Female dietary bias towards large migratory moths in the European free-tailed bat (Tadarida teniotis)

Abstract

In bats, sexual segregation has been described in relation to differential use of roosting and foraging habitats. It is possible that variation may also exist between genders in the use of different prey types. However, until recently this idea was difficult to test owing to poorly resolved taxonomy of dietary studies. Here, we use high-throughput sequencing to describe gender-related variation in diet composition of the European free-tailed bat (Tadarida teniotis), while controlling for effects of age and season. We analysed guano pellets collected from 143 individuals mist-netted from April to October 2012 and 2013, in northeast Portugal. Moths (Lepidoptera; mainly Noctuidae and Geometridae) were by far the most frequently recorded prey, occurring in nearly all samples and accounting for 96 out of 115 prey taxa. There were significant dietary differences between males and females, irrespective of age and season. Compared to males, females tended to consume larger moths and more moths of migratory behaviour (e.g. Autographa gamma). Our study provides the first example of gender-related dietary variation in bats, illustrating the value of novel molecular tools for revealing intraspecific variation in food resource use in bats and other insectivores.

1. Introduction

Sexual segregation in resource use is common in vertebrates [1]. Segregation is often associated with morphological and behavioural differences between sexes, which in turn affect a number of ecological and life-history traits such as home range, habitat selection, diet, foraging behaviour and survival rates [1]. Therefore, research on sexual segregation and its underlying causes is important to understand vertebrate ecology, demography and evolution, with implications in wildlife management and conservation.

In bats, most species do not exhibit obvious sexual dimorphism, but segregation between sexes has been described in relation to roosting and foraging habitat use, particularly during the maternity season [1]. In temperate bats, females tend to use warmer roosts for maximizing fetal growth rate and milk production [2], while males tend to choose colder roosts to make use of torpor and maximize energy saving [3]. In some species, females also tend to forage closer to roosts [4–6], as this seems to be more cost-efficient and can lead to lower infant mortality [7], where males seem to be forced to feed away from breeding areas, thereby reducing potential competition with females [8]. As a result, it is possible that segregation may also occur in the use of different prey types [9]. Testing this hypothesis, however, has been hindered by poor taxonomic resolution of most bat dietary studies, though the recent development of molecular tools for dietary analysis provides the opportunity to examine this issue in great detail [10].

We used high-throughput DNA metabarcoding to examine dietary sexual segregation in the European free-tailed bat (Tadarida teniotis). This is a medium-sized bat without obvious sexual dimorphism, which hunts at high altitude and has a large foraging range [11], and forms mixed colonies composed mainly of females [12]. The species has a highly specialized diet composed predominantly of nocturnal moths, but no dietary variation between sexes has been described [13]. Yet, it might be expected that owing to their high energetic requirements during breeding, females should feed more than males on large prey and on prey with high energetic value such as migratory moths [14]. To test this idea, we provide a detailed description of the diet of European free-tailed bats. Additionally, while controlling for potentially confounding factors related to age and sampling season, we assess differences between males and females in relation to (i) prey species composition, (ii) prey species richness, (iii) prey size, and (iv) prevalence of migratory moth species.

2. Material and methods

In April–October 2012 and 2013, we mist-netted bats at their roost in five bridges located in northeast Portugal (N41°09′–42°00′, W7°15′–6°15′); electronic supplementary material, Methods and figure S1), under an on-going monitoring programme for this species (e.g. [14]). Bats were placed in clean individual cotton bags, whence guano pellets were subsequently collected. We recorded gender, age (juveniles versus adults) and sampling date (1st of April = day 1) of each individual. Pellets were stored in tubes containing silica-gel and refrigerated at 4°C until DNA extraction.

We extracted DNA from one pellet per individual (55 adult females, 47 adult males, 14 juvenile females and 27 juvenile males) using the QIAamp DNA Stool Kit (Qiagen) following standard protocol with adjustments suggested by Zeale et al. [10]. We amplified DNA using arthropod general cytochrome oxidase subunit 1 (COI) primers ZBJ-ArtF1c and ZBJ-ArtR2c [10], modified to contain Illumina adaptors and a small identification barcode. Library preparation followed the manufacturer's protocol for metagenomic sequencing (Illumina). Independently amplified samples were individually tagged and sequenced using a MiSeq desktop sequencer (Illumina).

