Introduction
Intraspecific variation in diet or foraging behaviour among populations in different geographic locations can occur due to differences in external factors between locations, including abiotic factors such as temperature and biotic factors such as habitat type, prey availability, and interspecific competition (
Freckleton et al. 2003;
Kaupas and Barclay 2018). Among bats, populations of a single species in different locations may experience different prey availability, or individuals may selectively forage for prey that are more nutritionally or energetically beneficial under different environmental conditions, leading to geographical variation in diet (
Moosman et al. 2012;
Clare et al. 2014;
Aspetsberger et al. 2003).
The little brown bat (
Myotis lucifugus (Le Conte, 1831)) is a widespread insectivorous bat across North America (
Fenton and Barclay 1980). It typically uses aerial hawking to forage, although it also has the ability to glean prey (
Ratcliffe and Dawson 2003); and across its range, its diet varies (e.g.,
Belwood and Fenton 1976;
Clare et al. 2014;
Kaupas and Barclay 2018). In northern parts of its range (Alaska (USA), Yukon (Canada), Northwest Territories (Canada)),
M. lucifugus consumes spiders, in addition to its typical aerial prey (
Whitaker and Lawhead 1992;
Talerico 2008;
Shively 2016;
Kaupas and Barclay 2018). This varies seasonally and geographically. In Northwest Territories, cool temperatures in spring and late summer are associated with consumption of spiders (
Kaupas and Barclay 2018), while in Alaska, it consumes low quantities of spiders all season (
Whitaker and Lawhead 1992;
Shively 2016). The difference in timing of spider consumption is hypothesized to be due to interspecific competition (
Kaupas and Barclay 2018). In Northwest Territories, the northern long-eared bat (
Myotis septentrionalis (Trouessart, 1897)) also occurs and due to its lower wing loading and therefore greater maneuverability compared with
M. lucifugus (
Norberg and Rayner 1987;
Ratcliffe and Dawson 2003;
Kaupas and Barclay 2018) is able to glean more efficiently, which may allow it to consume spiders all season.
Myotis septentrionalis was not recorded in the Alaskan studies, suggesting that when both species are present, there may be niche partitioning between
M. lucifugus and
M. septentrionalis, such that
M. lucifugus forages for different prey than
M. septentrionalis when temperatures are warmer and aerial prey is abundant, and only resorts to consuming spiders when aerial prey abundance is reduced in colder temperatures (
Kaupas and Barclay 2018).
During colder temperatures, aerial insect abundance declines because insects are less able to fly (
Norberg 1978;
Belkair et al. 2018). Some spiders, however, are still active (e.g.,
Avery and Krebs 1984) or remain in the centre of their web (orb weavers (family Araneidae);
Dondale et al. 2003), which may make them available as prey to bats in colder temperatures.
Myotis lucifugus is abundant in the Kananaskis Valley of the Rocky Mountains in Alberta, Canada, where another long-eared bat species, the long-eared myotis (
Myotis evotis (H. Allen, 1864)), also occurs (
Barclay 1991). In that area, wing loading is significantly different between the two species (
Barclay 1991). Lower wing loading for
M. evotis suggests that it is more maneuverable than
M. lucifugus, and it is adept at gleaning stationary prey (
Faure and Barclay 1994).
Myotis evotis has been recorded consuming spiders in Idaho, USA (
Lacki et al. 2007), but temporal or spatial variation in consumption of spiders by
M. evotis have not been investigated.
Competition between
M. lucifigus and
M. evotis could lead to similar seasonal variation in spider consumption by
M. lucifugus, as seen in the Northwest Territories when
M. septentrionalis is present (
Kaupas and Barclay 2018). However,
M. lucifugus commonly forages over open water in the Kananaskis Valley, whereas
M. evotis forages along paths within the forest (
Barclay 1991), so different foraging habitat may also influence consumption of spiders.
We examined the consumption of spiders by
M. lucifugus and
M. evotis in the Kananaskis Valley, where mean temperatures early and late in the summer are low (i.e., below 10 °C;
Environment and Climate Change Canada 2020). Our goal was to gain a better understanding of the factors that influence consumption of spiders by bats, examining whether ambient temperature, time of year, and foraging habitat influence the diet of
M. lucifugus and
M. evotis.
We hypothesized that there is variation in spider consumption between
M. lucifugus and
M. evotis throughout the summer due to competition between them for insect prey, resulting in niche partitioning. We predicted that
M. evotis would consume spiders all season long, given its adaptations for gleaning and the amount of time it spends foraging near or within the foliage, especially in riparian zones (
Barclay 1991;
Faure and Barclay 1992). We predicted that
M. lucifugus would consume spiders at the beginning and end of the summer due to low aerial insect abundance associated with colder temperatures, while preying on aerial insects during the middle of the summer due to their abundance.
