Introduction :
A key question in the study of animal cognition is whether sexual selection has contributed to its evolution (Darwin 1871; Boogert, Fawcett, et al. 2011). A related but subtly different question is whether signals used in mate choice reflect cognitive abilities that are under natural selection, and would therefore offer direct or indirect benefits in a potential mate. The extension of this question is whether the utility of certain assessment signals in selecting a mate with beneficial cognitive abilities has led to their continued use and evolution.
The hypothesis that signals used in mate choice (“assessment signals”) serve as indicators of male quality has been extensively supported in a wide variety of taxa (reviewed in Wilgers and Hebets 2015), including fish (Houde and Torio 1992; Karino et al. 2005), birds (Hill 1991; Doucet and Montgomerie 2003), insects (David et al. 2000; Holzer et al. 2003), and even humans (Hume and Montgomerie 2001; Little et al. 2008). The hypothesis that assessment signals can serve as an indicator of general brain function has also been supported (Boogert, Fawcett, et al. 2011), in particular by studying the effects of developmental stress on bird song (Peters et al. 2014). Developmental stress lowers cognitive performance in a range of taxa (Levitsky and Strupp 1995; Erhard et al. 2004; Kitaysky et al. 2006; Santos de Souza et al. 2008; Peters et al. 2014; Kriengwatana et al. 2015) and, specifically within birds, on a range of tasks including novel foraging tasks (Kitaysky et al. 2006), spatial learning tasks (Pravosudov et al. 2005; Kriengwatana et al. 2015), associative learning tasks (Fisher et al. 2006; Farrell et al. 2016), and auditory learning tasks (Farrell et al. 2016). Developmental stress also negatively impacts preferred features of song, such as complexity (Spencer et al. 2003; Buchanan et al. 2004), bout duration (Buchanan et al. 2003), and accuracy in relation to tutor song (Nowicki et al. 2002), and is less preferred by females when presented in comparison with song from developmentally robust males (Spencer et al. 2005; Searcy et al. 2010).
These findings suggest that song can serve as an indicator of gross brain quality and function in the case of developmental inequity between potential mates. However, in the absence of conditions like developmental stress or genetic abnormalities that cause widespread disruption in the body and brain, there is increasing evidence that avian intelligence is more modular than general in nature (Searcy and Nowicki 2019). For instance, there are seldom correlations between performance on different cognitive tasks within individuals (Nettle et al. 2015; van Horik and Madden 2016; Anderson et al. 2017; Medina-García et al. 2017; DuBois et al. 2018; MacKinlay and Shaw 2018), with some exceptions (Shaw et al. 2015; Ashton et al. 2018). Notably, exceptions tend to occur in wild populations where unequal developmental conditions are more likely. Because of the apparent modularity of avian intelligence, it is unclear whether song is a reliable indicator of the specific cognitive skills that would offer females fitness benefits if found in a potential mate.
One such cognitive skill is foraging ability. Foraging efficiency has been shown to increase fitness in a range of avian species (Orians 1969; Weathers and Sullivan 1989; Lemon and Barth 1992; Lescroël et al. 2010; Cole et al. 2012), and is a particularly important skill in a mate in monogamous species that engage in biparental food provisioning. Indicative of its importance, foraging ability has also been shown to directly affect mate preferences in crossbills (Loxia curvirostra ), budgerigars (Melopsittacus undulatus ), and zebra finches (Taeniopygia guttata ) when females directly observe males solving a foraging task (Chantal et al. 2016; Chen et al. 2019) or more efficiently extracting food from a source (Snowberg and Benkman 2009; Chantal et al. 2016). Foraging ability is also correlated with problem-solving on artificial tasks in the wild (Cole et al. 2012; Cauchard et al. 2013), which has in turn been correlated with reproductive success (Keagy et al. 2009; Cole et al. 2012; Cauchard et al. 2013; Ashton et al. 2018).
The ability to solve a novel foraging problem is thus related to reproductive success and has been shown in some species to be preferred by females when observed directly. However, it remains unclear whether the same signals that can indicate gross brain function, as determined by developmental conditions, can reflect this specific skill. Do songs contain information about male foraging ability, given the apparent modularity of avian cognition, and could this help explain the utility of song as an assessment signal? Or, conversely, does song only contain information about gross developmental conditions, and not more specific skills that would offer females direct fitness benefits?
To answer these questions, we tested whether female zebra finches would discriminate between males with superior versus inferior performance on a novel foraging task based solely on their song. Using an operant conditioning assay, we compared female preference for the songs of males that quickly solved the novel foraging task to the songs of males that were incapable of solving the task. The females had no exposure to the males whose songs they were tested with, thus isolating their knowledge of the males to information contained in their songs. We predicted that if song is a reliable signal of novel foraging skill, females would be able to identify and prefer songs from males who were capable of quickly solving a novel foraging task over songs from males who were unable to solve the task.