THERE鈥橲 a German expression which translates as 鈥渁ll cats are grey at night鈥�. It鈥檚 certainly true for humans. As night falls, the colour-detecting cone cells in our eyes switch off, the rod cells take over and the world turns to fuzzy black and white 鈥� until we go indoors and switch on the lights.
It has always been assumed that nocturnal animals also see the world in black and white, albeit far more clearly than us. So when animal biologist and vision specialist Almut Kelber began studying nocturnal vision in geckos and moths, she was intrigued to discover that some species were actually seeing in colour.
Kelber and her colleagues at the vision research group at Lund University in Sweden now believe that nocturnal colour vision may be far more common than anyone imagined and could be found in toads, frogs, bees, wasps, fireflies and creatures of the deepest oceans. What鈥檚 more, finding out how they鈥檙e doing it is helping the Lund group design technologies to make our night life as colourful as theirs.
Advertisement
Having colour vision at night makes a lot of sense. Mates, food and shelter are all easier to find if you can see them in colour, and danger is easier to avoid. To see in colour at all, let alone at night, an animal鈥檚 eye must have at least two different types of photoreceptor, each sensitive to different wavelengths of light. The eyes of vertebrates like us have two types of photoreceptors, found on the retina, called rod and cone cells, but only the cones are important in colour vision. Humans, in common with many primates, are trichromats, which means we have three types of cone capable of detecting long, mid and short wavelengths of light. We perceive these as red, green and blue respectively. By combining and comparing signals from the cones, our brains interpret different combinations of these wavelengths, which allows us to distinguish all the colours of the rainbow.
鈥淢ates, food and shelter are all easier to find if you have colour vision at night鈥�
Many other mammals, such as horses, are dichromatic, with just two types of cones in their eyes. They see the world in combinations of two colours, much like a person with red-green colour blindness.
The other type of photoreceptor, rods, can work at lower light intensities than cones, so they come into play when night falls and fewer photons of light are available. 鈥淚t turns out you can discriminate more shades of grey [in low light] if you only use one type of receptor,鈥� Kelber explains. 鈥淭he signal-to-noise ratio is worse in a colour channel, and rods have been optimised to filter out noise, whereas cones have not.鈥�
The downside is that our rods respond to the different wavelengths of visible light in the same way and so cannot distinguish colours. This is why we see only in black and white after dark.
If only rods are active in low light, how do nocturnal geckos see in colour at night? The answer goes back to an early stage in the gecko evolutionary history, when no lizard was active after sunset. 鈥淟izards are very diurnal animals. They have got four different types of cone in their retina,鈥� says Kelber. 鈥淭hey can see all the different types of colour that we can see, plus ultraviolet.鈥� Over millions of years, the diurnal lizards used their rods so little that they simply evolved away, leaving the animals with only cone cells.
This was fine while they were purely diurnal, but at some point in the past one type of lizard, the geckos, became active at night. This left them in a bit of a mess, since they couldn鈥檛 see in the dark without rods. Then, says Kelber, two important things happened. First they got rid of their red-sensitive cones, which are the first to fail in low light, leaving them with cones sensitive to blue, green and ultraviolet. Then the outer segment of these remaining cones, the part that absorbs the light, altered to become longer and highly sensitive 鈥� more rod-like.
The cones in diurnal geckos are just five micrometres long, while those in nocturnal geckos measure 10 times that. Scientists studying nocturnal gecko vision in the 1970s thought the elongated cone cells were actually rods and that the geckos had black-and-white night vision. The biochemistry of these photoreceptors, however, is that of a cone, and recent studies by Victor Govardovsky of the I. M. Sechenov Institute of Evolutionary Physiology and Biochemistry in St Petersburg, Russia, and Kristian Donner of the University of Helsinki in Finland have confirmed that the gecko photoreceptors contain a particular type of opsin 鈥� a light-sensitive protein 鈥� that is only found in cones.
