Most flowering plants have evolved arrays of tiny cone-shaped cells on their petals. But the adaptive value of these structures for plants has never been clear. Why did they evolve? Heather Whitney, Lars Chittka, Toby Bruce and Beverley Glover recently tested the idea that these conical structures have evolved to act as `handles' for pollinators to grab on to as they approach from the sky.

First the team wanted to see if bumblebees could distinguish between flowers just by the feel of conical cells. To do this, they used mutant white snapdragon plants. These mutants are identical in sight and smell to common white snapdragons but have flattened cells instead of conical cells on their petals. Whitney and her colleagues provided the bees with smooth flowers (from the mutant snapdragons), containing a small tube of sweet nectar at the base of the petal, and bumpy normal snapdragon flowers flavoured with bitter quinine. The team let single bees wander over the flowers and repeatedly sample nectar from both types of flower. Each time an individual landed on a flower, the researchers noted how long the bee lingered and whether or not she made any effort to drink. If the bees really could feel the texture differences between the smooth and bumpy flowers, they would associate landing on a bumpy flower with bad tasting nectar and vote with their wings, taking off to find a better tasting flower. Indeed, after a few trials, individuals started taking off as soon as they came into contact with the bumpy surface. The animals were clearly feeling their way to a decision by gauging the shape of petal bumps.

In the next set of experiments, the researchers tested whether bees could discriminate fine textures without any other biological or chemical cues present. They made epoxy discs with bumpy and flat surfaces closely approximating those found on actual petals. Then they used these artificial flowers in place of actual snapdragons. The researchers tried coupling bitter nectar with bumps and sweetness with flat cells (and vice versa). Regardless, the little foragers learned not to waste time on any surface that they had learned to associate with a bitter reward.

So it seems that bees can discriminate between petals on the basis of conical cell texture. But why have plants evolved these structures? After all,there are more efficient ways to attract the attention of pollinators without forcing them to land. The team wondered whether these structures might provide good grip for insects landing on otherwise slippery plant surfaces. To test this, Whitney and her colleagues presented bees with horizontal conical and flat-celled epoxy `flowers', each providing a sugar reward. The bees showed no preference for either surface. Then the team gradually tilted the landing sites. As the steepness of the landing surfaces increased, the bumblebees shifted their foraging to food sources with bumps. The team then went back and did the same experiment with actual snapdragon flowers. Again, when the landing surface was tilted, the bees showed an increasing preference for flowers with conical textures. In both cases, high-speed videography confirmed that the bees were `slipping and sliding' on steep surfaces with flat cells,confirming that the conical cells make it easier for approaching pollinators to grab onto a flower.

The experiments in this study are very simple but also very elegant. This work stands out as a great example of how excellent science can be still done by doing simple manipulations without a lot of fancy equipment. This work is also significant because it effectively tests the question of `why?' in evolution. This is not always an easy question to unambiguously answer. But in the end, it's one of the most important questions that we can ask as biologists.

Whitney, H. M., Chittka, L., Bruce, T. J. A. and Glover, B. J. (
). Conical epidermal cells allow bees to grip flowers and increase foraging efficiency.
Curr. Biol.