Striking similarity of predation pressures across Heliconius species and populations
Our results of predator attacks on Heliconius models correspond to the attack patterns observed in mark-release-recapture experiments, where naïve predators significantly attack novel, exotic patterns relative to native, common warning patterns (Mallet and Barton, 1989; Langham, 2004). As may be expected, attacks of live prey showed much greater differences for exotic and native morphs (36-37% versus 0%, respectively), than we observed with artificial models (5-11.5% versus 4-6.5%, respectively; see Figure 3). This difference between model and live prey experiments likely results from the inability of models to replicate physical and behavioral cues recognized by potential predators (i.e. crawling, wing flapping, flight). Although the attack numbers on models likely do not reflect realized predation rates of live butterflies, they do provide reliable means to compare relative predation pressures due to differences in visual cues.
Importantly, there is a remarkable similarity in attacks ofHeliconius models in studies that span a variety of species and geographic locations over the past decade (Arias et al., 2016; Chouteau et al., 2016; Finkbeiner et al., 2012; Finkbeiner et al., 2014; Finkbeiner et al., 2018; Merrill et al., 2012; Seymoure et al., 2018). Attack percentages in these studies range from 4-15%, which overlaps our observed attack percentages that ranged from 4-12% on H. doris morphs. This consistency in attacks of models may reflect the similarity in avian predation pressures amongHeliconius species and populations. The Rufous Tailed Jacamar (Galbula ruficauda ) has been reported as a common predator ofHeliconius in western South America (Mallet & Barton, 1989) and Central America (Dell’aglio et al., 2016; Langham, 2004). In line with these reports, we observed a Rufous Tailed Jacamar successfully attack a Heliconius (likelyHeliconius sara ) at one of our inland experimental sites in French Guiana. Collectively, this supports that Jacamars may be a common predator driving similar attack rates in the various Heliconiusmodel experiments. Therefore, we can make direct comparisons among these studies and general inferences about the relative effectiveness of specific aposematic color patterns and corresponding selective pressures.