i.e. multiplied) the proportion of each vascular plant family of each sample in the g-h dataset by the corresponding, overall proportion of vascular plants determined with the c-h primer pair. Finally, we combined those weighted proportions of vascular plant families with the proportions of moss families determined with the c-h primer pair. For one collared lemming sample, amplification by the g-h primer pair failed and we thus used data from the c-h primer for proportion of vascular plant families. Even though DNA metabarcoding data for plants probably reflects small rodent diets well , some biases may still occur [34, 40] and we therefore also report the number of samples in which a given taxon was found (S4 Table).
We tested for differences in proportion of various food items (i.e. monocots, dicots and plant families accounting for >2% of the diet) between habitats (three levels) and lemming species (two levels) using ANOVAs. Data were rank-transformed before analysis because variances were heterogeneous and residuals were not normally distributed. Interactions between habitat and species were examined with the aligned rank transformation  but were never significant (P > 0.05) and are thus not reported.
We assessed diet overlap between the lemming species using Schoener’s diet overlap index . The index varies between zero and one, zero indicating no diet overlap and one indicating complete overlap. We calculated index values at plant family level in two different ways; 1) including all plant families and 2) excluding Salicaceae as they dominated the diets (see results). To assess diet diversity, we calculated an index of trophic niche width using the Shannon entropy (denoted hereafter as TNW) (equation given in ).
We further assessed whether intraspecific diet overlap (i.e. among individuals of the same species) differed between lemming species, using the individual specialization index (IS), which is an extension of the Schoener’s overlap index , assuming that pellets found in different nests originated from different individuals. To ensure that the index values of the two species were comparable in spite of different sample size, we resampled 100 times a random sample of brown lemming individuals corresponding to the sample size of collared lemmings (n = 22). We then compared the simulated IS for brown lemmings (i.e. average of the resampled IS) to their observed IS (i.e. acquired by including all individuals). As these values were identical, we did not use the resampled data in further analyses. Finally, we tested whether the lemming species differed in terms of IS using ANOVA. Diet overlap and diversity analyses were done in the software R 3.0.3. , and package RInSp  was used for the within species analysis.
We evaluated food selection by combining data on diet composition and plant availability. We used the selection ratio of Manly , which is the ratio of mean proportion of food item i in the diet over mean availability of that item. We calculated variance using the formulas of Manly for a situation where both use and availability were sampled. Because plant availability was only estimated in stream gullies and mesic tundra habitats, we excluded lemming pellet samples collected in wetlands from food selection analysis. Pellet samples from the two remaining habitats were pooled because diet did not differ between these two habitats (see results) and availability is similar . Sample sizes for this analysis were 47 for brown lemmings and 20 for collared lemmings. Vascular plants and mosses were sampled using different techniques, and thus availability is not comparable between these two groups. Selection was therefore analyzed separately for vascular plants and mosses; hence, availability and use sum up to 1 within each of these taxonomic groups. We excluded Ericaceae, an abundant vascular plant family not consumed by lemmings (see results), from the calculations of plant availability because its inclusion would have biased selection for all other vascular plant families towards positive.
Bryophyte reference library
For the bryophyte taxonomic reference library, sequences covering the complete P6 loop in the chloroplast trnL (UAA) intron were obtained from 498 specimens representing two divisions: Bryophyta (mosses) and Marchantiophyta (liverworts). The library includes 18 orders of mosses, 45 families, 137 genera, and 340 species. Liverworts were represented by 86 species from 45 genera, 28 families and three orders (see S2 Table, for a complete taxon list).
Lemming species identification
We were able to test genetically the lemming species identified in the field based on pellet size, shape and color for 74 of our 76 pellet samples (54 brown and 20 collared lemmings). The genetic identification was based on the difference between amplicon size. The amplicone sizes estimated by the QIAxcel System were on average 128 bp for collared lemming and 146 bp for brown lemming. While these were longer than presumed (see methods), the relative difference remained. (S1 Fig.). The genetic analysis confirmed field species identification in 98.6% of the cases. Only one pellet sample identified as brown lemming in the field turned out to be a collared lemming according to the genetic analysis. All muscle samples (n = 12) were identified to the correct species.
Taxonomic precision of lemming diet data
A total of 45,633 sequences were obtained with the g-h primer pair (608 sequences/sample on average) and 22,707 with the c-h primer pair (299 sequences/sample on average). Overall, 99.5% of the sequences were identified at the family level, 32.6% at the genus level and 7.5% at the species level. The low resolution at genus and species levels was largely caused by Salicaceae, a common family in our samples (see results) for which the g-h primer pair has in general low resolution . Excluding this family for the g-h primer pair, 72.7% and 16.7% of the sequences were identified at the genus and species levels, respectively.
