Introduction
Biotic interactions such as mutualism, commensalism, competition and predation, affect diversity and distribution of species at different scales in different ways (Schemske et al. 2009; Wisz et al. 2013; Araújo and Rozenfeld 2014; Lany et al. 2018; Nelsen et al. 2018). While the role of these interactions in shaping large-scale biodiversity patterns is often acknowledged, it is rarely tested (Schemske et al. 2009; McCain and Grytnes 2010; Wisz et al. 2013). This paucity of studies may be attributed to the fact that biotic interactions are often difficult to quantify (McCain and Grytnes 2010) and that cause and effect can be difficult to distinguish in the relationship between biotic interactions and species diversity (Fischer 1960; Schemske et al. 2009). However, there is a growing need to study the role of these interactions in shaping diversity patterns in the light of predicted shifts in species ranges due to climate change (Blois et al. 2013; Wisz et al. 2013; Valiente-Banuet et al. 2015; Anderson 2017; Gavish et al. 2017), especially as the strength and effects of these interactions are themselves susceptible to climate change (Tylianakis et al. 2008; Faldyn et al. 2018).
One of the key biotic interactions that can influence diversity patterns in various ways is competition. Competition may reduce diversity by competitive exclusion (Goldberg and Barton 1992; Valone and Brown 1995) or enhance diversity through increasing specialization of species (Abbott et al. 1977; Futuyma and Moreno 1988; Emerson and Kolm 2005). Closely related taxa often compete for similar resources and these competitive interactions can influence range limits of these taxa (Terborgh and Weske 1975; Jankowski et al. 2010; Pasch et al. 2013). However, interactions between distantly related taxa also have major impacts on diversity patterns. Although a few studies have demonstrated competition between distantly related taxa (Brown and Davidson 1977; Eriksson 1979; Palmeirim et al. 1989; Hochberg and Lawton 1990; Jennings et al. 2010) and the effect of competition between distantly related taxa on diversity patterns has been inferred from the fossil record (Jablonski 2008), this subject has not received much attention in macroecology. In fact, the presence of closely related species in a community has been used as evidence against the role of competition, emphasizing instead abiotic filtering (Webb et al. 2002; Gómez Juan Pablo et al. 2010; Tucker et al. 2017). However, recent advances in ecological coexistence theory imply that strong differences in competitive ability between distantly related competitors could cause competitive exclusion even if niches are substantially different (Chesson 2000; Mayfield and Levine 2010; Gerhold Pille et al. 2015). Indeed, many empirical studies support this model for ecological coexistence (Venail et al. 2014; Germain Rachel M. et al. 2016).
Here, we present observational and experimental evidence that suggests a role for ecological competition between two distantly related clades in shaping their complementary diversity patterns: songbirds (Phylum Chordata, Class Aves, Subfamily Oscines) and ants (Phylum Arthropoda, Class Insecta, Family Formicidae). While several studies have presented evidence for competition or amensalism between ants and birds (Supplementary table S1), they have not assessed the effect of these interactions on patterns of species diversity. Ants are important predators of other arthropods, especially in tropical and sub-tropical lowland forests (Floren et al. 2002; Sam et al. 2014). They have been experimentally shown to reduce numbers of other arthropods (Piñol et al. 2010; Karban et al. 2017). However, ants are absent or in very low in abundance in tropical and sub-tropical montane cloud forests across the world, for reasons that remain unclear (Janzen 1973; Samson et al. 1997; Longino et al. 2014). The elevation at which cloud forests are found varies with latitude and other geographic factors, but the lower elevational limit usually lies between 1500m-2500 m and upper limit ranges from 2400m-3300m (Stadtmüller 1987). One possible explanation is that ground-nesting ants cannot persist in places that are wet throughout the year whereas arboreal nesting ants cannot survive freezing temperatures (Wheeler 1917; Janzen 1973; Samson et al. 1997). On the other hand, although capable of living at low and higher elevations, carabid beetles (Wilson 1987; Maveety et al. 2013), songbirds (Price et al. 2014) and small mammals (Heaney 2001) are often very diverse and abundant in these cloud forests. One mechanism for this diversity and abundance maxima in cloud forests could be the lack of competition with ants for arthropod prey (Heaney 2001; Price et al. 2014). We investigated this hypothesis in the eastern Himalaya.
In the eastern Himalaya, breeding songbird diversity peaks in the cloud forests at elevations between 1200m and 2000m. Various historical and dispersal hypotheses, including greater area, dispersal from both above and below, and greater time for diversification at mid-elevations (associated with climatic niche conservation) have little support as an explanation for the mid-elevation peak in bird diversity (Price et al. 2014). The peak consists largely of small insectivorous bird species and is associated with greater resource (i.e. arthropod) abundance, potentially supporting more individuals and hence, more species (Price et al. 2014; Schumm et al. 2019) (Fig. 1, Fig. S1 and Fig. S2). By contrast, ants are almost absent in the cloud forests at mid-elevations, even though they are highly abundant and diverse at the low elevations (Fig. 1, Fig. S1 and Fig. S2) (Ghosh-Harihar 2013; Price et al. 2014). Among ants, low elevations are dominated by an arboreal insectivorous species, the Asian weaver ant Oecophylla smaragdina which disappears at about 900m elevation (K. Supriya pers. obs.). Weaver ants forage both on the trees and on the ground, move between trees through canopy connections and are highly aggressive (Basu 1997; Peng and Christian 2005; Van Mele 2008). We evaluated the possibility that the dominance of weaver ants at low elevations contributes to the lower diversity of birds at these elevations due to competition for food resources.
To test whether weaver ants and birds might compete for resources at the low elevation, we first assessed if there is dietary overlap between weaver ants and birds. A necessary pre-condition for competition between two taxa is significant overlap in the use of the same limiting resource (Brown and Davidson 1977). Previous research suggests that arthropods are a limiting resource for songbirds in the eastern Himalaya (Price et al. 2014) (also see Fig.1 and Fig. S2). Here, we compared the diet of weaver ants and birds at the low elevation where they co-occur, and at higher elevations where weaver ants are absent, to assess dietary overlap. Next, we compared arthropod abundance and leaf damage due to insect herbivory on trees with and without weaver ants and conducted a weaver ant removal and exclusion experiment using a paired design to assess whether weaver ants significantly reduce arthropod abundance on trees.