Materials and methods
Fifteen field sites were located in the Kananaskis Valley near the University of Calgary Biogeosciences Institute (BGI), Barrier Lake Field Station (51°01′39.9″N, 115°02′03.6″W; Supplementary Fig. S1
1) and we obtained temperature data from the weather station at the BGI Institute. Bats were captured with mist nets and harp traps (
Kunz et al. 2009) set across trails and roads. Captured bats were placed individually into cloth bags for an hour before they were processed to allow time to digest and excrete previously consumed prey. Sex, reproductive condition, age (based on the fusion of metacarpal–phalangeal joint epiphyses;
Brunet-Rossinni and Wilkinson 2009) and species was determined for each captured bat. Bats were captured and fecal samples were collected between May and September from 1985 to 1988 by R.M.R. Barclay, D. Solick, and P. Faure (personal communication), and in 2018 by D.G. Maucieri. Additionally, we hand fed spiders and moths to some bats to test whether they would consume them when offered. Fecal samples from hand-fed bats were not examined since bats did not produce fecal samples between time of hand feeding and time of release. All procedures were approved by the Life and Environmental Sciences Animal Care Committee at the University of Calgary.
Habitat type of capture sites was determined using a straight-line measure of distance on a satellite map, from the netting site to a lake, pond, or river. We classified sites as riparian areas (near water) if they were within 50 m of these water sources and non-riparian (far from water) if they were more than 50 m away.
One or two pellets from each fecal sample were weighed and samples were randomly numbered so that species and time of year was unknown when fecal analysis was conducted. Samples were softened with 70% ethanol and examined under a dissecting microscope. We identified insect parts to order and estimated their proportions in each sample (
McAney et al. 1991;
Whitaker et al. 2009). Spider tarsal claws were used as evidence of consumption of spiders (
Kaupas and Barclay 2018).
We compared our data on
M. lucifugus and
M. evotis with data from Northwest Territories (
Kaupas and Barclay 2018) for
M. lucifugus and
M. septentrionalis. All statistical analyses were performed using RStudio version 4.0.1 (
R Core Team 2020).
Due to non-normal data, we used a PERMANOVA (1000 permutations) to test the difference in diet (seven arthropod orders) between
M. lucifugus and
M. evotis from Kananaskis. We then used a Dufrêne–Legendre Indicator Species Analysis (DLISA) to determine which arthropod orders were indicators for each bat species (
Dufrêne and Legendre 1997;
Roberts 2016). DLISA identifies important orders without the use of multiple pairwise comparisons (
Dufrêne and Legendre 1997;
Roberts 2016). A significant indicator order for one bat species indicates that that bat species consumes larger volumes of the order or consumes it at greater frequencies than does the other species (
Dufrêne and Legendre 1997;
Roberts 2016).
We used a two-factor ANOVA to analyze whether time of year and location and year group had an effect on mean daily temperature. Location and year groups included temperature from Kananaskis in 1985 to 1988, Kananaskis in 2018, and Fort Smith, Northwest Territories, in 2011 and 2012. All examined feces were collected from 19 May to 6 Sept. Data from Kananaskis were split into two groups, because there was a long time between the two and mean temperature may have changed due to climate change or other environmental factors. We used a percentage overlap calculation to determine dietary overlap for the entire sampling season between
M. lucifugus and
M. evotis (
Krebs 1999); it was calculated by summing the lowest percent volume of each prey order between the two species. To examine the factors influencing spider consumption in Kananaskis, we created 14 models based on a priori biological hypotheses, plus a global model, and used Akaike’s information criterion (AIC) for model selection (Supplementary Table S1).
1 All models within 6 AIC units from the best model were considered. There was one best model and a two-factor ANOVA was performed on that model. Finally, a generalized linear model with a binomial distribution was performed on the spider occurrence in fecal samples to determine whether there was a difference in the consumption of spiders between species and locations (Kananaskis and Fort Smith). Model performance was checked using the “DHARMa” package (
Hartig 2020), and the data were checked for over-dispersion, under-dispersion, and zero inflation.