For collared/brown lemmings, we collected 8/15 pellet samples in mesic tundra, 12/22 in stream gullies and 2/7 in wetlands and 0/10 in unknown habitat. The proportion of monocotyledons/dicotyledons (and mosses for brown lemmings) did not differ between mesic tundra and stream gullies for either of the species (collared, p > 0.329; brown, p > 0.291; wetland excluded due to small sample size) or for any individual plant family that we examined (collared, p > 0.08; brown, p > 0.06 for all tests); therefore, habitats were pooled for subsequent analyses.
The winter diet of collared lemmings was composed of 86% dicotyledons, 14% monocotyledons and <1% mosses (n = 22). In comparison, the diet of brown lemmings was composed of 65% dicotyledons, 9% monocotyledons and 26% mosses (n = 54). Even though the diet of brown lemmings had less dicotyledons than the one of collared lemmings (F1,74 = 21.4, p < 0.001), there was a high overlap in diet composition (overlap index = 0.75) between the two species with respect to these two broad groups of vascular plants.
At the family level, diets of both lemmings were clearly dominated by Salicaceae (Fig. 1). Among vascular plants, Poaceae was also found to be relatively abundant in both species diets. Even though collared lemmings consumed more Salicaceae (F1,74 = 28.8, p < 0.001) and Poaceae (F1,74 = 8.9, p = 0.004) than brown lemmings (n = 54) overall, we noted again a high overlap in vascular plant families (overlap index = 0.67) between the two species. Salicaceae is represented at the site only by the genus Salix, which was thus identified as the dominant food item of both lemming species. See Table 1 for the full list of genera and species identified and S4 Table for the number of samples in which a given taxon was found.
Diets (mean ± SE) of collared (n = 22) and brown lemming (n = 54) during the winter 2010–11 on Bylot Island, Nunavut, Canada, based on DNA metabarcoding sequences extracted from pellets. Vascular plant and moss families are separated by a dashed line. Other or unidentified items accounted for 0.5% and 1.2% of the collared and brown lemming diet, respectively (not shown on graph). Taxa without dot on the graph were not found. Families are Caryophyllaceae (Caryo), Fabaceae (Faba), Juncaceae (Junca), Poaceae (Poa), Polygonaceae (Polyg), Rosaceae (Rosa), Salicaceae (Salic), Saxifragaceae (Saxif), Aulacomniaceae (Aulac), Bryaceae (Bryac), Dicranaceae (Dicra), Grimmiaceae (Grimm), Polytrichaceae (Polyt) and Timmiaceae (Timm).
The high diet overlap between lemming species was mainly caused by the dominance of Salicaceae as the remaining proportions of taxa in diets differed between species (overlap index = 0.32). In addition to Salicaceae and Poaceae, the diet of collared lemming included 10 other vascular plant families and 3 moss families but these comprised only 4% of the diet (Fig. 1). In the diet of brown lemmings we found more families; in addition to Salicaceae and Poaceae, 12 vascular plant families and 10 moss families composed on average 37% of the diet (Fig. 1). Of these, the moss families Polytrichaceae and Aulocomniaceae and the vascular plant families Polygonaceae and Saxifragaceae were most prominent (Fig. 1). Thus, diet diversity of brown lemmings was higher in terms of numbers of families, but also when measured with the diet diversity index TNW; index value for collared lemmings was 0.60, while it was 1.60 for brown lemmings. On the other hand, the two lemming species did not differ significantly in terms of intraspecific diet specialization. Although the IS index was 0.64 for brown lemming and 0.85 for the collared lemming, this difference was not significant (F1,74 = 0.31, p = 0.58).
The most abundant vascular plant family, Ericaceae, accounted for 58% of the plant biomass in the winter habitats of lemmings on Bylot Island . However, it was not consumed by either lemming species and the sole species of this family present, Cassiope tetragona, was avoided; it was thus considered non-food for lemmings. Within the plant taxa that were eaten, no taxa were strongly selected or avoided. For collared lemming, Salicaceae was consumed in proportion to its availability, Poaceae tended to be selected for whereas Juncaceae, Rosaceae, and possibly Fabaceae were avoided (Table 2). For brown lemmings, Salicaceae was also consumed in proportion to availability and Juncaceae was avoided. Polygonaceae and Saxifragaceae had high selection ratios due to their very low availability but these were not significant due to the high variance. Among mosses, Aulacomniaceae and Dicranaceae had high selection ratios but these were not significantly different from 1, also due to their high variance (Table 1). Polytrichaceae, the most common moss family, tended to be selected whereas other abundant moss families like Amblystegiaceae, Hylocomiaceae and Scapaniaceae were not consumed.
