Home & Garden Columns

Thinking Like a Bird: Jays, Hummingbirds and Memory

By Joe Eaton Special to the Planet
Tuesday April 11, 2006

The more scientists learn about non-human cognition, the blurrier the boundary between the human mind and various animal minds seems to become. And I’m not just talking about tool-making, intention-guessing, empathetic chimps. Some remarkable findings have emerged from the study of birds—and not necessarily the kinds of birds you’d expect. 

Episodic memory—memory that encodes particulars of what, where, and when—used to be considered exclusively human. Your recall of where you were when you heard about the attacks on the World Trade Center is an episodic memory (in my case, waiting for the elevator in the lobby of my office building). 

So is your recollection of where you last left your car keys. Animals, even bright ones like the great apes, were not supposed to be able to store and retrieve this kind of information. 

But biologists working with corvids—birds in the family that includes jays, crows, nutcrackers, and magpies—began to wonder about that. Some corvid species are food-cachers: they hide stashes of acorns or pine nuts in summer or fall to sustain themselves during winter and early spring, when other food is scarce. 

When they returned to their cache, had they made a random search or had they remembered where the items had been hidden? Better-than-chance retrieval performance suggested that birds like western scrub-jays, pinyon jays, and Clark’s nutcrackers had a well-developed spatial memory. 

The brains of these species have a larger-than-average hippocampus, an area thought to be responsible for processing memory. 

Do their memories have a time dimension, though—a “when” associated with the “what” and “where”? Nicola Clayton, working with western scrub-jays at Cambridge, thinks so. In an ingenious series of experiments, Clayton allowed captive scrub-jays to cache perishable food items—mealworms—and relatively non-perishable items—peanuts—in the lab. When given the opportunity to retrieve the goodies right after caching, her jays displayed a strong preference for the mealworms. 

But she found that the preference changed over time. If a jay had recovered a worm that was past its prime, in a subsequent trial five days after caching it would go for the more reliable peanuts.  

To rule out the possibility that, for whatever reason, more worm memories were lost than peanut memories, Clayton “taught” some of her subjects that worms did not in fact go bad by substituting fresh ones in the caches. Those birds continued to choose the worms at the five-day mark. In another variation, she accelerated the apparent decay rate of crickets, and found this resulted in a preference for nuts within a shorter time frame. 

Clayton has been careful to call what she has observed “episodic-like memory”, to avoid equating the human and corvid thought processes. But it sure looks like her jays were recalling not only where they had stashed the different kinds of food but when they had done so. 

(She has also documented something very like a theory of mind in scrub-jays, the ability to put oneself in the mental space of another individual. Jays that are more prone to pilfer other birds’ caches are correspondingly more likely to move their own caches if they were observed in the act.) 

Corvids have relatively big brains for birds (and scrub-jays have large hippocampi even for corvids). If you’d expect any kind of bird to be capable of memnonic prodigies, it would probably be a scrub-jay. However, episodic-like memory may not be unique to jays. Very similar processes have now been documented in, of all things, hummingbirds.  

In a study that recently appeared in Current Biology, Susan Healy and Jonathan Henderson of the University of Edinburgh describe their fieldwork with rufous hummingbirds in the Canadian Rockies.  

(The rufous hummer is an early spring migrant through the Bay Area; its close relative, the Allen’s hummer, stays to nest). Healy and Henderson placed eight artificial flowers in an alpine meadow patronized by hummingbirds. Some of the “flowers” were refilled with hummer food at 10-minute intervals, others at 20-minute intervals.  

Tallying visits by three male rufous hummers, the researchers found the birds could distinguish between the 10-minute and 20-minute “flowers” and remember their locations and when they had last drained them. Over several days, they reliably returned to the “flowers” just after they had been refilled; once again, a matter of what, when, and where.  

It makes sense for hyperactive birds like hummers to maximize their foraging efficiency. Return to a flower too soon, and the nectar won’t have been replenished; too late, and a rival may have beaten you there. With a long migration route and a short breeding season, rufous hummers can’t afford to waste time and energy in the search for food.  

Healy and Henderson point out that their male hummers were able to track the timing of nectar supplies while defending their territories and courting females. So you have not only episodic (oh, OK, “episodic-like”) memory but serious multitasking, all with a brain the size of a grain of rice.  

I don’t know how large a hummingbird’s hippocampus is, absolutely or relatively. But the bird doesn’t have a whole lot of neurons to work with. It may turn out to be not the size of the brain that enables these kinds of mental processes, but the complexity of the wiring. Smaller does not necessarily equate to dumber: the minuscule brain of the hummer appears to have the bandwidth to do what it needs to do.