The sun’s final rays filter through the leaves as night falls in the dense, muggy rain forest. The descending gloam over Panama’s Barro Colorado Island obscures the towering, spiky kapok trees, the palms and the shrubs until human eyes can’t see much more than the small patches of starlit sky through the canopy above. Crickets commence their chorus as the howler monkeys hush for the night.

In the twilight, a nocturnal sweat bee, with bulging eyes, a metallic green head and a pale brown abdomen, emerges from her nest in a foot-long, hollowed-out stick. She’s hungry for nectar and pollen. But before she flits off, she turns to look back at the stick, which has a black-and-white-striped card above it, placed there by scientists. Nearby stick nests also have cards, but these are simply a flat gray.

After the bee flies off, zoologist Eric Warrant and his colleagues at Lund University in Sweden switch things around, moving the striped card to another nest. When the bee returns, she zooms right into the nest with the stripes, assuming it is hers and demonstrating that sweat bees spot and use such visual signals. “Even in the very dimmest intensities, they have no problem seeing this,” Warrant says. (He notes that if the humans tracking the insects did not wear night-vision goggles, they would “literally crash into trees” because it is so dark.)

The remarkable night vision of these bees (Megalopta genalis) stems only in part from eye adaptations such as larger lenses. Those do improve sweat bees’ light sensitivity. Still, the nocturnal insects find their nests at light levels where even those peepers should not be sufficient. Warrant has concluded that in addition to the bees’ eyes, the way in which their brain processes the little light available allows them to navigate after sunset.

For decades scientists assumed that most creatures must see the same dim, colorless nightscape that people do. They thought that nocturnal animals relied on other senses, such as smell and hearing. Today a new wave of research is overturning that assumption. “We always thought we knew how well animals saw in the dark, but very few people had actually looked,” Warrant explains. Once researchers started peering into this dark world, they discovered that a wide variety of species see a startlingly clear nightscape.

Moths, frogs and geckos, for example, can distinguish colors at night when researchers themselves see nothing but shades of gray. Being more sensitive to color variations gives them an advantage because hue is a much more reliable way to distinguish objects, in bright or dim light, than noncolor indicators such as intensity. It can help them find food, nests or mates in the dark. “It’s just amazing that so many animals can be active in dim light and still perform behaviors when we can’t,” says Almut Kelber, a sensory biologist at Lund.

The secrets to night navigation reside between eye and brain. Nerve cells in the optical systems of these animals add up scarce bits of light to create a brighter picture and carefully prune away other, noisy signals that would muddle the image. The cells perform these summations by grabbing input from neighboring spots in their visual field. They also sum up input from single spots over a long time period, essentially slowing down visual perception to make things much brighter.

In living color

The eyes of people, along with those of most other vertebrates and invertebrates, have cells that work as photoreceptors, detecting light coming from outside. The cells are called cones and rods. During the day, we use mainly cones, which send signals back toward the brain when hit by incoming photons of red, green or blue light. They give humans excellent color vision, but they do not respond much in the dark. In dim light, we rely on rods, which are more sensitive because they work together in groups, pooling the information from scant incoming light. They tend to distinguish only shades of gray, however.

Credit: Tami Tolpa; Source: “The Remarkable Visual Capacities of Nocturnal Insects: Vision at the Limits with Small Eyes and Tiny Brains,” by Eric J. Warrant, in Philosophical Transactions of the Royal Society B, Vol. 372; April 5, 2017

Warrant, Kelber and another Lund colleague, Anna Balkenius, were the first to show, in a 2002 study, that an animal had color vision at night. The researchers put insects called hawkmoths in a cage in the laboratory and trained them to associate either a blue or yellow artificial flower with a sugar-water reward. The zoologists started the tests under dusklike illumination, then turned down the light to levels as low as dim starlight. As dark as the surroundings became, the moths could still tell yellow from blue. Since that study, Kelber’s team has found nocturnal color vision in carpenter bees and geckos. She hopes to test for color vision in fruit bats and in owls, whose nocturnal hunting prowess has usually been ascribed to keen hearing or big eyes.

Frogs can see color in the dark as well, distinguishing blue from green. Animal physiologist Kristian Donner of the University of Helsinki in Finland and his colleagues tested European common frogs for phototaxis, a behavior by which the frogs typically hop toward light. Donner wondered if they would be choosy about the color of the light. Decades ago lab tests on frogs’ rod cells had shown that some specifically reacted to blue light, whereas others responded to green. To find out what the cell differences meant for frog behavior, Donner’s group placed 17 amphibians, one at a time, in a bucket with two windows on opposite sides. The scientists shined blue light in one side and green light in the other. Then they measured the frequency and direction of the frog hops at different light levels.