Our study is the first to examine the winter diet of lemmings using DNA metabarcoding techniques, as all previous studies have relied on microhistological analysis. Due to this novel method, we were able to elucidate lemming winter diets at an unprecedented level of details. Our analysis of two sympatric lemming species revealed similarities with previous studies but also some startling differences. Interestingly, our results do not fit our prediction that these sympatric species should have clearly different winter diets, as the diet of both species showed a high degree of overlap. Diets of both lemming species were by far dominated by Salix and moss consumption was relatively low.
Lemming winter diets
On Bylot Island, mosses were barely consumed by collared lemmings and their winter diet was dominated by dicotyledons, in line with previous studies [8, 30, 32]. However, within dicotelydons, variable patterns of consumption have been found among studies. A dominance of Salix was found both in Northern Alaska and Northern Greenland [19, 30] but of Dryas at Pearce Point and Igloolik in Northern Canada [8, 32]. These dietary differences seem to largely reflect differences in availability among sites as willow was abundant and Dryas scarce in Alaska but the reverse was true at Pearce Point and Igloolik . Generally, Salix was consumed by collared lemmings in higher proportion than its availability or preferred in feeding trials [8, 31]. On Bylot Island, Salix was very abundant in snowbeds and consumed in proportion to its availability whereas Rosaceae (Dryas and Potentilla) were not very common and were actually avoided by collared lemmings. The abundant use of Salix on Bylot Island thus fits well with the previous observations on the use of Salix vs Dryas being determined by their availability.
Mosses were consumed by brown lemmings, but accounted for a lower proportion of their winter diet on Bylot Island (26%) than at Barrow, Alaska (40%; ) and Igloolik, Nunavut (>80%, ). In sharp contrast with previous studies, we found that dicotelydons (primarily Salix), instead of monocotyledons, dominated the winter diet of brown lemmings. Grasses and sedges were the dominant vascular plant food items eaten by brown lemmings in both summer and winter at all other sites [30, 32, 67] with Salix being a negligible component of their diet. Furthermore, feeding trials with captive animals have shown that brown lemmings find Salix rather unpalatable [31, 67]. On Bylot Island, grasses and sedges are abundant in the wet summer habitat of the brown lemmings, but scarce in their preferred winter habitat, i.e. snow beds in stream gullies . In contrast, Salix is very abundant in the stream gullies and the surrounding mesic tundra and its availability pattern may explain this unexpected result. A gradual switch to willows in fall may allow the digestive tract of brown lemmings to adapt to the relatively high content of secondary compounds present in Salix . This hypothesis is supported by the findings of Rodgers and Lewis  who noted that naïve animals born and raised in captivity consumed more shrubs (and especially Salix) than animals captured in the wild. Moreover, DNA metabarcoding analysis of the stomach contents of five individuals collected from Bylot Island indicates that Salicaceae and Rosaceae are important food items during the summer as well (Soininen and Gauthier, unpublished data). Finally, because microhistological methods have a tendency to overestimate monocotyledon proportions in diets [68, 69], their proportion may have been overestimated in previous studies. However, our findings reveal that brown lemming diets may be more flexible and spatially variable than previously believed.
In summary, we found indication that diet of both lemming species on Bylot Island is heavily affected by food availability, which adds to increasing evidence showing that availability is an important determinant of small rodent diets [3, 8, 41]. Furthermore, the large differences between locations revealed by our study may imply that both competitive interactions between lemmings species and lemming-vegetation interactions may vary greatly across the arctic tundra.
Species interactions and food web dynamics
The observed interspecific overlap index is clearly very high and the same level as within herbivore species in other studies . While herbivores may segregate diets also by means of habitat selection  and selection for different plant parts , this is unlikely to be the case for collared and brown lemmings. First, the species have similar winter habitat preferences . Second, small rodents share many characteristics of digestive morphophysiology and are thus unlikely to have preferences for different plant parts. Brown and collared lemmings hence present a high potential for exploitation competition during winter. Whether this potential translates into actual competition would depend on food limitation. Salix is abundant (forming up to 80% of non-Ericaceae biomass) on Bylot Island, and lemming winter grazing has a negligible impact on snowbed vegetation, even during a year of peak lemming abundance . This suggests that the high potential for exploitation competition is not currently expressed due to a lack of food limitation. This is further suggested by the low interspecific diet overlap of the non-Salix portion of the diet. Our results thus highlight that even in the High Arctic, food resources may be abundant enough for herbivores to cope with high diet overlap.
