During the 2017 eclipse, giraffes stampeded, Galápagos tortoises started mating, gibbons sang unfamiliar songs, and gorillas quietly got ready for bed, all within the same two and a half minutes. More than three-quarters of the animals studied reacted, and the reason behind it surprised the scientists more than the behaviour itself.

On the afternoon of 21 August 2017, at approximately 2:41 in the afternoon local time, the sun disappeared entirely behind the moon above Columbia, South Carolina. The state of near-total darkness lasted for roughly two minutes and thirty-six seconds. During that specific window, at a mid-sized zoo on the outskirts of the city, a team of researchers stood in coordinated positions with clipboards and stopwatches, waiting to see what a collection of seventeen different species of mammals, birds, and reptiles would do when the middle of the day briefly became the middle of the night.

What they saw, on the surviving primary-source record of the study, was one of the more genuinely strange episodes ever documented in modern comparative zoology.

What the animals actually did

According to the peer-reviewed study published in March 2020 in the journal Animals by Adam Hartstone-Rose, Edwin Dickinson, Lisa M. Paciulli, Kaitlyn Leonard and colleagues at North Carolina State University, alongside Ashley Deutsch of the University of Florida, Leon Tran of the University of Hawaii, and Grace Jones of the University of South Carolina, titled “Total Eclipse of the Zoo: Animal Behavior during a Total Solar Eclipse”, the researchers had established baseline behavioural observations for each of their seventeen species in the days leading up to the event. They knew what each animal ordinarily did in the middle of an August afternoon in South Carolina. They then compared those baselines with what each animal did when totality actually arrived.

Thirteen of the seventeen species, roughly seventy-five per cent of the sample, changed their behaviour in a measurable way. Not slightly. Dramatically.

The Galápagos tortoises in the reptile enclosure, which had been sitting in the sun in the way Galápagos tortoises sit in the sun, began mating. Actively. In the middle of the day. This is not what Galápagos tortoises ordinarily do at that hour on a summer afternoon. The siamang gibbons in the primate house began producing extended vocalisations of the kind their species uses at unusual times of day, on tones the researchers observing them had not previously recorded from those particular individuals during their baseline observations. The gorillas, on the other side of the primate compound, gathered together and began the sequence of settling behaviours they normally perform at bedtime. The Hamadryas baboons ran and paced in coordinated groups, clustering around their juveniles. The flamingos gathered in the centre of their enclosure and vocalised. The rainbow and coconut lorikeets became loud. And the giraffes, which are large, slow, ordinarily very placid animals that spend their days walking calmly around and eating leaves, briefly ran.

Giraffes do not usually run. The lead author of the study, Dr Adam Hartstone-Rose, later told reporters that giraffes are, in his precise words, kind of delicate, and that when they run, it is because they are running from a predator. There was no predator. There was only, briefly, an unscheduled darkness in the middle of the sky.

Related watch – what happens to our bodies during a solar eclipse?

What the scientists were looking for

The research team had, before the eclipse arrived, worked out a classification framework for the responses they expected to observe. Any behaviour a given animal displayed during the eclipse would be sorted into one of four categories. The first was normal, meaning the animal did essentially what it had been doing all afternoon. The second was evening or nighttime behaviour, meaning the animal did what it would ordinarily do at dusk or at the transition into sleep. The third was novel behaviour, meaning the animal did something new that it did not appear to do at any regular time of day. The fourth was apparent anxiety, meaning the animal displayed movements, vocalisations, or clustering that would be classed, on the standard ethological literature, as fear responses.

The team’s private expectation, on their published pre-registration of the study, had been that they would see a considerable amount of category three. A total solar eclipse is, in the strict evolutionary sense, an event no living animal on Earth has ever ancestrally encountered as a recurring stressor. Total eclipses over any specific point on the planet’s surface are separated by, on average, roughly three hundred and seventy-five years. No captive animal in any zoo has ever previously experienced one. Neither has any of their evolutionary ancestors, on any timescale their genomes could reasonably have adapted to.