When the bucket was completely dark, the hops were random. But as soon as the researchers let in the least possible amount of light, the frogs showed a clear preference for green. “At the very limit for vision, they can still differentiate between blue and green,” Donner says. For human comparison, his students stuck their heads in the bucket and could not see any light, much less tell green from blue.

It’s not certain why the amphibians jumped toward green light. Perhaps, Donner speculates, frogs get clues from the stars. Starlight is made up of relatively long wavelengths, and green light wavelengths are longer than blue, so green coming into the bucket might hint at starlit open spaces and a route to escape from the container.

Starlit paths

If frogs indeed follow the stars, they would not be the only animals that do so. Dung beetles travel in a perfectly straight line on moonless nights, when the only light comes from stars. The movement is a good strategy for a beetle with a nice fresh bit of dung, says James Foster, a sensory biologist at Lund. It wants to leave the scrum of other beetles at the dung pat and find a quiet patch of ground to dig in with its prize. Going straight, rather than weaving or turning about, will get the beetle away from the pat as quickly as possible.

How do the beetles do it? Foster’s Lund adviser Marie Dacke, Warrant and other researchers had already discovered that the insects use what they can see above them to find their way around. The scientists put cardboard visors on the critters so they could not see the sky. Then they let the insects loose in a circular arena and tracked the way each traveled to the edge. When capped, the beetles took much more circuitous routes, indicating that something in the sky was important to them.

The researchers suspected the beetles might use the pattern of stars for orientation, like six-legged sailors navigating with constellations. To test this idea, Dacke and her colleagues brought the beetles, with their dung balls, to a planetarium where skylight patterns could be easily controlled. Under either a simulation of a full starry sky or just the bright streak of the Milky Way, the beetles sped straight to the circle’s edge in under a minute. They took longer if the galaxy was absent. It was the first time any animal was shown to orient itself using this band of stars. (After publication in the journal Current Biology in 2013, the work earned a tongue-in-cheek Ig Nobel Prize in Biology and Astronomy.)

More recently, Foster investigated how dung beetles might use the Milky Way to go in one particular direction. Seen from our planet, the galaxy’s thick band of stars is a fairly symmetrical line. From the beetles’ perspective, the line would look just the same when they are moving forward or backward. Yet the insects do not get turned around.

Foster suspected that the beetles kept track of subtle differences in light intensity between one end of the Milky Way and the other. When he analyzed photographs of the galaxy taken from the beetles’ South African habitat, he found that the intensity of light from the northern and southern ends of the Milky Way indeed differed by at least 13 percent and sometimes much more, depending on how he processed the images.

To test this effect on the beetles themselves, Foster built a simplified, artificial Milky Way out of single-file LED lights on an arch over an arena. He could vary the intensity of light on each side. The beetles could go straight if he gave them a 13 percent contrast between one end of the bright line and the other but wavered if the contrast dropped below that. This result indicated the animals should be able to tell the two ends of the real Milky Way apart.

Credit: Tami Tolpa; Source: “Neural Summation in the Hawkmoth Visual System Extends the Limits of Vision in Dim Light,” by Anna Lisa Stöckl et al., in Current Biology, Vol. 26, No. 6; March 21, 2016 (flower images)

Signal boosters

In addition to beetles and bees, a number of other animals are now known to see remarkably well in dark environments: cockroaches, lantern fish, cuttlefish, frogs and nocturnal primates such as owl monkeys. So neuroscientists are turning to the question of how they do it. Bigger eyes collect more light, for example, but do not gather enough photons to explain the highly sensitive night vision that scientists have documented. Other visual processing must take place after the rods have absorbed incoming light. In particular, animals must be able to overcome or filter out visual “noise” created by photoreceptor activity that does not reveal anything useful about the visible world.

Noise in the visual system comes from a few different sources. One, called photon shot noise, happens when only a few photons come into photoreceptors. Because those light packets tend to arrive sporadically, they create a variable, unreliable picture. It’s as if you shone three or four flashlights around the ceiling of the Sistine Chapel at night. You would hardly be able to appreciate Michelangelo’s complete masterpiece.

A second source of noise arises from the molecular interactions in the photoreceptors themselves. A photoreceptor senses light when an incoming photon hits a molecule called rhodopsin. But every so often—once a minute, at most—a rhodopsin molecule is triggered by accident, or another part of the pathway misfires. This is called dark noise because it can happen even in pitch-black conditions with your eyes closed. A third source, transducer noise, results from variation in the timing and strength of the visual system’s response to a single real photon.