Results
There was a significant effect of date (two-factor ANOVA,
F[1,327] = 10.1,
p = 0.002) and location and year group (two-factor ANOVA,
F[2,327] = 47.8,
p < 0.001) on the average mean daily temperature (
Fig. 1). However, there was no significant interaction (two-factor ANOVA,
F[2,327] = 1.22,
p = 0.296) between date and location and year group (
Fig. 1). Fort Smith had the highest average mean daily temperature, whereas Kananaskis from 1985 to 1988 had the lowest (
Fig. 1). All locations exhibited parabolic patterns in temperature, with the highest temperatures in mid- to late July (
Fig. 1).
Of the bats who were hand-fed spiders, 6 out of 10 ate them. Of these bats, 3/4 M. lucifugus and 3/6 M. evotis consumed the spiders. Three bats did not consume spiders or moths when hand-fed.
Fecal samples from 59
M. evotis (21 female, 37 male, 1 unknown) and 62
M. lucifugus (16 female, 46 male) were analyzed, although samples that did not contain identifiable insect prey (4
M. lucifugus, 5
M. evotis) were removed. Seven orders of arthropods were consumed by the two species (
Table 1). There was a significant difference in the diet composition of
M. lucifugus and
M. evotis (PERMANOVA,
F[1,105] = 4.91,
p = 0.002). Presence of the orders Lepidoptera (DLISA, indicator value = 0.488,
p = 0.008) and Araneae (DLISA, indicator value = 0.222,
p = 0.001) were significant indicators for
M. evotis, whereas Diptera (DLISA, indicator value = 0.359,
p = 0.032) was the significant indicator order for
M. lucifugus. Dietary niche overlap between
M. lucifugus and each location’s long-eared bat species was comparable between Fort Smith (80.4%) and Kananaskis (74.9%).
Both
M. evotis in Kananaskis and
M. septentrionalis in Fort Smith (
Kaupas and Barclay 2018) consumed spiders all summer long (
Figs. 2a and
2b). In contrast,
M. lucifugus did not consume spiders at all in Kananaskis, although it consumed spiders early and late in the season in Fort Smith (
Figs. 2c and
2d). Occurrence of spiders in Kananaskis fecal samples was lower than in those from Fort Smith (GLM,
z = −4.67,
p < 0.001), and lower in
M. lucifugus fecal samples than those of long-eared bats species (
M. evotis in Kananaskis and
M. septentrionalis in Fort Smith; GLM,
z = −3.28,
p = 0.001;
Figs. 2a–2d). In Kananaskis, 22.2% (12/54) of
M. evotis and 0% (0/58) of
M. lucifugus fecal samples contained spiders while in Fort Smith, 70% (35/50) of
M. septentrionalis and 37.5% (21/56) of
M. lucifugus samples contained spiders. However, there was no interaction between occurrence of spiders in fecal samples of bat species and location (GLM,
z = −0.019,
p = 0.985;
Figs. 2a–2d). GLM validation did not show any significant deviations and data were not over-dispersed, under-dispersed, or zero-inflated (Supplementary Fig. S2).
1The model that best described variation in the proportion of spiders consumed by Kananaskis bats included species and distance from capture location to water as factors, with the global model including species, date, sex, mean temperature, and proximity to water (Supplementary Table S1).
1 There were main effects of species (two-factor ANOVA,
F[1,108] = 9.89,
p = 0.002) and distance to water (two-factor ANOVA,
F[1,108] = 7.73,
p = 0.006), as well as a significant interaction (two-factor ANOVA,
F[1,108] = 9.09,
p = 0.003), on the proportion of spiders consumed. In Kananaskis, a greater proportion of
M. evotis consumed spiders than
M. lucifugus did, and more
M. evotis consumed spiders when they were caught less than 50 m from a source of water (0.147 ± 0.058 (mean ± SE) proportion of fecal samples,
n = 19) compared with when they were more than 50 m away (0.020 ± 0.011 proportion of fecal samples,
n = 35).
Discussion
As we predicted, the diets of
M. lucifugus and
M. evotis in our study area differed, although not entirely as we expected. Lepidoptera and Araneae were indicator prey for
M. evotis, whereas Diptera was the indicator for
M. lucifugus. This difference in diet is not surprising given the different foraging behaviours employed by each species.
Myotis lucifugus commonly forages for aerial insects over bodies of water and catches insects close to, and occasionally from, the water’s surface (apparently using its mouth) (
Fenton and Bell 1979;
von Frenckell and Barclay 1987;
Barclay 1991), whereas
M. evotis forages primarily within the forest along flyways, gleaning within the clutter (
Barclay 1991;
Holloway and Barclay 2000;
Faure and Barclay 1992,
1994). The availability of dipterans will be greater for
M. lucifugus than for
M. evotis since many dipterans emerge from an aquatic larval state and mate in the open air over or on the surface of the water (
Barclay 1991;
Drake 2001;
Laeser et al. 2005).
