The data set contains Gas Chromatography and Mass Spectrometry (GCMS) analysis data from headspace collections of volatiles from the herb Collinsia heterophylla (Plantaginaceae) covered in Toppits oven bags. Plants were sampled for 5 hours with Porapak Q filters. 200 ng of anethole was added as an internal standard to each sample. Each GCMS file contains a sample from an individual potted plant or a control with only flower pots and soil. Individual plants analyzed by GCMS were either in the bud stage or in the flowering stage, enabling comparisons to identify flower-specific volatiles. Flowering plants belonged to either of two floral morphs, with either a white upper lip or with a dark band on the upper lip of the flower. The data set also contains observations of pollinator visitations to different flower morphs in a population polymorphic for the two flower traits with white or banded upper lip. This includes cross-visitations of pollinators between flower morphs, based on transfer of fluorescent marker dyes between flowers.

We investigated flower scents in Collinsia heterophylla, a bee-pollinated, annual herb with mixed-mating system combining self- and outcross-pollination. In Collinsia, floral size and development variation is related to mating system, ranging from large-flowered mixed-mating species to small-flowered self-pollinated species. To our knowledge, this is the first description of flower scent from any species in the genus. We performed headspace collections of volatiles in the greenhouse from potted flowering plants, and compared these to controls in the bud stage. Flower-specific volatiles were highly dominated by terpenoid compounds typical of bee-pollinated plants, such as β -myrcene, (Z)- and (E)-ocimene and sesquiterpenes (E)-α-bergamotene and β-sesquiphellandrene. The aliphatic ester methyl hexanoate was also prominent, together with additional esters, whereas methyl cinnamate constituted the only aromatic compound. Additionally, we studied whether flower-emitted volatiles were coupled to presence vs. absence of a coloured band on the upper lip within a population in C. heterophylla. Floral colour morphs showed no qualitative difference in volatiles, but the coloured morph produced significantly higher quantities for seven of the 26 individual flower compounds. We also studied if floral colour morphs affected pollinator visitation. A field experiment performed within a natural population, using behavioural observations and florescent dyes dusted on the flowers, could not detect any differences in pollinator visitation between colour morphs. We conclude that C. heterophylla flowers emit volatile compounds commonly associated with attraction of their most important pollinators.