We used OBITools (https://git.metabarcoding.org/obitools/obitools) for general sequence processing. From each pellet, we removed haplotypes representing less than 1% of the total number of reads and those containing stop codons. We then compared the remaining haplotypes retrieved from the overall sample (n = 315) against known reference sequences within the BOLD database (www.boldsystems.org). Haplotypes that best matched the same species were collapsed into a single taxon unit. But all eight of these units (n = 115) contained at least one sequence whose similarity between known species was higher than 98.5%, and were identified to species level. When the same haplotype matched more than one species, we only considered species known to occur in the Iberian Peninsula, or classified them into a species group. Species identifications were screened by a specialist on Portuguese Lepidoptera (Martin F. V. Corley), and discarded from statistical analyses if the putative identification conflicted with known distribution or flight time. For moth species, we estimated the size (wingspan in millimetres) and migratory behaviour (migratory or not migratory) from available literature (electronic supplementary material, Methods and table S1).

PerMANOVA was used to compare diet composition between gender, age classes and sampling day, using the vegan package (http://cran.r-project.org/package=vegan) for R (www.r-project.org). Prey contribution to differentiation between groups was assessed with a similarity percentage analysis, also using vegan. Generalized linear models (GLM) were used to estimate the effects of sex, age and sampling date on diet species richness (negative binomial errors; log link) and proportion of migratory species (binomial; logit) per individual. Generalized Linear Mixed Models (GLMMs) were used to examine the effect of these variables on the size of moth species found in pellets (Gaussian, identity), using individuals as random factors. Model building and inference were based on the information-theoretic approach [15], using as candidates all model subsets built with three predictors and all their interactions (electronic supplementary material, Methods). GLM analyses were conducted with MuMIn (http://CRAN.R-project.org/package=MuMIn).

3. Results

In the diet of T. teniotis, we identified 115 prey items within five insect orders (table 1). Lepidoptera accounted for 83.5% of prey items and occurred in 99% of pellets. Most Lepidoptera were Noctuidae (47 prey items) and Geometridae (14). Diptera and Neuroptera each occurred in about 10% of pellets, while the occurrence of Coleoptera and Hemiptera was much lower. Each pellet contained on average 4.1 ± 2.2 prey items, of which 56.9±36.7% were migratory moth species.

Table 1.

Prey items recorded in the diet of female (n = 69) and male (n = 74) Tadarida teniotis. N.I., Not identified.

order-family	species (common name)	no. prey items	% samples
			F	M
Coleoptera		2	1.4	1.4
Diptera		6	10.1	12.2
Tipulidae	Tipula oleracea (crane fly)	1	5.8	9.5
	other Tipulidae	3	1.4	2.7
other Diptera		1	1.4	1.4
Diptera N.I.		1	1.4	0.0
Hemiptera		1	2.9	8.1
Lepidoptera		96	98.6	100.0
Crambidae	Nomophila noctuella (rush veneer)	1	7.2	14.9
	other Crambidae	6	5.8	9.5
Gelechiidae	Mirificarma mulinella	1	7.2	14.9
	other Gelechiidae	1	0.0	2.7
Geometridae	Aspitates ochrearia (yellow belle)	1	4.3	10.8
	Rhodometra sacraria (vestal)	1	14.5	36.5
	other Geometridae	12	20.3	13.5
Noctuidae	Agrotis ipsilon (dark sword-grass)	1	21.7	4.1
	Agrotis puta (shuttle-shaped dart)	1	13.0	18.9
	Agrotis segetum/trux (turnip/crescent dart)	1	37.7	36.5
	Autographa gamma (silver y)	1	43.5	23.0
	Hoplodrina ambigua (vine's rustic)	1	18.8	35.1
	Mythimna albipuncta (white-point)	1	5.8	9.5
	Mythimna vitellina (delicate)	1	24.6	28.4
	Noctua pronuba/janthe (large/lesser broad-bordered yellow underwing)	1	14.5	13.5
	Peridroma saucia (pearly underwing)	1	24.6	12.2
	Phlogophora meticulosa (angle shades)	1	18.8	8.1
	other Noctuidae	37	31.9	36.5
Tortricidae	Tortrix viridana (European oak leafroller)	1	5.8	13.5
	other Tortricidae	4	2.9	5.4
other Lepidoptera		14	17.4	21.6
Lepidoptera N.I		7	13.0	23.0
Neuroptera		7	7.2	12.2
Insecta N.I.		2	1.4	1.4
Arthropoda N.I.		1	7.2	5.4

PerMANOVA revealed significant variation in diet composition related to gender, but not in relation to age or sampling day (electronic supplementary material, table S2). Gender-related differences were primarily owing to a higher frequency of Rhodometra sacraria and Hoplodrina ambigua in males, and of Agrotis ipsilon and Autographa gamma in females (figure 1). GLM and GLMM models (electronic supplementary material, tables S3 and S4) provided strong support (w+ > 0.90) for females consuming larger prey and a higher proportion of migratory moth species than males (figure 2). There was moderate support (w+ = 0.82) for juveniles consuming smaller moths than adults, but the confidence interval overlapped zero and the effect size was small (figure 2).

Figure 1.