Which meant the sensible prediction was that the animals would treat the eclipse as what it in fact was, a genuinely unprecedented environmental event, and respond with novelty behaviour of a kind ordinarily reserved for the appearance of something the animal has never encountered before.

What actually surprised the researchers

That is not what the animals did.

The single largest response category across the thirteen behaviourally-changed species, on the study’s final analysis, was not novelty. It was evening behaviour. Eight of the seventeen species performed some version of their standard dusk-and-bedtime routine, roughly nine hours ahead of schedule. The second largest category was apparent anxiety, exhibited most notably by the baboons and the gorillas. Novel behaviour, meaning genuine what-on-earth-is-this responses, was the least common category. Almost none of the animals treated the eclipse as a new event. Most of them treated it as evening arriving early.

What that finding suggested, on Hartstone-Rose’s analysis, is genuinely counter-intuitive and worth sitting with. The animals were not, in the way the researchers had expected, recognising the total solar eclipse as an unprecedented astronomical phenomenon. They were reading it as an ordinary environmental signal. Light was fading. Temperature was dropping. The combination of cues the animal’s internal circadian system used to identify the transition into dusk was arriving, in a compressed form, in the middle of the afternoon. And in most cases, the animal’s internal machinery simply performed the routine it had evolved to perform when those cues arrived. Time to gather the group. Time to head toward the sleeping quarters. Time to sing the evening call. Time, in the case of the Galápagos tortoises, to do the thing tortoises do in the twilight cool.

The picture is not, however, uniform across the whole animal kingdom. According to a separate peer-reviewed study of the same 2017 eclipse published in Biology Letters in November 2018 by Cecilia Nilsson, Kyle G. Horton, Adriaan M. Dokter, Benjamin M. Van Doren and Andrew Farnsworth at the Cornell Lab of Ornithology and the University of Oxford, titled “Aeroecology of a solar eclipse”, the response of wild flying animals across the United States looked different. Using data from 143 weather radar stations, eight of which sat inside the path of totality, the Nilsson team measured the airborne movement of birds, bats, and flying insects across the continent during and around the eclipse. What they found was that daytime activity decreased sharply as totality approached, but that nighttime activity, meaning the specific patterns of migratory flight and roosting movement that usually appear at dusk, did not increase. The wild animals overhead did not treat the eclipse as dusk. They simply stopped.

Nilsson’s own interpretation, published alongside the study, was that the flying animals may have been reading the eclipse conditions as something closer to an incoming storm rather than the transition into night. According to Cornell University’s own reporting on the study, published in the Cornell Chronicle in November 2018, Nilsson noted that the specific combination of rapidly darkening sky and dropping temperature is closer, in terms of the environmental cues a bird’s or insect’s nervous system uses, to weather deterioration than to the ordinary end of the day.

Which taken together, on the combined evidence of both studies, points at a specific and quietly interesting proposition. The animals were not, in any case, seeing the eclipse. They were reading the environmental signals it produced, and the reading depended entirely on what environment they were in. Inside a safe zoo enclosure, with familiar surroundings and no immediate threat, the signal read as dusk. Overhead, in the wild, with no safe roost immediately in reach, the signal read as bad weather. Same event. Same cues. Different context, different interpretation.

Which leaves the whole story with a small and quietly humbling observation about what animal senses actually do. They do not, on the accumulated evidence of the past century of comparative ethology, understand what is causing the environmental signals they respond to. They understand only the signals themselves, and only in the specific context they find themselves in when the signals arrive. A dropping temperature and a fading light is either dusk or a storm, depending on where you happen to be standing. It is not, on any animal nervous system on Earth, the moon briefly moving in front of the sun.

Kiran Athar is not a zoologist or an ethologist. She writes about biology, animal behaviour, and the ordinary corners of the natural world where the two intersect, drawing on peer-reviewed research and primary-source scholarship.