Discussion
Our findings suggest that climate plays a major role in dictating anole community diversity, and that bottom-up forces predominate, such that both bird and anole richness is strongly correlated with mean annual temperature (a proxy for ecosystem productivity). Further, more lizard-specialized predator occurrence was positively correlated with anole community metrics, even more than climate effects, suggesting that these specialist predators may be exerting a stronger top-down effect. Nevertheless, top-down effects from predators seem to leave some signatures in anole communities. We found that areas with higher predator presence also tended towards higher evenness and greater overall diversity in anoles, whereas in general these relationships were not observed with pigeons, our non-predatory “control” group. Our finding that anole species richness was highest where bird predation pressure was highest could be interpreted either as support for bottom-up forces controlling bird occurrence, or as greater top-down pressure on dominant competitors facilitating higher lizard diversity (although, interestingly, without reducing total lizard abundance). Either way, it demonstrates no evidence that greater predator pressure results in a net loss of species richness, as would be expected in the case of extreme top-down forcing inducing local extirpation of prey species.
Our observation that more lizards and more bird species are present in areas where there is higher resource abundance is consistent with the idea that bottom-up forces drive diversity. Favorable environmental conditions promote abundance at lower trophic levels, which thus allow for greater abundance and diversity of consumer species. Indeed, productivity has been used as an indicator of resource availability (Evans et al. 2005, Novosolov et al. 2016). Higher temperature and precipitation are tied to ecosystem productivity and are shown to drive an increase in plant biomass, and can also sustain a larger arthropod population (Siemann 1998, Haddad et al. 2001, Wenninger and Inouye 2008, Bragazza et al. 2015). Lizards rely on vegetation for microhabitat and arthropods for food, and a greater abundance of these would therefore support a greater abundance of lizards in these areas. Under bottom-up control, this higher resource availability would also allow for greater abundances of bird species (both lizard predators and insectivorous competitors, as well as granivores and frugivores that do not directly interact with lizards).
Our findings show that there are more avian predators and anole prey species in warmer climates. In these cases, predator presence and anole community richness appear to be dictated by resource availability and environmental suitability. This is especially true for cuckoo occurrence, which even when controlling for climate shows a strong positive correlation with anole community abundance. It is more likely that we could detect a direct bottom-up effect for these more specialized predators, since lizard abundance will be more limiting for them than for other predators that may also be consuming mammals, insects, amphibians, and birds in addition to lizards.
While our findings suggest that diversity is driven by bottom-up forces, some of our results can be interpreted as showing signatures of top-down effects that could be operating in conjunction. On both islands, we saw that areas with more intense predation also had lizard communities with higher evenness (Figure 3b). Interestingly, there did not appear to be any relationship between predation pressure and lizard abundance, despite there being a positive correlation with community evenness. This suggests that while predator presence seems to track with prey presence, there may be some feedback where predators alleviate competition between anole species, potentially by reducing the abundance of the numerically dominant competitor, and thereby granting more resources for other anole species. This would then result in both minimal changes in community wide abundance, while also increasing both anole species richness and community evenness. Prey switching has been noted in birds, and this would especially serve to increase evenness such that any common competitor eliminated can be replaced by another single less common one (Murdoch 1969, Fitzpatrick et al. 2009). The presence of specialist predators seems to have particularly accentuated effects. We found that while cuckoo presence had a positive relationship with lizard abundance and richness, it was also a significant predictor of community evenness – more so than temperature or precipitation. It is thus possible that cuckoos specifically exert some top-down control of island lizard communities.
While our results may be indicative of predators playing some role regulating diversity, we also found some significant relationships when considering our non-predatory guild: pigeons. Higher pigeon presence was weakly associated with less even lizard communities on both islands and alternatively lower and higher diversity on Hispaniola and Jamaica respectively. Because there is no direct causal link by which pigeons (which are not-insectivores, and therefore neither consume nor compete with anoles) we suspect that pigeon presence captures some other aspect of environmental variation that is not completely accounted for by our climate variables. Importantly however, the significant relationship between pigeons and anole evenness was negative (beta = -0.04), suggesting that the positive relationship found with predatory species is not a characteristic found between birds and anoles generally. Similarly, the effect of pigeons on anole community diversity was generally modest and differed in sign and magnitude on the two islands – its effect was weak and negative on Hispaniola (beta = -0.03), and positive of moderate size on Jamaica (beta = 0.16; Table 1). This was in contrast to all other analyses of predator effects, which, even when a significant interaction effect between island identity and predation pressure existed, was always of the same sign and order of magnitude in strength on both islands. Together, these patterns suggest that the inferred signature of avian predation on these islands is likely real, rather than purely driven by unmeasured features of the environment that affect anoles and birds in tandem.
In addition to birds, snakes, bats, and introduced mammals such as rats and mongoose also prey on lizards, and were not considered in this study. Other potential predators include large spiders, andSolenopsis ants which feed on anole eggs (Reyes-Olivares et al. 2020, Andrews and Rand 2022). While lists of mammal and snake occurrence on these islands as a whole exist, our study focused on birds because the eBird platform provides timed and tracked presence-absence surveys of bird communities at a local scale. As a result, our measures of bird predation are lower overall than what lizards likely experience from all predator sources. Hispaniola has more species of predator snakes than Jamaica (4 and 2 respectively), but fine-scale data on local coexistence is limited, so this this does not necessarily guarantee that snake occurrence is greater at individual sites. Regardless, our measurements consequently may have underestimated the effect of total predation in shaping lizard communities, though we have no a priori expectation that these other forms of predation would have different consequences than those of birds. Our analytical approach represents an improvement over previous whole-island predator lists in that it estimates probability of predator occurrence at the local scale. However, the ideal metric of predation pressure would incorporate local predator density, and then further integrate information on per capita predator effects on prey. Unfortunately, our preliminary examinations of eBird data with abundance-based models resulted in convergence issues, or extreme predictions, likely due to high variance in observed abundance in some checklists (e.g. many hundreds or thousands of individuals, presumably when distant flocks were observed). As such eBird and similar databases at best represent a imperfect substitute for broad-scale time and area-standardized ecological surveys.
Although previous studies generally report negative correlations between anole abundance and predation pressure, thereby implicating predators presence as important drivers of Anolis community structure (Waide and Reagan 1983, Buckley and Jetz 2007, Calsbeek and Cox 2010, Pringle et al. 2019), our results show no evidence that predation pressure either reduces overall anole abundance or leads to local extirpation and declines in species richness. The reasons for this divergence is likely in part due to the differences in focal scale between past studies and our own analyses. Many past studies focused on small, bounded experimental islands, which often found that predation may quickly lead prey to go locally extinct. On much larger islands, however, there are more opportunities for species to find refuge, which may mitigate a predator’s impact on prey abundance, and therefore extinction. On the other end of the spectrum, studies comparing lizard abundances versus predator richness across multiple islands account for multiple species of both prey and consumer, but they rarely consider which species on an island actually co-exist with one another on a local scale. Buckley and Jetz (2007), for instance, conclude that sharp reductions of predators are a major driver of increased lizard densities. However, in such cross-island comparisons larger islands will nearly always have more predator species, even if such predator species do not occur everywhere on the islands. As such, actual local predator pressure may be unrelated to total number of predators on the island as a whole. If some other aspect of island size correlates with average lizard density, then a negative relationship between island-wide predator number and locally measured lizard density may be uncovered, even if the mechanisms driving predator number and lizard density are not directly related.