A recent analysis of the Bylot Island food web , shows that lemmings consume a very small proportion of the annual primary production. Our results suggest that Salicaceae could be exposed locally to heavy winter browsing by lemmings, especially during peaks in population density. On the other hand, little evidence has been found that lemming grazing during winter has a strong impact on Salix biomass in snowbed vegetation on Bylot Island . However, only total biomass was sampled by cutting plant material in the latter grazing impact experiment. Therefore, other potentially important effects of lemmings on Salix demography, such as mortality of new recruits [73, 74], was not quantified. Thus, the impact of overwintering lemmings on Salix demography could still be substantial in spite of a low effect on total biomass.
Recently, increasing growth of erect shrubs, and especially Salix, has been observed in response to climate warming in many locations throughout the circumpolar Arctic (reviewed in ). On the other hand, herbivory appears to be a factor limiting shrub encroachment in many areas [75, 76]. Small herbivores, such as voles and lemmings, can have a substantial impact on Salix shrubs as saplings can suffer up to 90% mortality in Finnmark, Northern Norway, in peak years of small rodent abundance  (V. T. Ravolainen unpublished data). Even though prostrate Salix species (S. arctica and S.herbaceae) dominate in the snowbeds used by lemmings at our study site, erect shrubs (S. richardsonii) are occasionally present. As snowbeds are generally favorable to the growth of erect shrubs [74, 77], the consumption of Salix by both lemming species present on Bylot Island is a factor that could potentially limit encroachment of erect shrubs in this habitat despite a warming climate .
Shape, size and color of fecal pellets collected in the field have been used as criteria to identify lemming species in previous studies when both brown and collared are present [20, 32, 52]. For the first time, we validated this technique using genetic techniques and showed that it was highly reliable (>98% correct identification). Thus, misidentification of lemming fecal pellets was not an issue in our study.
DNA metabarcoding of feces has been successfully used to describe diets of several large herbivores, including gazelles, chamois and brown bear [37, 38, 79]. Still, inference of the quantity of each ingested taxon from the number of DNA sequences retrieved should be done with some caution. Several potential biases, such as taxon-specific numbers of chloroplasts in the consumed tissue and differential digestion may influence the observed patterns . However, for stomach contents, results of food item proportions in small rodent diets gained by DNA metabarcoding correspond rather well to those gained by microhistological methods . Furthermore, Willerslev et al.  have recently demonstrated that DNA metabarcoding results of sheep rumen content corresponded well to known proportions in their diets. As previous studies on small rodents have used mainly stomach contents [33, 40, 41], no evaluations between food intake and DNA metabarcoding results of feces are published yet. Still, preliminary results of a study comparing small rodent stomach content and feces from the rectum of the same individual (n = 40) showed a good correspondence , indicating that differential digestion is among taxa unlikely to be a major issue in small rodents. In addition, the surprising abundance of Salicaceae in our results is unlikely to be an artifact. The DNA fragment amplified by the primer pair g-h for Salix is of no shorter length than for example the Poaceae genera we identified (Salix being 56bp and the grass genera 52–53bp). It is therefore unlikely that Salix DNA would have been better preserved during digestion than other taxa. We are thus confident that our results reflect actual diet proportions rather well.
The newly constructed bryophyte reference library comprising common arctic and north boreal species allowed us to achieve a high and reliable taxonomic resolution of the moss component of lemming diets. Mosses are a key plant group in the arctic ecosystems, both in terms of biomass and function [81, 82] and as a food item for many herbivores [33, 83, 84]. DNA-based identification of mosses species in the diets of arctic herbivores is currently developing [33, 41, 85], as microhistologic identification of moss species in diet samples is virtually impossible. Reliable reference data of bryophyte DNA is therefore essential especially for the Arctic. Existing public reference DNA databases such as GenBank provide more limited information on bryophyte taxa as compared to the vascular plant groups. Moreover, specimen identification errors, taxonomic complexities and discrepancies between different nomenclatures may lead to erroneous identification of the DNA sequence data. Local reference libraries, as used for both vascular plants [57, 58] and bryophytes in this study, are constructed based on material that is collected and verified by taxonomic experts, and archived and stored in museum collections for future reference. Hence, high quality of taxonomic assignment is ensured.
The taxonomic resolution achieved by the DNA metabarcoding method made it possible to assess lemming winter diet composition at an unpreceded level of detail. We can thus conclude that the two lemming species on Bylot Island show high diet overlap during winter and consequently have a high potential for competition for food. However, this potential apparently does not translate into actual competition because their main food item, Salix, is abundant, lemming grazing has little impact on the vegetation, and the non-Salix portion of diets overlaps little between the species. It seems therefore unlikely that the species would suffer from strong food competition in the focal system. Our results highlight that even during the long high arctic winter, plant food resources—in relation to their use—may be abundant enough for herbivores to allow for high diet overlap. Moreover, our study underlines that in order to understand interspecific resource competition, it is important to assess how patterns of use and availability are related.
