Premise of the study: Shifts in abiotic factors can affect many plant traits, including floral volatiles. This study examines the response of floral volatiles to water availability, and whether phenotypic plasticity to water differs among populations. Furthermore, it investigates genetic differentiation in floral volatiles, determines the effect of temperature on phenotypic plasticity to water, and assesses temporal variation in floral scent emission between day and evening, since pollinator visitation differs at those times.

Methods: Rocky Mountain columbine plants (Aquilegia coerulea), started from seeds collected in three wild populations in Colorado, Utah, and Arizona, were grown under two water treatments in a greenhouse in Madison, Wisconsin, USA. One population was also grown under the two water treatments, at two temperatures. Air samples were collected from enclosed flowers using dynamic headspace methods and floral volatiles were identified and quantified by gas chromatography (GC) with mass spectrometry (MS) detection.

Key Results: Emission of three floral volatiles increased in the wetter environment, indicating phenotypic plasticity. The response of six floral volatiles to water differed among populations, suggesting genetic differentiation in phenotypic plasticity. Five floral volatiles varied among populations, and emission of most floral volatiles was greater during the day.

Conclusions: Phenotypic plasticity to water permits a quick response of floral volatiles in changing environments. The genetic differentiation in phenotypic plasticity suggests that phenotypic plasticity can evolve but complicates predictions of the effects of environmental changes on a plant and its pollinators.

Plants from the three populations were grown from seeds in one room, kept at 20° C during the day and 7.7° C at night, at the Walnut Street greenhouses in Madison, Wisconsin (the cool room). Plants for the CO population were also grown in a warm room kept at 25.5° C during the day, and 13.3° C at night, which represented a difference from the cool room of 5.5° C during the day, and 5.6° C at night. The temperature regime was linked to the light regime in the greenhouse rooms. Within each room, plants within populations were randomly assigned to an every fifth (wet) or every eighth day (dry) water treatment. The same volume of water was added to each plant on a given day, and only the frequency of watering changed between the two water treatments. These two water treatments represented unstressed and stressed conditions. Plants were grown for about four months in their respective treatments, and when most plants had ten leaves, they were vernalized for 2 months at 4º C to trigger flowering before being returned to their respective greenhouse room. While flowering time varied with temperature and water treatment (Brunet and Larson- Rabin, 2012), most plants started flowering 2.5-3 months after being returned to their respective treatments.

To measure the volatiles emitted from columbine flowers, we collected air samples from enclosed flowers using dynamic headspace methods and identified and quantified compounds by gas chromatography (GC) with mass spectrometry (MS) detection. Volatiles were collected from a single receptive female flower per plant, and from up to seven plants per day for most days between March 22 and April 10. We used a chiral GC column in order to separate and measure optical enantiomers of some of the floral volatiles.

We examined the impact of population and water at low temperature; the impact of water at two temperatures for the Colorado population, and examined the temporal variation in floral volatile emission between day and evening (including dusk) because hawkmoths tend to visit these plants at dusk.

Variables are explained in the Read me file