Two additional consensus shapes were generated from the subsets of Nicotiana species pollinated primarily by hawkmoths [CHawkMoth (CHM)] or hummingbirds [CHummingBird (CHB)], respectively
(see Fig. 1).
In initial behavioural trials, only the shapes CAll and CHM were used.
The inner diameter (of area within the corolla lobes) of all paper shapes was held constant (2ฦ0 cm, extrapolated from the average measured N. alata overall corolla diameter of 5ฦ5 cm).
However, because total surface area was not equal for these different shapes (5ฦ95 ± 0ฦ19 cm2 vs. 4ฦ28 ± 0ฦ15 cm2, calculated by using paper corolla mass as a proxy for surface area; t18 = 21ฦ82, P < 0ฦ0001) and is known to affect hawkmoth foraging performance (Goyret & Raguso 2006), subsequent trials used artificial flowers of all three shapes (CAll, CHM and CHB) with surface area (determined bymass) held constant.
All paper corollas were slightly creased radially (from the apex of each petal inwards towards the corolla opening) as are most Nicotiana flowers in nature.