Many herbivores employ reward-based mutualisms with ants to gain protection from natural enemies. We examine the evolutionary dynamics of a tetra-trophic interaction in which gall wasp herbivores induce their host oaks to produce nectar-secreting galls, which attract ants that provide protection from parasitoids. We show that, consistent with other gall defensive traits, nectar secretion has evolved repeatedly across the oak gall wasp tribe and also within a single genus (Disholcaspis) that includes many nectar-inducing species. Once evolved, nectar secretion is never lost in Disholcaspis, consistent with high defensive value of this trait. We also show that evolution of nectar secretion is correlated with a transition from solitary to aggregated oviposition, resulting in clustered nectar-secreting galls, which produce a resource that ants can more easily monopolize. Such clustering is commonly seen in ant guard mutualisms. We suggest that correlated evolution between maternal oviposition and larval nectar induction traits has enhanced the effectiveness of this gall defense strategy.
Cynipid species inducing nectar-secreting galls
Cynipid species inducing galls known to secrete nectar, with information on their galls. Data collected from the literature and personal field observations.
Online Supplementary Table S1.xls
Specimen collection and sequencing details, including GenBank accession numbers.
Specimen collection and sequencing details, including GenBank accession numbers. Species names in bold are nectar secretors; individual codes in italics indicate Disholcaspis species included in the Cynipini-wide analyses. Four species currently classified as belonging to the genus Disholcaspis (D. chrysolepidis, D. corallina, D. plumbella and D. sulcata) were excluded from the Disholcaspis study as sequence data (see Figure 2) and morphology (see Melika and Abrahamson 2002) both indicate they are not true Disholcaspis.
Online Supplementary Table S2.xls
Posterior probabilities of nectar secretion or no nectar secretion at ancestral nodes within the Disholcaspis phylogeny.
Posterior probabilities of nectar secretion or no nectar secretion at ancestral nodes within the Disholcaspis phylogeny (nodes numbered in insert). Posterior probabilities of ancestral character states are presented for the full set of models visited in the rjMCMC run (columns 2 & 3), those under a symmetric reversible model (columns 4 & 5) and under a model where the rate of change of nectar to no nectar was fixed at zero (columns 6 & 7). The final set of columns indicate likelihoods of models run with the respective node fixed to one or other character state, and the result of a Bayes factor test comparing the two models.
Online Supplementary Table S3.xls
Cynipini cytb sequences
Alignment of Cynipini cytochrome b sequences
Cynipini_cytb.nex
Cynipini D2 sequences
Alignment of Cynipini 28S D2 sequences
Cynipini_D2.nex
Cynipini opsin sequences
Alignment of Cynipini opsin sequences
Cynipini_opsin.nex
Disholcaspis cytb sequences
Alignment of Disholcaspis cytochrome b sequences
Disholcaspis_cytb.nex
Disholcaspis D2 sequences
Alignment of Disholcaspis 28S D2 sequences
Disholcaspis_D2.nex
Disholcaspis ITS2 sequences
Alignment of Disholcaspis internal transcribed spacer 2 sequences
Disholcaspis_ITS2.nex
Disholcaspis opsin sequences
Alignment of Disholcaspis opsin sequences
Disholcaspis_opsin.nex