Fig 1.
Flowers of Campanula species used for the study and their oligolectic pollinator Chelostoma rapunculi.
(A) Ca. glomerata, (B) Ca. lactiflora (C) Ca. persicifolia alba, (D) Ca. persicifolia, (E) Ca. rapunculoides, (F) Ca. rotundifolia, and (G) Ca. trachelium. (G, H) Females of Ch. rapunculi gathering pollen on Ca. trachelium and Ca. glomerata, respectively. All photos by Paulo Milet-Pinheiro.
Table 1.
Mean absolute and relative amount of volatile compounds in headspace samples collected from inflorescences of six Campanula species (one species with two morphs).
Fig 2.
Corolla color loci of Campanula in the bee color hexagon.
Corolla color loci of six Campanula species (one species with two color morphs) and artificial flowers plotted in a hexagon color space. EU, EB, EG: excitation of the UV, blue, and green receptor, respectively. The pairwise interspecific distances in color loci and the distances to the centre of the single species can be found in Supporting Information (S1 Table).
Fig 3.
Behavioral assays testing attractiveness of inflorescences and artificial flowers to Chelostoma rapunculi bees.
Responses of flower-inexperienced bees of Chelostoma rapunculi (males and females together) to visual cues of either inflorescences of Campanula persicifolia or artificial flowers in behavioral assays performed in an experimental flight cage. Numbers above bars indicate the absolute number of responding bees. Difference in responses for each assay was assessed by an exact binominal test, P < 0.05 (*), P < 0.001 (***), not significant (n.s).
Fig 4.
Nonmetric multidimensional scaling (NMDS) of floral scent composition in Campanula.
Comparison of floral scent bouquets among six Campanula species (one species with two color morphs) based on qualitative Bray—Curtis similarities plotted in a non-metric multidimensional scaling (NMDS). The NMDS plot based on semi-quantitative similarities is not shown since samples clustered similarly (see text).