S1 Text. Additional methodological details of the DNA metabarcoding analysis.
S1 Table. Primer pairs used for detecting lemming species.
Primer pairs and corresponding COI-fragment used for detection of lemming genera Lemmus and Dicrostonyx.
S2 Table. Bryophyte reference library.
Taxonomic content of the Arctic-boreal bryophyte reference library Version 1.0 with sequences of the short P6 loop of the trnL plastid region (available from the Dryad Digital Repository, http://datadryad.org/). Column “lib_refN” refers to the reference number in the library
S3 Table. Plant family availability data.
Plant biomass (g/m2) in August 2010 sampled on Bylot Island. In column “Plot ID” M refers to mesic habitat and R to river gully habitat. Plant family names are abbreviated (-ceae removed), in columns “Other” and “Total” v refers to vascular plants and b to bryophytes.
S4 Table. Plant families found in lemming diets.
Frequency of occurrence of plant families in the winter diets of collared and brown lemmings during the winter 2010–11 on Bylot Island based on DNA metabarcoding of pellets. For vascular plants, data are based on primer pair g-h, for mosses on primer pair c-h.
S1 Fig. An example of a capillary electrophoresis output from QIAxcel System.
Colums represent 12 different samples. Horizontal bands represent DNA fragments, numbers along the edge of columns show scale in bp length. Two band sizes can be seen along 150bp line, indicating samples of Lemmus (146bp, samples 1–4 and 11–12) and Dicrostonyx (128bp, samples 5–10).
We thank Carole Poillot for help with the laboratory work; Lucie Zinger and Frédéric Boyer for help in bioinformatics analyses; Patrick Morissette for help during the field work; Arnaud Tarroux, Marie-Helene Truchon and Elise Bolduc for help with the lemming muscle samples; Marie-Christine Cadieux for conducting some statistical analyses and assistance with logistics; Virve Ravolainen for discussions on Salix and Benoit Tremblay for his advice with plant taxonomy.
Conceived and designed the experiments: EMS G. Gauthier FB DB NGY LG PT. Performed the experiments: EMS FB LG PT. Analyzed the data: EMS G. Gauthier NGY. Wrote the paper: EMS G. Gauthier FB DB NGY LG PT G. Gussarova EB. Created the reference library for bryophytes: G. Gussarova EB CB KH HKS LE ASN.
- 1. Rosenzweig ML (1991) Habitat selection and population interactions—the search for mechanism. Am Nat 137: S5–S28.
- 2. Sundell J, Church C, Ovaskainen O (2012) Spatio-temporal patterns of habitat use in voles and shrews modified by density, season and predators. J Anim Ecol 81: 747–755. pmid:22325037
- 3. Soininen EM, Ehrich D, Lecomte N, Yoccoz NG, Tarroux A, et al. (2014) Sources of variation in small rodent trophic niche: new insights from DNA metabarcoding and stable isotope analysis. Isotopes Environ Health Stud 50: 361–381. pmid:24830842
- 4. Krebs CJ (2009) Ecology: the experimental analysis of distribution and abundance. San Francisco: Pearson Benjamin Cummings.
- 5. Brown JS (1989) Desert rodent community structure—a test of 4 mechanisms of coexistence. Ecol Monogr 59: 1–20.
- 6. Heske EJ, Brown JH, Mistry S (1994) Long-term experimental study of a chihuahuan desert rodent community—13 years of competition. Ecology 75: 438–445.
- 7. Galindo C, Krebs CJ (1985) Habitat use by singing voles and tundra voles in the Southern Yukon. Oecologia 66: 430–436.
- 8. Bergman CM, Krebs CJ (1993) Diet overlap of collared lemmings and tundra voles at Pearce Point, Northwest Territories. Can J Zool 71: 1703–1709.
- 9. Batzli GO, Lesieutre C (1995) Community organization of arvicoline rodents in Northern Alaska. Oikos 72: 88–98.
- 10. Saetnan ER, Gjershaug JO, Batzli GO (2009) Habitat use and diet composition of Norwegian lemmings and field voles in Central Norway. J Mammal 90: 183–188.
- 11. Krebs C, Danell K, Angerbjörn A, Agrell J, Berteaux D, et al. (2003) Terrestrial trophic dynamics in the Canadian Arctic. Can J Zool 81: 827–843.
- 12. Ims RA, Fuglei E (2005) Trophic interaction cycles in tundra ecosystems and the impact of climate change. BioScience 55: 311–322.
- 13. Gauthier G, Berteaux D, Bety J, Tarroux A, Therrien JF, et al. (2011) The tundra food web of Bylot Island in a changing climate and the role of exchanges between ecosystems. Écoscience 18: 223–235.
