Researchers want to understand the cacophony that erupts when 16,000 of the birds congregate at night before roosting in a Bothell wetland.

Share story

BOTHELL — If a congregation of crows is called a murder, then the nightly gathering at North Creek Wetlands could be described as a massacre.

More than 16,000 of the birds converge on the streamside thickets every evening, descending in a raucous whirlwind of feathers, claws and rushing black shadows. Picture Tippi Hedren being mobbed in Alfred Hitchcock’s horror masterpiece “The Birds.” Subtract the malice — but multiply the intensity by a hundred — and you’ve got a sense of what it’s like to stand in the midst of the maelstrom.

Students at the adjacent UW Bothell campus dash between buildings to avoid being spattered. The cacophony of caws is so loud it drowns out the roar of Interstate 405 a scant quarter-mile away.

“It’s really spectacular,” said UW Bothell biologist Douglas Wacker. “They’re coming from all directions, and the sky is just black with crows.”

Intrigued by the living laboratory in their backyard, Wacker and acoustic engineer Shima Abadi recently launched a project to eavesdrop on the birds and try to figure out what all the ruckus is about.

Before the crows settle into their treetop perches for the night, they mill around on campus. Big groups gather on the rooftops and athletic fields and line the eaves like gargoyles. The din is intense — but why are the birds so talkative at that time of day? Are they greeting old friends? Discussing their day? Jockeying for dominance?

“We want to see if there’s a relationship between their calls and their behavior,” Abadi explained. “When they say caw-caw-caw three times, does that mean food, or when they say it four times, does that mean danger?”

Studying crow vocalizations isn’t new, but the UW team is using a sophisticated array of instruments to tease out subtleties in pitch, timing and social setting that likely influence the messages crows are trying to convey.

“I’m not ready to say crows have language,” Wacker said. “But I do think they have the capability to transmit complex information via their vocalizations.”

To the casual listener, all “caws” might sound the same. But crows can produce an impressive array of noises, from rattles and bell-like notes to rapid-fire caws or drawn-out croaks. Studies at The Cornell Lab of Ornithology found individual crows have distinctive voices, just like humans.

Why would crows make so many different sounds if they don’t carry meaning?, Wacker asked. “It doesn’t make sense to me that they would have evolved such elaborate vocalizations if it didn’t serve some communication function.”


Scientists have inferred meanings for only a few categories of vocalizations, like the guttural caws that incite crows to mob predators and the plaintive squawks young birds employ to beg for food. The majority of crow talk remains impenetrable to human observers, said UW wildlife biologist John Marzluff.

“There’s definitely quite a mystery in terms of how much information is encoded in these calls.”

Marzluff isn’t directly involved in the new project, but his own research has helped reveal the surprising brain power of crows and ravens. In one study, he showed that crows can recognize human faces and pass the information across generations.

That intellect, coupled with a natural wariness, makes wild crows a tough subject for study. Most sound recordings are made by people with microphones, but the presence of a person can change the way crows respond, Wacker pointed out.

Abadi drew on her earlier work recording whale calls to devise a remotely operated system with four recorders. Software filters out background noise and triangulates each call to the crow who uttered it. By comparing the audio recordings with video images, the researchers hope to see how crows respond to specific calls in a natural setting.

Even though the instruments don’t require a human operator, it’s still easy to spook the crows. When the team set up the equipment recently on the roof of the science building, not a single crow landed there that evening. Eventually, the birds got used to the gear and resumed their pre-roost, rooftop soirees, but curious crows pecked holes into one microphone’s foam casing.

The researchers have also started conducting experiments in which they play the recorded calls and watch how other crows react.

Much of the work and data analysis is done by undergraduate students, which gives them valuable hands-on research experience, Wacker said. It’s also convenient to have the roost so close at hand.

“We know exactly where the crows will be,” he said. “They’re like clockwork, showing up in the same spot every evening.”

Habitat changes

It’s only been in the past decade, though, that the birds started flocking to the area.

The historic flood plain was a network of meandering channels and backwater lakes where North Creek flowed into what was called Squawk Slough.

Beginning in the late 1800s, the land was logged, then later drained and diked for farming. More than 50 acres were restored and replanted with native vegetation as part of the UW Bothell construction in the late 1990s. By 2008, the trees were tall enough to entice the crows to abandon their old roost on Foster Island near Washington Park Arboretum.

Most of the crows who spend their days foraging in neighborhoods from Seattle to Everett and the Snohomish Valley wing their way to Bothell as dusk descends during the fall, winter and early spring, when they’re not breeding, Marzluff said.

“If you put about a 30-mile radius around it, you’re going to bring birds in from that far away.”

Some crow roosts in Oklahoma and Nebraska attract as many as 75,000 birds, with the biggest on record topping out at about 2 million, said Cornell crow specialist Kevin McGowan.

The birds probably flock together because there’s safety in numbers, he said, but there’s more to it than that.

“There’s a social component that we don’t understand very well,” McGowan said.

He’s convinced crow language is largely tonal, like Chinese. “Caw-caw-caw might mean three very different things to the crows, depending on inflection, or changes in pitch,” he said. Posture and body language may also add nuance.

Crows don’t think like humans, so word-for-word translation of crow speech will probably never be possible, McGowan said. But he’s excited to see how much the UW team will be able to decipher.

“Crow communication is probably a lot more complex than we give them credit for,” he said. “Even after 30 years, I see new things every day, so there’s a lot to learn about these birds.”