Prey species with the highest contributions to dietary differences between female (F) and male (M) Tadarida teniotis according to the analysis of similarity percentages. Species whose frequencies of occurrence are significantly different between genders are marked as follows: *p < 0.1; **p < 0.05; ***p < 0.01. (Online version in colour.)

Figure 2.

Boxplots representing the average, standard error, and 95% confidence interval of prey richness, prey wingspan (in millimetre) and proportion of migratory prey in diets of female (F) and male (M), and adult (Ad) and juvenile (Juv) Tadarida teniotis.

4. Discussion

Our results showed for the first time the occurrence of gender-related dietary differences in an insectivorous bat species. Although there was a substantial overlap in diet composition, the average size (wingspan) and the frequency of migratory moth species were much higher in females than in males. Together with previous studies on the use of roost and foraging habitats, these results suggest that gender-related ecological segregation may be more frequent in bats than previously recognized [3,5].

The higher consumption of large and migratory moths by females may be a consequence of their high energy demands during pregnancy and lactation [2]. Feeding on large moths may be particularly rewarding because they provide a large energy intake per individual captured, though this should be weighed against the effort needed to catch each prey and its digestibility, which are unknown in T. teniotis. Likewise, migratory moths may be rewarding because they tend to be large and to build energetic reserves to sustain migration [14]. Finally, migratory moths may provide attractive foraging patches, as heavily eaten species such as Autographa gamma migrate during the night in very large swarms [16]. Tadarida teniotis may be particularly adapted to explore this valuable resource, because it is a fast and high-flying species with low manoeuvrability, which feeds in open areas, mostly above canopy level. Studies on the congeneric Tadarida brasiliensis have shown that large numbers of individuals track migratory moth swarms at 400–500 m above ground level [17].

If large and migratory moths are particularly rewarding prey, it might be expected that males should use them heavily as well, instead of resorting to relatively smaller and sedentary species. As suggested in previous studies on bat sexual segregation in foraging habitats [4,5,8], this may be driven to some extent by intraspecific competition, with males avoiding foraging close to putatively dominant females. Irrespective of current competition, however, dietary segregation may also be a consequence of gender-related differences in morphology [18], echolocation [19,20], or social and physiological needs [21], resulting in distinct foraging habitats or prey types. Although the single radio-tracking study on T. teniotis did not find any evidence of gender-related differences in habitat use [11], this could be a consequence of small sample sizes and short tracking periods.

The gender-related differences documented here may also be a consequence of vertical segregation in space use, with females foraging more frequently at the high altitudes used by large migrating moths [16,17], and males foraging closer to the ground where encounters with sedentary species should be more frequent. Interestingly, the only migratory moth that males fed on more often was Rhodometra sacraria, a small species known to migrate at relatively low altitudes [22]. The pattern of altitudinal segregation suggested here for T. teniotis is inverse from that described in other species [5,23], where females were found in lower elevations, associated with resource-abundant riparian habitats. However, vertical segregation of genders over the same foraging grounds has never been analysed or found before. Clarifying the occurrence of vertical segregation between male and female T. teniotis, and how this affects predation on high-flying migratory moths, should be the subject of further research, using for instance altimeter tags to estimate the altitude of foraging individuals.

Overall, our study points out the importance of understanding gender-related ecological segregation in bats, and the unique opportunities raised by DNA metabarcoding [10]. For instance, our results suggest that female T. teniotis may be more susceptible to large-scale climate changes driving moth migrations [24], whereas males may be more vulnerable to local moth declines arising for example owing to land use changes. Conservation actions targeting this species should thus consider the different gender requirements. These conservation implications would have gone unnoticed using conventional diet analysis, because low taxonomic resolution would have blurred dietary differences between sexes. Future metabarcoding studies are needed to describe gender-related dietary variation in other bats and insectivores, which should contribute to our understanding of species ecology and evolution, and aid in the design of conservation strategies.

Ethics

Sampling was approved by Instituto da Conservação da Natureza e das Florestas.

Data accessibility

The dataset supporting this article has been uploaded in Dryad (http://dx.doi.org/10.5061/dryad.m8t72).

Authors' contributions

V.A.M, F.A., H.R. and P.B. designed the study. V.A.M and F.A. collected the data. V.A.M., M.F.V.C., G.F.M., H.R. and P.B. analysed the data. V.A.M. led the writing with substantial contributions from all authors. All authors are accountable for the content of this publication and approve its final version.

Competing interests

The authors have no competing interests.

Funding

The study was funded by Fundação para Ciência e Tecnologia (FCT) (Project LTER/BIA-BEC/0004/2009) and EDP Biodiversity Chair. FCT funded Vanessa Mata (PD/BD/113462/2015), Francisco Amorim (PD/BD/52606/2014) and Hugo Rebelo (IF/00497/2013).