A group of British elementary school pupils, some aged just eight, have become the youngest science writers to have a paper accepted by a peer reviewed journal. They discovered that bumblebees are able to use a combination of color and spatial patterns to recognize flowers that will provide nourishment.

The project was organized by University College London neuroscientist Beau Lotto and Dave Strudwick, head teacher of Blackawton Primary School, but carried out by 25 students aged between eight and ten. The idea was to allow the students to choose their own question for study, with the aim being that they learned about the process of science rather than simply learning facts.

The students first picked bumblebees as a subject, then hunted for a question that hadn’t been explored in a previously published paper, settling on whether the bees could use patterns rather than merely colors in looking for flowers that still had a healthy supply of nectar.

To test this, the students set up four grids, each of 16 circles, with the central four lit in either blue or yellow each time, and the surrounding 12 lit in the other color. The central four of a grid always contained sugar water, with surrounding 12 containing salt water, regardless of the color combination. The idea was that the five bees used in the test would always want to locate the sugar water and would learn to hunt by position rather than color.

After a training period for the bees, the students began measuring the bees’ performance and discovered they were able to select a central circle 90.6% of the time. They also discovered that all but one of the bees would overwhelmingly go for circles of a particular color each time. That suggested that the bees were capable of detecting color, and had a personal preference, but were still able to deduce that positioning was key.

However, in a second experiment, the students lit the central (sugar water) circles of each grid in green. This time round, only two bees preferred the green, and as a whole the group only selected green 30.9% of the time, not much above the 25% they’d have achieved by selecting at random. (Results illustrated above.)

In a final experiment, the students reconfigured the grids so that in each case the four correct colored circles were placed in the corners rather than the center. This time the bees got the correct circle 40.1% of the time, suggesting that they weren’t able to decipher the one consistent rule across all three experiments: that the sugar water was always behind the color that appeared the least often in the grid. The two bees that did get the right answer in the second experiment also failed here, suggesting their previous correct answers were down to positioning rather than color.

The overall conclusions? That the bees were able to solve puzzles by learning pattern-based rules, but not perfectly; that different bees adopted different strategies; and that some bees may allow a personal preference of color to override the “logical” solution each time.

The adults were so impressed with the study and the resulting work from the children that they submitted the paper to Biology Letters, a journal of the Royal Society. The journal sent the paper for review by a neural science psychologist at New York University and an animal behavior expert from the University of Exeter. They concluded that, although simplistic, the paper was sound and contained genuinely fresh scientific results:

The experiments are modest in scope but cleverly and correctly designed and carried out with proper controls to avoid possible artefacts. They lack statistical analyses and any discussion of previous experimental work, but they hold their own among experiments carried out by highly trained specialists. The experimenters have asked a scientific question and answered it well.