Coevolutionary arms race versus host defense chase in a tropical herbivore-plant system

Abstract

Coevolutionary models suggest that herbivores drive diversification and community composition in plants. For herbivores, many questions remain regarding how plant defenses shape host choice and community structure. We addressed these questions using the tree genus Inga and its lepidopteran herbivores in the Amazon. We constructed phylogenies for both plants and insects and quantified host associations and plant defenses. We found that similarity in herbivore assemblages between Inga species was correlated with similarity in defenses. There was no correlation with phylogeny, a result consistent with our observations that the expression of defenses in Inga is independent of phylogeny. Furthermore, host defensive traits explained 40% of herbivore community similarity. Analyses at finer taxonomic scales showed that different lepidopteran clades select hosts based on different defenses, suggesting taxon-specific histories of herbivore-host plant interactions. Finally, we compared the phylogeny and defenses of Inga to phylogenies for the major lepidopteran clades. We found that closely related herbivores fed on Inga with similar defenses rather than on closely related plants. Together, these results suggest that plant defenses might be more evolutionarily labile than the herbivore traits related to host association. Hence, there is an apparent asymmetry in the evolutionary interactions between Inga and its herbivores. Although plants may evolve under selection by herbivores, we hypothesize that herbivores may not show coevolutionary adaptations, but instead "chase" hosts based on the herbivore's own traits at the time that they encounter a new host, a pattern more consistent with resource tracking than with the arms race model of coevolution.

Keywords: Inga; coevolution; defensive traits; herbivores; plant–herbivore interactions.

Fig. S1.

Leaf production of 25 Inga species in Los Amigos between 2007 and 2008 as the proportion of individuals flushing (A) continuously, (B) moderately continuously, and (C) episodic (highly synchronized).

Fig. S2.

Lepidopteran herbivore families associated with Inga in Los Amigos. Percentages represent the relative abundances. All individual herbivores were barcoded with the mitochondrial locus COI. In addition, nuclear genes (elongation-factor 1α, EF-1α, and wingless Wg) were sequenced for the Gelechioidea, Riodinidae, and Erebidae.

Fig. 1.

Relationship between the similarity of lepidopteran communities (1, Bray–Curtis index) on host plants vs. (A) phylogenetic distance between Inga hosts and (Mantel r = −0.25, P = 0.02), (B) distance in defenses between Inga hosts for all pairwise combinations of plants (partial Mantel r = −0.50, P = 0.001).

Fig. 2.

Distance-based redundancy analysis plot of the most parsimonious model for the lepidopteran community similarity measured by the Bray–Curtis index (R²_adj = 0.40, P = 0.001). Each point represents an Inga species and is color-coded by defense chemistry.

Fig. 3.

Results of best-fit distance-based redundancy analyses (dbRDAs) models for the three most abundant lepidopteran families. Significant values (P < 0.05) are marked with an asterisk.

Fig. S3.

Stacked circular histogram of peak flushing events for Inga species at Los Amigos between 2007 and 2008. Flushing events are grouped by defense chemistry.

Fig. 4.

Bipartite trophic network of Inga hosts and herbivores. (A) Phylogenies of Inga and Lepidoptera plotted in the margins. (B) Phylogenies of Lepidoptera and Inga defensogram plotted in the margins. For each network, lower bars represent host abundance and upper bars represent herbivore abundance.

Fig. S4.

Comparison between the phylogenetic tree (Left) and the defensogram (defense dendrogram; Right) for Inga species.