Fig. 2

Genetic assessment of parasite species boundaries and parasite host preferences. Coloration depicts species that were also studied in behavioral and chemical analysis (red = specialists; green = generalists; orange = E. burchellii foreli; blue = E. hamatum; see also other figures). Grey boxes depict species that were not considered further. a Neighbor-joining trees based on Tamura-Nei distances (scale bars) of the mitochondrial gene fragment COI detected four genetic clusters for Ecitophya and Ecitomorpha (specialists) and five genetic clusters for Tetradonia (generalists). We excluded the genetic outgroups from the NJ trees for better visibility (indicated by dashed lines; outgroups: Ecitoglossa sp. for specialists and Ecitomorpha cf. nevermanni for generalists). Numbers of analyzed specimens is given in boxes. Bootstrap support values are shown at major branches (1000 replicates). b Host specificity of Tetradonia beetles (generalists) and Wasmannian mimics (specialists). Differential shading corresponds to the proportion of colonies of a given host species in which a given parasite species was collected. White boxes depict missing associations between parasites and army ants. Black boxes show myrmecophiles that were collected in all colonies of a given host species. Photographs depict frontal head views of Eciton soldier workers and dorsal views of beetles for the different species. Data on Tetradonia species boundaries and host specificity have been published previously [20]. Specimen images are not to scale in both graphs