Myotis evotis uses passive prey detection by listening for prey-produced sounds while it is foraging, aided by its large ears (
Faure and Barclay 1992). Moths make fluttering sounds as they warm up their wing muscles to take off for flight, making them easier for
M. evotis to detect passively (
Faure and Barclay 1992). Some moths are able to hear the echolocation calls of bats, so the cessation of echolocation by
M. evotis as they approach fluttering moths (
Faure et al. 1990) allows individuals to be effective foragers on moths (
Faure and Barclay 1992).
Myotis lucifugus may not be as effective since it continues to echolocate while attacking fluttering moths (
Ratcliffe and Dawson 2003), which may be alerted by its echolocation calls. Lepidopterans are an indicator species for
M. evotis likely because of this difference in foraging behaviour.
In Kananaskis, there was an overall dietary niche overlap of 74.9% between
M. lucifugus and
M. evotis, similar to that in the Northwest Territories between the diets of
M. lucifugus and
M. septentrionalis (80.4%;
Kaupas and Barclay 2018). The similar niche overlap between the two locations suggests that the predatory role of each long-eared bat species (
M. septentrionalis in Northwest Territories and
M. evotis in Kananaskis) is similar and that equivalent dietary competition occurs between the species in each location.
In Kananaskis, as we predicted, M. evotis consumed spiders all season long and date was not included in the best fitting models examining spider consumption. However, we also predicted that M. lucifugus would consume spiders during early and late summer, but we found that it did not consume spiders at any time. We predicted that spiders would be consumed by M. lucifugus when aerial insect abundance was reduced by cold temperatures, but that M. lucifugus would preferentially feed on aerial insects in warmer temperatures. This would reduce competition between the two species of bats and result in niche separation.
Despite similar niche overlap and a similar amount of interspecies competition to that of the Northwest Territories bats, M. lucifugus did not eat spiders in Kananaskis. Indeed, there was lower spider consumption in Kananaskis than in the Northwest Territories, regardless of species. This is surprising because mean daily temperatures in Kananaskis were lower than those in Northwest Territories, suggesting lower aerial insect abundance.
Myotis evotis and
M. lucifugus in Kananaskis forage in different habitats and therefore encounter different prey, which may explain why there were differences between them in spider consumption.
Myotis evotis forages in the trees, whereas
M. lucifugus forages over open water (
Barclay 1991). Abundance of orb weaver spiders increases with proximity to water because the transition from water-bound juvenile stages of insects to aerial adults means there is an increase in aerial insects near bodies of water (
Laeser et al. 2005).
Myotis evotis often forages within the forest in riparian zones, which may increase the rate these bats encounter spiders, whereas
M. lucifugus foraging over water may not encounter spiders.
Among
M. evotis, bats caught at sites within 50 m of a body of water consumed spiders in greater proportions than those caught more than 50 m away from water. We suggest this is due to an increased chance of encountering spiders when they forage in a location with greater spider abundance, such as riparian forests. However, this does not explain the variation in spider consumption found in bats in Northwest Territories (
Kaupas and Barclay 2018).
Avoiding open habitats, such as over water, is common for foraging
M. lucifugus in Yukon, Canada (
Talerico 2008), perhaps to avoid predation by visual predators (birds), under relatively bright northern conditions at night. It is also possible that
M. lucifugus forages in the darker interior of forests because their prey are there as a way of escaping predation by birds (
Talerico 2008;
Lima and O’Keefe 2013). When
M. lucifugus forages in the interior of forests, as
M. septentrionalis does, it may also encounter spiders, which are opportunistically consumed during colder temperatures and lower aerial insect abundance, such as at the beginning and end of summer (
Talerico 2008;
Lima and O’Keefe 2013). This difference in foraging locations may explain why there is a difference in spider consumption by
M. lucifugus during colder temperatures between our study area and northern sites.
How bats detect and capture spiders needs further study. Future analysis of the types of spiders consumed by bats could help to further explain the differences among locations and species. A better understanding of which spiders are consumed, which habitats they occur in, and whether they balloon seasonally could allow for a better understanding of how bats detect and capture spiders. Additionally, investigating the diet of M. lucifugus and M. evotis in different geographical locations, and in different habitats and temperatures, will help to further understand variation in spider consumption by bats.