Noise isn’t a big problem in broad daylight, because the tremendous volume of photons hitting the eyes overwhelms these slight variations. In the dark, however, animals need a strategy to boost the signal to similarly noise-overwhelming levels. They do so by summing up the signals they get from individual photoreceptors across space and time.

Spatial summation works like this: Imagine you are at a concert where 1,000 fans are waving their illuminated cell phones with excitement. You can’t see the light from each individual phone all that well. If every group of 50 concertgoers combined the light of their phones into a single, brighter spotlight, you’d see those 20 spotlights really well. The retina—the sheet of tissue that contains rods and cones—does the same, pooling the input from numerous rods into a single, bigger signal that gets sent to the brain. At the concert, you lose the picture of each individual person waving a phone, and the same thing happens in spatial summation; the resulting image is brighter but also coarser.

Temporal summation also increases brightness. Rods slow their activity down, summing up the input from incoming photons over, say, 100 milliseconds. Again, there’s a trade-off. This type of summation makes it easier to detect objects, but it blurs them when they move.*

In some insects, spatial and temporal summation happen in parallel, and it occurs in cells farther back toward the brain, according to biologist Anna Stöckl, now at the University of Würzburg in Germany. Stöckl, when she was a graduate student under Warrant, positioned hawkmoths in front of a computer screen showing a pattern of scrolling black-and-white stripes. Then she cut a tiny hole in the back of each moth’s head and poked electrodes into its cells. Her goal was to stimulate the photoreceptors with each alternating stripe and compare their activity with that of other nerve cells deeper in the brain, in the optic lobe. This area gets the signal after any processing or summation has occurred, so differences between the unprocessed “input” at the photoreceptor and the “output” in the optic lobe would indicate that the brain altered the visual signal.

Comparing these input and output values, Stöckl calculated that when she transferred moths from light to dark, the size of a “pixel” in their optic lobe quadrupled, showing that they used spatial summation. She also found that moths used temporal summation, slowing their vision in the dark so they added up input over 220 milliseconds. The combination allowed the hawkmoths to see well at light levels 100 times dimmer than when summation was not in use, Stöckl reported in a 2016 paper.

“This hasn’t been shown in any other animal apart from hawkmoths, but the principle is so basic that it would be hard to believe it isn’t widespread,” Warrant says.

Another approach that animals use is to filter out noise, say scientists who have investigated visual noise-canceling methods used by mice and monkeys. While not on a par with hawkmoths, these mammals do reasonably well at night. Researchers have found there are at least two threshold points on a path between their photoreceptors and the brain that allow only strong signals through and reject those likely to be noise. Midway along this path are gatekeepers called rod bipolar cells. These cells, it turns out, are tuned to send the “photon detected” signal onward only if they receive significant input from rods. Several incoming photons at once are strong enough. But single photons, and much of the noise in the system, might not be. A second cellular gate lies deeper in the optic system on this same path. This gate blocks errant signals that are missed by the first one or that arise after that point. The result is nearly noiseless vision, says Petri Ala-Laurila of the University of Helsinki, one of the scientists who identified the process.

In dim light, sweat bees (top) detect detailed patterns, dung beetles (middle) navigate by starlight, hawkmoths (bottom left) blend visual signals to brighten images, and southern pig-tailed macaques (bottom right) filter interference from what they see. Credit: Alamy (sweat bee); Chris Collingridge (dung beetle); Malcolm Schuyl Alamy (hawkmoth); Anup Shah Getty Images (macaque)

Forward-looking

Despite all this research, Warrant says, scientists are just beginning to understand animals’ ability to see in the dark and how they manage to do so. Studies of the genes and light-sensitive molecules that nocturnal animals possess can offer new clues. For example, some night-active lemurs have genes and pigments that indicate their eyes might be sensitive to blue or green, which could help them distinguish blue seeds and green leaves in twilight. And some bats—which, contrary to popular wisdom, are not blind—also possess genes tied to color vision.

Still, having the genes and molecules to detect color does not prove an animal’s brain uses that information after twilight. For example, some light-sensitive molecules are involved in maintaining bodily rhythms that have nothing to do with vision. Therefore, scientists still need to perform behavioral experiments, such as those carried out on the hawkmoths and frogs, to show those molecules play a role in night sight. That work may indicate that the molecules are not used in the dark—or it could reveal sight-enhancing tricks researchers have not yet envisioned.

Read how animal night vision was used to develop car cameras.

*Editor’s Note (6/24/19): This paragraph was revised after posting. It originally referred to a shooting star as one point of light at any given moment. A shooting star, or meteor, is a streak of light rather than a single point, created when a meteoroid enters Earth’s atmosphere.