Pigeons rival primates in number task

Trained on one-two-three, birds can apply the rule of numerical order to such lofty figures as five and nine

Pigeons, who aren’t even distant uncles to a monkey, have matched primates in a test of learning an abstract numerical concept.

WORKING BIRD Pigeons rival rhesus monkeys at putting groups of shapes in numerical order. The birds also show the primatelike quirk of taking longer and making more mistakes when clusters hold close to the same number of shapes.  William van der Vliet

Trained on one-two-three, the pigeons then had to put pairs of numbers up to nine in order, says comparative psychologist Damian Scarf of the University of Otago in New Zealand. Pigeons rivaled rhesus monkeys tested earlier at the same task, Scarf and his colleagues report in the Dec. 23 Science.

The results “suggest that despite completely different brain organization and hundreds of millions of years of evolutionary divergence, pigeons and monkeys solve this problem in a similar way,” says Elizabeth Brannon of Duke University, a coauthor of the original study of numerical order in monkeys.

Humankind may be pretty proud of its numerical prowess, but numbers — four succulent fruits versus eight, one lurking lion versus three — matter very much in animal life, too. Research is uncovering various kinds of number-related abilities in animals as diverse as the honeybee, mosquitofish, grey parrot, Plethodon salamanders and a waterbird called a coot.

So pigeons could be compared with other species, Scarf used Brannon’s numerical-order test, which baboons and lemurs as well as some monkeys have passed. For training, pigeons saw computer screens displaying sets of three images, each with one, two or three shapes. The shapes varied so that a bird couldn’t get the number order right just by pecking at increasing surface area. Scarf then rewarded birds for pecking in one-shape, two-shapes, three-shapes order.

It took Scarf more than a year of sessions every day, workday and weekend, to get three pigeons’ success rates up to 30 to 40 percent on ordering the three images. The way the exercise was set up, mere chance would have given the birds the correct answer only a bit more than 8 percent of the time, he says.

And no, he’s not sick of pigeons. “They’re very diligent workers,” he says.

To see if all that training gave the birds an abstract sense of numerical order, he offered them pairs of shape clusters to peck. Birds had trained only up to three, but Scarf included all pairs possible with numbers one through nine. For the toughest challenge — when both numbers in a pair were unfamiliar from training — birds pecked the images in ascending numerical order about 74 percent of the time, matching the monkey score.

Seeing such similarity on the same task, Scarf says, “helps us with the puzzle of identifying the origins of numerical ability.”The comparable success is a surprise considering that pigeons don’t seem to process order the way primates do in non-numerical matters, says comparative psychologist Dustin Merritt of Duke University. The difference shows up when pigeons and monkeys learn images in an order — say, A-B-C-D-E. Monkeys can put selected pairs in order (B-D, A-C), but pigeons usually need one of the end images for success.

All this numerical ordering isn’t technically counting, a complex business requiring specific tags for each number. Irene Pepperberg of Harvard University, however, has reported that a grey parrot named Alex that she worked with actually did count and could indicate which of two Arabic numbers was larger or smaller. It’s hard to compare Alex’s numerical prowess with that of other species, though: He indicated his choices in experimental tests by speaking in English.

Susan Milius is the life sciences writer, covering organismal biology and evolution, and has a special passion for plants, fungi and invertebrates. She studied biology and English literature.

More Stories from Science News on Life