Discussion
In this study, we set out to ask if weaver ants were likely to compete with birds for arthropods at low elevations in eastern Himalaya. Our results show significant diet overlap between weaver ants and birds. Moreover, we found lower arthropod abundance on trees with weaver ants and confirmed experimentally that weaver ant removal led to increased abundance of arthropods. This is important because a greater abundance of arthropods at mid-elevations correlates with the presence of many small insectivorous bird species at these elevations (Price et al. 2014). Further, the two most common arthropod orders in bird diets at low elevations were the ones to increase most in abundance after weaver ant removal. Since weaver ants are found only at low elevations, they could contribute to the presence of fewer arthropods at these elevations rather than higher up. Together, these results suggest that weaver ants reduce food availability for birds at low elevations in the eastern Himalaya.
Our molecular diet analyses showed overlap in the diet of weaver ants and low elevation birds at all taxonomic levels and overlap in the diet of weaver ants and mid-elevation birds at higher taxonomic levels. While many previous studies have presented evidence for competition between birds and ants (Supplementary table S1), this appears to be the first study that has quantified overlap in diet between an ant species and insectivorous songbirds. Our work also demonstrates the utility of molecular diet analyses to examine dietary niche overlap between distantly related organisms. Studies of dietary niche partitioning among species are increasingly using molecular tools to get fine-scale taxonomic information on diet composition (Razgour et al. 2011; Brown et al. 2014; Kartzinel et al. 2015; Arrizabalaga‐Escudero et al. 2018). Due to PCR and sequencing biases, these methods may not give accurate information on the quantity of dietary items (Pompanon et al. 2012; Shokralla et al. 2012). Still, we think that information on the identity and frequency of dietary items obtained using these methods can be very useful to understand the diversity patterns of dietary guilds along environmental gradients.
Reduced herbivory and reduced abundance of large arthropods on trees with weaver ants have been demonstrated previously. Asian weaver ants have been long used as a biological control agent and are still used to control pest populations in orchards (Way and Khoo 1992; Peng and Christian 2005; Van Mele 2008; Thurman et al. 2019). A recent review found that weaver ants significantly reduce pest populations on tropical tree crops (Thurman et al. 2019). Offenberg et al. (2004b) showed that a leaf beetle species avoided eating leaves with weaver ant pheromones on them. In mangrove forests, the presence of weaver ants is negatively correlated with leaf damage due to herbivores (Offenberg et al. 2004a), as we also found in this study. More broadly, many plant species form strong mutualistic associations with ants to reduce the risk of herbivory by offering them rewards such as food bodies, extra-floral nectaries and domatia (i.e. nesting sites) (Janzen 1966; Fiala et al. 1989; Webber et al. 2007; Chomicki et al. 2016). Even facultative or opportunistic ant-plant interactions are known to reduce herbivory and deter other arthropods from plants occupied or visited frequently by ants (Bentley 1977; Fiala et al. 1994; Chamberlain and Holland 2009; Rosumek et al. 2009; Trager et al. 2010).
We suggest that the relatively high arthropod abundance at mid-elevations in the eastern Himalaya is partly a consequence of reduced ant predation, but this does not exclude contributions from other factors, including differences in primary productivity (Acharya et al. 2011), plant diversity and density (Acharya et al. 2011) and higher seasonality (Supriya et al. 2019). Overall, our results lend support the idea that competition from ants could contribute to mid-elevational peak in songbird diversity in the eastern Himalaya. A similar link between competition with ants and diversity patterns was recorded by Brown & Davidson (1977) who found complementary diversity patterns in response to annual precipitation in seed-eating ants and rodents along a north-south gradient in the US. Likewise Heaney (2001) suggested competition with ants may be responsible for the peak in small mammal diversity in cloud forests in the Philippines. Other studies have shown patterns of complementary diversity patterns between ants and other arthropod groups, such as staphylinid beetles, carabid beetles and spiders (Hölldobler and Wilson 1990; Halaj et al. 1997; Noreika and Kotze 2012). More generally, the near-absence of ants in cloud forests (Longino et al. 2014) could be an important explanatory factor for the high diversity of many other taxa there.
A recent meta-analysis of ant species diversity patterns along elevational gradients found some support for a model whereby temperature and precipitation interact to affect ant diversity, i.e. there is a significant relationship between temperature and ant diversity on 83% of wet mountains compared to only 25% of arid mountains (Szewczyk and McCain 2016). Still, the question of why ants are so rare in montane cloud forests remains largely unanswered. Previous studies have suggested that the combination of year-round cool temperatures and humidity make cloud forests unsuitable for ants (Wheeler 1917; Janzen 1973; Samson et al. 1997). However, competition from endotherms such as birds and mammals might also contribute to limiting ant distributions, in much the same way that we postulate ants affect birds in the warm and wet lower elevations. Experimental tests to compare the importance of abiotic versus biotic effects in shaping the pattern of ant abundance along elevational gradients are much needed.
Overall, our observational and experimental data suggest that birds and ants compete for arthropod prey at low elevations in the eastern Himalaya. As diversity patterns of taxa shift due to climate change, it is important to monitor these patterns and compare patterns of such distantly related but potentially competing taxa. We advocate for more studies on ecological interactions between distantly related taxa in shaping diversity patterns, because these interactions could dampen (Suttle et al. 2007) or enhance the effect of climate change on species abundance and range distributions, depending on the actors involved (Davis et al. 1998).