- 14. Gruyer N, Gauthier G, Berteaux D (2008) Cyclic dynamics of sympatric lemming populations on Bylot Island, Nunavut, Canada. Can J Zool 86: 910–917.
- 15. Krebs C (2011) Of lemmings and snowshoe hares: the ecology of Northern Canada. P Roy Soc B-Biol Sci 278: 481–489.
- 16. Moen J, Lundberg PA, Oksanen L (1993) Lemming grazing on snowbed vegetation during a population peak, Northern Norway. Arc Alp Res 25: 130–135.
- 17. Virtanen R (2000) Effects of grazing on above-ground biomass on a mountain snowbed, NW Finland. Oikos 90: 295–300.
- 18. Olofsson J, Tommervik H, Callaghan TV (2012) Vole and lemming activity observed from space. Nature Clim Change 12: 880–883.
- 19. Klein DR, Bay C (1994) Resource partitioning by mammalian herbivores in the High Arctic. Oecologia 97: 439–450.
- 20. Duchesne D, Gauthier G, Berteaux D (2011) Habitat selection, reproduction and predation of wintering lemmings in the Arctic. Oecologia 167: 967–980. pmid:21701915
- 21. Reid DG, Bilodeau F, Krebs CJ, Gauthier G, Kenney AJ, et al. (2012) Lemming winter habitat choice: a snow-fencing experiment. Oecologia 168: 935–946. pmid:22042523
- 22. Stenseth NC, Ims RA, editors (1993) The biology of lemmimgs. London: the Linnean Society of London by Academic Press.
- 23. Yoccoz NG, Ims RA (1999) Demography of small mammals in cold regions: the importance of environmental variability. Ecological Bulletins 47: 137–144.
- 24. Ims RA, Yoccoz NG, Killengreen ST (2011) Determinants of lemming outbreaks. PNAS 108: 1970–1974. pmid:21245340
- 25. Bilodeau F, Gauthier G, Berteaux D (2013) The effect of snow cover on lemming population cycles in the Canadian High Arctic. Oecologia 172: 1007–1016. pmid:23232938
- 26. Bilodeau F, Reid DG, Gauthier G, Krebs CJ, Berteaux D, et al. (2013) Demographic response of tundra small mammals to a snow fencing experiment. Oikos 122: 1167–1176.
- 27. Batzli GO, Pitelka FA, Cameron GN (1983) Habitat use by lemmings near Barrow, Alaska. Holarctic Ecology 6: 255–262.
- 28. Rodgers AR, Lewis MC (1986) Diet selection in arctic lemmings (Lemmus sibiricus and Dicrostonyx groenlandicus)—demography, home range, and habitat use. Can J Zool 64: 2717–2727.
- 29. Morris DW, Davidson DL, Krebs CJ (2000) Measuring the ghost of competition: insights from density-dependent habitat selection on the co-existence and dynamics of lemmings. Evol Ecol Res 2: 41–67.
- 30. Batzli GO, Pitelka FA (1983) Nutritional ecology of microtine rodents—food-habits of lemmings near Barrow, Alaska. J Mammal 64: 648–655.
- 31. Rodgers AR, Lewis MC (1985) Diet selection in arctic lemmings (Lemmus sibiricus and Dicrostonyx groenlandicus)—food preferences. Can J Zool 63: 1161–1173.
- 32. Rodgers AR, Lewis MC (1986) Diet selection in arctic lemmings (Lemmus sibiricus and Dicrostonyx groenlandicus): forage availability and natural diets. Can J Zool 64: 1684–1689.
- 33. Soininen EM, Zinger L, Gielly L, Bellemain E, Bråthen KA, et al. (2013) Shedding new light on the diet of Norwegian lemmings: DNA metabarcoding of stomach content. Polar Biol 36: 1069–1076.
- 34. Pompanon F, Deagle BE, Symondson WOC, Brown DS, Jarman SN, et al. (2012) Who is eating what: diet assessment using next generation sequencing. Mol Ecol 21: 1931–1950. pmid:22171763
- 35. Valentini A, Pompanon F, Taberlet P (2009) DNA barcoding for ecologist. Trends Ecol Evol 24: 110–117. pmid:19100655
- 36. Taberlet P, Coissac E, Hajibabaei M, Rieseberg LH (2012) Environmental DNA. Mol Ecol 21: 1789–1793. pmid:22486819
- 37. Valentini A, Miquel CN, M. A., Bellemain E, Coissac E, Pompanon F, et al. (2009) New perspectives in diet analysis based on DNA barcoding and paralle pyrosequencing: the trnL approach. Mol Ecol Resour 9: 51–60. pmid:21564566
- 38. Raye G, Miquel C, Coissac E, Redjadj C, Loison A, et al. (2011) New insights on diet variability revealed by DNA barcoding and high-throughput pyrosequencing: chamois diet in autumn as a case study. Ecol Res 26: 265–276.
