Animal intelligence is a fascinating topic and there have been many attempts at studying it. Many of the individual studies look at one or other specific trait that we associate with intelligence in one species and the traits studied can differ from species to species, making general conclusions hard to arrive at. Ed Yong reports on a massive multinational study that looked across many species at one aspect of intelligence (self control) as demonstrated by two specific tasks. (You can read the paper on which his article is based here.)

Evan MacLean, Brian Hare, and Charles Nunn from Duke University … led a international team of 58 scientists from 25 institutes in studying the evolution of one mental skill—self-control—in 567 animals from 36 species.

Chimpanzees, gorillas, baboons, marmosets, lemurs, squirrels, dogs, elephants, pigeons, parrots and more tried their hands (or trunks or beaks or snouts) at the same two tasks. “It was literally a mouse-to-elephant study,” says MacLean, “or at least a Mongolian-gerbil-to-elephant study.”
…

The team focused on self-control—the ability to stop doing that, put that down, eat that later. Animals exercise it when they stop themselves from mating in the presence of a dominant peer, when they forgo an existing source of food in favour of foraging somewhere new, or when they share resources with their fellows. In humans, a child’s degree of self-control correlates with their health, wealth, and mental state as adults. It’s important.

What were the two tests for self-control that they used? One was based on a test done by Jean Piaget where he studied babies when he “repeatedly put a toy under a box in front of some infants, and then moved it to a second box.”

He found that babies under 10 months of age would keep on searching under Box A, despite what they had seen. They couldn’t resist their old habit to do something flexible and different; that ability only kicks in around our first birthday. MacLean, Hare and Nunn’s team gave this “A-not-B” test to their animals, using food rather than a toy.

They also tried a second task, where animals had to reach round the side of an opaque cylinder to get at food within. The team then swapped the opaque cylinder for a transparent one. Now, the animals had to hold back their natural instinct to reach directly for the food (which would have knocked the cylinder over), and go around as before.

The team tested all their animals on one or both tasks, and compared their performance to traits like brain size or group size.

So what were the results?

They found a few surprises. For example, the animals’ scores correlated with the absolute but not relative sizes of their brains. In other words, it didn’t matter whether the animals’ brains were big for their size, but whether they were big, full-stop.

The dependence on absolute brain size is both surprising and not surprising. Surprising because there has been a long-standing belief that beyond a certain size, what matters is the size of the brain relative to the body. Not surprising because if one thinks of the brain as a computer, then a larger brain means potentially more neurons and more complex neural networks and thus greater brain power. Another possibility is that what may be important is the size of specific regions of the brain, not its overall size.

The team also looked at what drives the development of intelligence.

The team also tested two leading explanations for the evolution of primate intelligence. One idea says that our smarts evolved so we could keep track of the relationships within our complex social groups.
…

Instead, the team found more support for a second idea: that primate intelligence was driven by the need to keep track of a wide range of food like fruit, which vary by place and season.

Of course, there are criticisms and caveats. Intelligence is a complex phenomenon and hard to pin down and it is quite possible that the two tasks being looked at were too narrow to be truly representative of intelligence.