Chickens are loquacious creatures, and Kevin Mitchell would know. He oversees the care of about a million of them on Wilcox Farms properties in Washington State and Oregon.
Mitchell says the birds have “patterns of speech” that reveal a lot about their well-being. They are usually noisiest in the morning—a robust concert of clucks, chortles and caws. “When I hear that, I know they are pretty healthy and happy,” Mitchell says. In the evenings when they’re preparing to roost, the chickens are much more mellow, cooing softly. When a hen lays an egg she celebrates with a series of staccato clucks, like drumbeats, culminating in a loud “buck-caw!” If chickens detect an aerial predator—say, by spotting the shadow of a hawk or eagle—they produce a short, high-pitched shriek. And they have a distinct warning for terrestrial threats: The repetitive clucking most people associate with chickens is in fact a ground predator alarm call.
One morning many years ago Mitchell entered a chicken house and found it oddly calm and quiet. Instead of making the usual ruckus, the birds were murmuring and shuffling lethargically. He soon discovered that an automated lighting system had failed and the lights had not switched off the night before; the chickens were sleep-deprived. If he had only been able to eavesdrop on the flock, he might have known much sooner that something was amiss.
On supporting science journalism
If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.
Over the past five years, engineers and poultry scientists at The University of Georgia and Georgia Institute of Technology have been collaborating to help farmers like Mitchell make better use of the information latent in chicken chatter. In a series of studies published between 2014 and 2016, Georgia Tech research engineer Wayne Daley and his colleagues exposed groups of six to 12 broiler chickens to moderately stressful situations—such as high temperatures, increased ammonia levels in the air and mild viral infections—and recorded their vocalizations with standard USB microphones. They then fed the audio into a machine-learning program, training it to recognize the difference between the sounds of contented and distressed birds.
When chickens detect an aerial predator such as a hawk, they give a short, high-pitched alarm call. Credit: Courtesy of Carolynn “K-lynn” Smith, Macquarie University
So far, the software can detect when the chickens are uncomfortable due to heat stress as well as identify their “rale” sounds—a soft gurgling produced when mucus from a respiratory infection clogs airways—with near perfect accuracy. “A lot of poultry farmers we have worked with say they can hear when something is wrong with a flock, but they can’t tell us exactly how they know that,” Daley says. “There’s a lot of subtlety. We’re learning that there are changes in the frequency of the sounds and the levels—the amplitude or loudness—that the machines can pick up on.” The IEEE (aka Institute of Electrical and Electronics Engineers) published the results of these ongoing studies in conjunction with the annual IEEE Global Conference on Signal and Information Processing.
“It’s really interesting work, fairly ingenious and logical,” says Wallace Berry, a poultry scientist at Auburn University’s College of Agriculture who was not involved in the studies. “Chickens are a very vocal species, and as a poultry farmer you can always use more data to make better decisions. This is a great way to continuously filter all the information available in a chicken house and learn as soon as possible that something is wrong.”
Carolynn “K-lynn” Smith, a biologist at Macquarie University in Australia and a leading expert on chicken vocalizations, says that although the studies published so far are small and preliminary, they are “a neat proof of concept” and “a really fascinating approach.” It’s “incredibly important to find new ways of monitoring the health of chickens,” she adds. “There are billions of animals in intensive farming around the world. We need more ways to carefully define what it means, and what it sounds like, for a chicken to be stressed.”
When chickens notice a predator on the ground, they warn each other with a series of rapid clucks. Credit: Courtesy of Carolynn “K-lynn” Smith, Macquarie University
People have been living with chickens for at least 6,000 years, and the global domesticated chicken population now exceeds 19 billion. But despite our lengthy shared history, few people have given serious consideration to the potential meaning of chicken vocalizations. Between the 1950s and 1980s ornithologists Nicholas and Elsie Collias of the University of California, Los Angeles, catalogued more than 24 distinct chicken calls and their probable meanings, based on careful observation. It was not until the 1990s that researchers such as Chris Evans (Smith's adviser) at Macquarie began controlled experiments to investigate the functions of these calls.
Evans, Smith and other scientists have discovered chicken communication is far more complex than previously realized. By attaching wireless microphones to chickens (using bra straps), for example, Smith and her colleagues revealed roosters do not reflexively shriek every time they spot an aerial predator—after all, calling out makes them more vulnerable. Instead, the roosters survey the situation: If there are females nearby, they usually sound the alarm; if they are alone or surrounded by other males, they often stay quiet—and they are much more likely to cry in warning if they are also able to take cover beneath, say, a bush. Could there be more nuance in chicken vocalizations that we do not yet understand? “It’s definitely possible,” Smith says.
Next, Daley and his collaborators want to challenge their algorithms to extract even more information from chicken sounds by changing an array of environmental variables such as access to food and water. But in recent trials, after shifting from small, highly controlled experimental setups to commercial chicken houses, they encountered a problem. Most industrial chicken houses have high levels of background noise, largely from heaters and giant air-circulation fans. The computer program Daley and his colleagues designed sometimes has trouble detecting subtle changes in chicken vocalizations amidst the acoustic chaos. The new challenge is teaching the software to zero in on the chickens—somewhat like focusing on a single person’s speech at a crowded party.
Chickens make this sound when "tidbitting," a behavior in which they repeatedly pick up and drop morsels of food to either attract mates or teach chicks what to eat. Credit: Courtesy of Carolynn “K-lynn” Smith, Macquarie University
If they eventually overcome this obstacle, the technology should be easy to incorporate into existing poultry farms. A modern chicken house is already a high-tech affair wired with all sorts of sensors, notes Wilcox Farms’ Mitchell. Poultry farmers can monitor and modify lighting, temperature, ventilation and automated feeding systems from their phones and laptops. Mitchell thinks adding an audio component could be useful.
“Some farmers tell me that despite all the gadgets, their most important piece of technology is a five-gallon bucket,” Berry says. “They turn it over and sit on it, and watch the birds for hours. They learn what a normal house looks and sounds like. If the chickens are content, there is a certain way they sound. If they are cold or hot, there are certain sounds they make. This is what Daley is trying to do in an automated way. It makes perfect sense. The kinds of systems he is designing are so unobtrusive that I don't see any problem integrating them.”
“In Georgia poultry is a big industry, and almost every company that raises chickens has a presence here,” Daley says. “We’ve talked with a lot of them and they all think it would be useful. The ecosystem to do this already exists. We just need to perfect the technology.”