- 39. Kowalczyk R, Taberlet P, Coissac E, Valentini A, Miquel C, et al. (2011) Influence of management practices on large herbivore diet—case of European bison in Bialowiez a primeval forest (Poland). Forest Ecol Manag 261: 821–828.
- 40. Soininen EM, Valentini A, Coissac E, Miquel C, Gielly L, et al. (2009) Analysing diet of small herbivores: the efficiency of DNA barcoding coupled with high-throughput pyrosequencing for deciphering the composition of complex plant mixtures. Front Zool 6:16. pmid:19695081
- 41. Soininen EM, Ravolainen VT, Bråthen KA, Yoccoz NG, Gielly L, et al. (2013) Arctic small rodents have diverse diets and flexible food selection PLoS ONE 8: e68128.
- 42. Hughes RJ, Reed A, Gauthier G (1994) Space and habitat use by greater snow goose broods on Bylot Island, Northwest Territories. J Wildlife Manage 58: 536–545.
- 43. Gauthier G, Rochefort L, Reed A (1996) The exploitation of wetland ecosystems by herbivores on Bylot Island. Geoscience Canada 23: 253–259.
- 44. Bilodeau F, Gauthier G, Fauteux D, Berteaux D (2014) Does lemming winter grazing impact vegetation in the Canadian Arctic? Polar Biol 37: 845–857.
- 45. Elven R, Murray DF, Razzhivin VY, Yurtsev BA (2011) Annotated checklist of the panarctic flora (PAF): vascular plants http://nhm2.uio.no/paf/ Natural History Museum, University of Oslo.
- 46. Brassard GR, Fife AJ, Webber J (1979) Mosses from Baffin Island, Arctic Canada. Lindbergia 5: 99–104.
- 47. Brassard GR (1976) The mosses of northern Ellesmere Island, Arctic Canada III. New or additional records. Bryologist 79: 480–487.
- 48. Brassard GR (1971) The mosses of northern Ellesmere Island, arctic Canada II. Annotated list of the taxa. Bryologist 74: 282–311.
- 49. Missouri Botanical Garden. Tropicos.org http://www.tropicos.org.
- 50. Bilodeau F, Gauthier G, Berteaux D (2013) Effect of snow cover on the vulnerability of lemmings to mammalian predators in the Canadian Arctic. J Mammal 94: 813–819.
- 51. Bilodeau F, Kenney AJ, Gilbert BS, Hofer E, Gauthier G, et al. (2013) Evaluation of a technique to trap lemmings under the snow. Arctic 66: 32–36.
- 52. Maclean SF, Fitzgerald BM, Pitelka FA (1974) Population cycles in arctic lemmings: winter reproduction and predation by weasels. Arc Alp Res 6: 1–12.
- 53. Tarroux A, Bety J, Gauthier G, Berteaux D (2012) The marine side of a terrestrial carnivore: intra-population variation in use of allochthonous resources by arctic foxes. PLoS ONE 7: e2427.
- 54. Hebert PDN, Cywinska A, Ball SL, DeWaard JR (2003) Biological identifications through DNA barcodes. P Roy Soc B-Biol Sci 270: 313–321.
- 55. Taberlet P, Coissac E, Pompanon F, Gielly L, Miquel C, et al. (2007) Power and limitations of the chloroplast trnL (UAA) intron for plant DNA barcoding. Nucleic Acids Res 35: e14. pmid:17169982
- 56. Taberlet P, Gielly L, Patou G, Bouvet J (1991) Universal primers for amplification of 3 noncoding regions of chloroplast DNA. Plant Mol Biol 17: 1105–1109. pmid:1932684
- 57. Sønstebø JH, Gielly L, Brysting A, Elven R, Edwards M, et al. (2010) Using next-generation sequencing for molecular reconstruction of past Arctic vegetation and climate. Mol Ecol Resour 10: 1009–1018. pmid:21565110
- 58. Willerslev E, Davison J, Moora M, Zobel M, Coissac E, et al. (2014) Fifty thousand years of Arctic vegetation and megafaunal diet. Nature 506: 47–51. pmid:24499916
- 59. Ficetola GF, Coissac E, Zundel S, Riaz T, Shehzad W, et al. (2010) An In silico approach for the evaluation of DNA barcodes. Bmc Genomics 11: 434. pmid:20637073
- 60. Salter KC, Fawcett RF (1993) The art test of interaction—a robust and powerful rank test of interaction in factorial models. Commun Stat Simul Comput 22: 137–153.
- 61. Schoener TW (1970) Nonsynchronous spatial overlap of lizards in patchy habitats. Ecology 51: 408–418.
