Introduction
Sexual partitioning of food resources is known to occur in many animal species, but the extent and ecological significance of this phenomenon are still poorly understood (Ruckstuhl & Neuhaus, 2006). In birds, differences in diet indicative of resource differentiation have mostly been studied in birds with considerable sexual dimorphism in body size (Bravo, Ponce, Bautista, & Alonso, 2016; Catry, Alves, Gill, Gunnarsson, & Granadeiro, 2012; Donals et al., 2007; Gonzalez-Solis, Croxall, & Wood, 2000; Thalinger, Oehm, Zeisler, Vorhauser, & Traugott, 2018) or in bill size or shape (Smith, 1990; Summers, Smith, Nicoll, & Atkinson, 1990; Temeles, Mazzotta, & Williamson, 2017; Temeles & Roberts, 1993). As a consequence, intraspecific dietary differentiation in birds has been largely attributed to morphological differences, with more sexually dimorphic species expected to show higher resource differentiation (Alarcón et al., 2017; Fonteneau, Paillisson, & Marion, 2009; Lewis et al., 2005; Phillips, McGill, Dawson, & Bearhop, 2011; Selander, 1966). However, it is possible that sexual food resource differentiation also occurs in monomorphic or only slightly dimorphic birds, but this idea remains little explored (but see Botha, Rishworth, Thiebault, Green, & Pistorius, 2017; Cleasby et al., 2015; Elliott, Gaston, & Crump, 2010; Hedd, Montevecchi, Phillips, & Fifield, 2014).
One of the obstacles to understand eventual sexual partitioning of food resources is related to limitations of widely used diet analysis methods, which often are unable to provide enough taxonomic resolution to detect subtle differences in prey consumption (e.g., Mata et al., 2016). This is the case, for instance, of methods widely used in avian ecology, including for instance the morphological identification of the remains of ingested food items (Bravo et al., 2016; Fonteneau et al., 2009; Hunter, 1983; Hunter & Brooke, 1992), direct observation (Catry et al., 2012), fatty acids and alcohols analysis (Owen et al., 2013), or stable isotope analysis (Blanco-Fontao, Sandercock, Obeso, McNew, & Quevedo, 2013; Cleasby et al., 2015; Elliott et al., 2010; Hsu, Shaner, Chang, Ke, & Kao, 2014; Ludynia et al., 2013; Paiva et al., 2018; Phillips et al., 2011). The advent of high-throughput DNA sequencing is making it possible to overcome the limitations of these methods, providing the ability to identify virtually all prey species consumed with unprecedent taxonomic resolution (Hope et al., 2014; Nielsen, Clare, Hayden, Brett, & Kratina, 2017; Razgour et al., 2011; Soininen et al., 2009). As a consequence, this approach has been increasingly used to describe the diets of a wide range of animals (Brown, Jarman, & Symondson, 2012; Kaunisto, Roslin, Sääksjärvi, & Vesterinen, 2017; Macías-Hernández et al., 2018; Mata et al., 2016; Soininen et al., 2009), including birds (Coghlan et al., 2013; Deagle, Chiaradia, McInnes, & Jarman, 2010; Jedlicka, Vo, & Almeida, 2017; Liu et al., 2018; Sullins et al., 2018; Trevelline et al., 2018). The high taxonomic resolution provided by high-throughput sequencing has already been used to describe sexual dietary differences that otherwise would be almost impossible to detect (Mata et al., 2016). However, previous studies have focused on specialists with a relatively narrow feeding niche, while this methodology remains underexplored in testing sexual dietary in more generalist species such as many omnivorous passerines. Dietary generalists are more challenging to study using metabarcoding because they require a combination of markers to fully encompass the full spectrum of food resources used (da Silva et al., 2019a).
Here we aim to show the power of multi-marker metabarcoding to investigate differences in diet between sexes, by focusing on a generalist passerine judged to have minimal sexual dimorphism, the black wheatear (Oenanthe leucura ). To address this general goal, the study first documents differences in morphology (bill and body features) between sexes, and then uses a previously developed approach for integrating metabarcoding dietary data across multiple markers (da Silva et al., 2019a) to describe the diets of both sexes. Using this data we then tested the hypothesis that diet varies between sexes in terms of (i) diet diversity and (ii) frequency of occurrence of the main food items, and that (iii) sexual dietary differentiation can only be detected at the high taxonomic resolution provided by metabarcoding. Results were used to discuss the potential of multi-marker metabarcoding to provide a detailed understanding of intraspecific variation in bird diets.