How do flowers change when the animals that carry their pollen change? A new color may attract a new visitor, but that does not necessarily mean the visitor can reach the nectar and carry pollen. Examining changes in scent and flower shape as well can reveal what may be an intermediate stage in evolution.
A study published in 2025 by Katherine E. Wenzell and colleagues examined two red-flowered Mimulus species pollinated mainly by hummingbirds. Both species have naturally occurring yellow forms. The researchers compared pigments, scent, shape, genes, and bumblebee behavior to investigate how similar changes arise in separate species and the early stages of a shift in pollinators.
Comparing red and yellow flowers within the same species
The species studied were Mimulus cardinalis and Mimulus verbenaceus. A total of four lines, red and yellow for each species, were grown in a greenhouse. These experimental lines were derived from wild plants and maintained over generations; this was not an experiment in which flowers were repainted.
Yellow flowers had not only less of the anthocyanins responsible for red coloration, but also more carotenoids, yellow-family pigments. In both species, the same three carotenoids increased, and expression of the same two genes involved in their biosynthesis was higher.
However, the patterns of DNA changes were not identical in the two species. The authors interpret this as different changes producing similar colors while drawing on shared genetic machinery.
Yellow flowers also emitted more scent
Scent was collected for 24 hours from newly opened flowers and analyzed by GC–MS. Compared with red flowers of the same species, yellow flowers emitted about 2.3 times as much in M. verbenaceus and about 1.7 times as much in M. cardinalis.
However, the two species did not converge on the same scent. Their baseline scent amounts differed substantially, and none of the compounds detected in this analysis were shared by the two species. In one species, overall composition was relatively similar between color forms, whereas in the other it differed greatly.
What they shared was the direction of change: greater emissions in yellow lines. The study did not assess which smelled better to people or how strong people perceived the scents to be.
Yellow flowers received about twice as many visits as red flowers
Next, flower-naive buff-tailed bumblebees were introduced individually into a flight cage and offered four red and four yellow flowers of the same plant species. Visit behavior was recorded for 30 minutes, using 10 bees for each plant species.
For M. verbenaceus, 59 of 89 visits were to yellow flowers and 30 to red flowers. For M. cardinalis, 99 of 150 visits were to yellow and 51 to red. In both cases, yellow flowers received more visits, at a ratio of about two to one.
These counts include repeated visits by the same bees. An analysis accounting for differences among individual bees still found more visits to yellow flowers. However, considering only each bee’s first choice, six chose yellow and four chose red for each plant species, with no clear preference.
Scent alone cannot be identified as the deciding factor
The greater number of visits to yellow flowers cannot be explained solely by increased scent. Modeling how flower colors appear suggested that yellow flowers stood out more than red flowers against a green background. Because the flowers differed simultaneously in color, scent, shape, and other features, the individual effects cannot be fully separated.
Scent-detection responses were also examined using antennae from nine bees. Scent extracts from all four flower lines produced larger electrical responses than air alone. However, comparisons with the extraction solvent alone were inconsistent, and yellow-flower scent did not consistently produce stronger responses than red-flower scent within the same species.
Antennal responses do not directly indicate which flowers bees prefer. There was also a measurement limitation: the concentration of the reference odorant was low. These results therefore cannot establish that scent was irrelevant either.
Attracting a visitor is not the same as having it carry pollen
Yellow lines had wider flower openings in both species. However, changes in tube length differed: the tubes were shorter in M. cardinalis and longer in M. verbenaceus. There was no clear difference in nectar volume between red and yellow flowers of the same species.
Although the bees visited yellow flowers more often, they struggled to enter them and reach the nectar. They were also observed failing to make adequate contact with the pollen-producing anthers or the pollen-receiving stigmas. Sometimes, they forced their way in and tore the flower tube.
The authors propose an evolutionary pathway in which traits such as color that strongly change attraction shift first, followed by changes in scent, opening width, and patterns that guide body positioning. These yellow flowers may attract bees more effectively while still being partway toward a shape that allows efficient pollen transfer.
However, the study did not follow the actual sequence of changes over a long period. Other animals, such as hawkmoths, may also be involved in the wild, and visits and reproductive success need to be studied in natural populations. Laboratory preferences alone do not establish that a shift from hummingbirds to bees is complete.
Summary
Yellow-line flowers differed not only in pigments but also in scent emissions and opening shape, and received more bumblebee visits. Even so, the bees’ bodies and the floral structures were not yet well matched.
The study illustrates the possibility of an intermediate stage in which flowers attract a new partner and move toward a form suited to it, rather than color, scent, and shape changing simultaneously into a finished configuration. Scent is part of this process, but a pollination relationship requires a combination of standing out, being detected, and allowing visitors to reach nectar and contact pollen.
References
Wenzell, K. E., Neequaye, M., Paajanen, P., Hill, L., Brett, P., & Byers, K. J. R. P. (2025). Within-species floral evolution reveals convergence in adaptive walks during incipient pollinator shift. Nature Communications, 16, 2721. DOI: 10.1038/s41467-025-57639-3 (Full original article).
The research was supported by the John Innes Centre, the UK BBSRC, and other sources, and one author received fellowship support from the US NSF. The authors declared no competing interests. Original article: © The Author(s) 2025, CC BY 4.0.