- 62. Bolnick D, Yang L, Fordyce J, Davis J, Svanbäck R (2002) Measuring individual-level resource specialization. Ecology 83: 2936–2941.
- 63. Schoener TW (1968) Anolis lizards of Bimini—resource partitioning in a complex fauna. Ecology 49: 704–725.
- 64. R Core Team (2014) R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing.
- 65. Zaccarelli N, Bolnick DI, Mancinelli G (2013) RInSp: an R package for the analysis of individual specialization in resource use. Methods Ecol Evol 4: 1018–1023.
- 66. Manly BFJ, McDonald LL, Thomas DL, McDonald TL, Erickson WP (2002) Resource selection by animals: statistical design and analyses for field studies. Dordrecht: Kluwer Academic.
- 67. Batzli GO, Jung H-JG (1980) Nutritional ecology of microtine rodents: resource utilization near Atkasook, Alaska. Arc Alp Res 12: 483–499.
- 68. Vavra M, Holechek JL (1980) Factors influencing microhistological analysis of herbivore diets. J Range Manage 33: 371–374.
- 69. Parker DM, Bernard RTF (2006) A comparison of two diet analysis techniques for a browsing megaherbivore. J Wildlife Manage 70: 1477–1480.
- 70. Mysterud A (2000) Diet overlap among ruminants in Fennoscandia. Oecologia 124: 130–137.
- 71. Darmon G, Calenge C, Loison A, Jullien JM, Maillard D, et al. (2012) Spatial distribution and habitat selection in coexisting species of mountain ungulates. Ecography 35: 44–53.
- 72. Legagneux P, Gauthier G, Berteaux D, Bêty J, Cadieux M-C, et al. (2012) Disentangling trophic relationships in a High Arctic tundra ecosystem through food web modeling. Ecology 93: 1707–1716. pmid:22919916
- 73. Moncrieff GR, Chamaillé-Jammes S, Bond WJ (2014) Modelling direct and indirect impacts of browser consumption on woody plant growth: moving beyond biomass. Oikos 123: 315–322.
- 74. Ravolainen VT, Brathen KA, Yoccoz NG, Nguyen JK, Ims RA (2014) Complementary impacts of small rodents and semi-domesticated ungulates limit tall shrub expansion in the tundra. J Appl Ecol 51: 234–241.
- 75. Myers-Smith IH, Forbes BC, Wilmking M, Hallinger M, Lantz T, et al. (2011) Shrub expansion in tundra ecosystems: dynamics, impacts and research priorities. Environ Res Lett 6: 045509.
- 76. Plante S, Champagne E, Ropars P, Boudreau S, Levesque E, et al. (2014) Shrub cover in northern Nunavik: can herbivores limit shrub expansion? Polar Biol 37: 611–619.
- 77. Tape KD, Hallinger M, Welker JM, Ruess RW (2012) Landscape heterogeneity of shrub expansion in Arctic Alaska. Ecosystems 15: 711–724.
- 78. Gauthier G, Bêty J, Cadieux MC, Legagneux P, Doiron M, et al. (2013) Long-term monitoring at multiple trophic levels suggests heterogeneity in responses to climate change in the Canadian Arctic tundra. P Roy Soc B-Biol Sci 368: 20120482.
- 79. Ait Baamrane MA, Shehzad W, Ouhammou A, Abbad A, Naimi M, et al. (2012) Assessment of the food habits of the Moroccan dorcas gazelle in M’Sabih Talaa, West Central Morocco, using the trnL approach. PloS ONE 7: e35643. pmid:22558187
- 80. Soininen EM (2012) Interactions between small rodents and their food plants in tundra habitats (PhD thesis). Tromsø, Norway: University of Tromsø.
- 81. Prins HHT (1982) Why are mosses eaten in cold environments only? Oikos 38: 374–380.
- 82. Gornall JL, Jonsdottir IS, Woodin SJ, Van der Wal R (2007) Arctic mosses govern below-ground environment and ecosystem processes. Oecologia 153: 931–941. pmid:17618466
- 83. Soininen EM, Hubner CE, Jonsdottir IS (2010) Food selection by barnacle geese (Branta leucopsis) in an Arctic pre-breeding area. Polar Res 29: 404–412.
- 84. Ihl C, Klein DR (2001) Habitat and diet selection by muskoxen and reindeer in Western Alaska. J Wildlife Manage 65: 964–972.
- 85. Stech M, Kolvoort E, Loonen MJJE, Vrieling K, Kruijer JD (2011) Bryophyte DNA sequences from faeces of an arctic herbivore, barnacle goose (Branta leucopsis). Mol Ecol Resour 11: 404–408. pmid:21429154