This, however, was only half the battle. Just because an animal has the physiological equipment to see in colour doesn鈥檛 necessarily mean it uses it. When an animal sees in colour, two separate neural pathways are at work. First, the receptor signals for two cone types are combined, which provides the brain with a signal about the brightness of whatever it is looking at but not about the object鈥檚 colour. For information about colour, as well as hue and saturation, the cone signals have to be compared. 鈥淎n animal has to compare or subtract the signals from two or more photoreceptors to see in colour,鈥� says Kelber. 鈥淭his comparison is an important step, and it is very hard to show with physiological measures.鈥�
鈥淣ewton pointed out in the 18th century that seeing in colour is something that happens in the brain and is not a property of an object or light source per se,鈥� explains Julian Partridge of the ecology of vision research group at the University of Bristol, UK. 鈥淭o demonstrate colour vision requires particular types of colour-mixing experiments.鈥�
Knowing that the geckos had the hardware to see in colour at night was one thing, but Kelber needed a behavioural experiment to prove that they were using it. Borrowing an experimental technique used in the 1930s in studies on diurnal lizards, she presented her subjects with crickets, some of which had been dipped in salt water and then dried. Crickets are a favourite food of geckos, but because they live in the desert, where water is scarce, they avoid salt to prevent dehydration.
Kelber and her colleague Lina Roth offered both types of cricket to the nocturnal geckos, always holding the salty ones in a pair of tweezers with a blue pattern on them and the tasty ones in grey-patterned tweezers. Once the crickets had learned which tweezers to go for, Kelber and Roth repeated the experiments in various low light levels, including starlight and dim moonlight, which was too dim for the researchers themselves to tell the colours apart. The geckos, however, unfailingly avoided the blue-tweezered, salty crickets.
After sunset
This was the first time that a vertebrate had been shown to have nocturnal colour vision, but Kelber thinks the phenomenon is much more widespread. 鈥淚 see no reason why all gecko species should not see [colour] in the dark,鈥� she says.
There may, however, be a good reason why colour night vision is not universal, and why evolution has left humans in the dark. Seeing at night is all about how many photons 鈥� units of light 鈥� your eyes can capture. A moonless night is about 100 million times darker than a bright sunny day, but unless you live in a cave or deep in the sea there is always some light available. Nocturnal animals generally make the most of the little light that is available to them by having large pupils to let in as many photons as possible 鈥� but as any photographer knows, large apertures result in shallow depth of focus, so the gain in sensitivity is counterbalanced by loss in detail.
鈥淓volution has left us humans in the dark鈥�
Seeing colour at night presents eyes with even more of a challenge. Colour vision requires the output of more than one cone cell, which means that in low light the eye has to share scarce photons between different cones. Each photon can only be absorbed by one receptor, so sharing them out like this means the eye detects even less detail. 鈥淭hat鈥檚 why colour vision in these dim light intensities isn鈥檛 all that common,鈥� says Kelber. It explains, too, why humans can鈥檛 see in colour at night 鈥� for us, the details are more important than the colours.
For insects, however, colours are often more important. Daytime flower visitors such as honeybees and butterflies have long been known to use colour to find, recognise and select the flowers that contain the sweetest nectar, and it now appears that some of their nocturnal counterparts, such as hawkmoths 鈥� known as sphinxes in the US 鈥� do the same.
Unlike geckos, which have lens eyes with large pupils to maximise the amount of photons able to reach their retina, invertebrates such as honeybees, butterflies and moths have compound eyes. This type of eye is multifaceted and made up of hundreds of units called ommatidia containing photoreceptor cells. At the centre of the eye, light-sensing tubes called rhabdoms absorb the incoming photons and perform the function of both rods and cones in vertebrates.
Hawkmoths prefer to feed on the nectar produced by blue or yellow flowers. The trouble is that after sunset the colour of natural light changes depending on how far the sun is below the horizon, whether or not the moon is out, whether the night is cloudy or starry and, increasingly, whether humans have built a light-polluting town nearby.
Discriminating between the colours of flowers in such changing conditions, especially against surrounding green foliage, is impossible with eyes that only use a brightness scale 鈥� a blue flower in very late twilight will look brighter than its green background, but in starlight it will look darker. Kelber figured that if the moths can pick the right colours at night, they must be using colour vision rather than brightness.
As with the geckos, Kelber knew that hawkmoths were physiologically capable of detecting colour but she needed to be sure they were using that ability rather than scent or some other sense to find their preferred blue and yellow flowers. She trained hawkmoths in twilight using five coloured discs, each with a small hole through which the moth could dip its proboscis. Four of the discs were different shades of grey and one was blue. The moths quickly learned that there was a sucrose reservoir behind the blue disc. Kelber then removed the reward, so there was no scent, and performed the experiment in different light levels. Each time, the moths selected the blue disc in anticipation of a reward.
So far, the geckos and hawkmoths are the only two nocturnal species that have been shown in behavioural experiments to see in colour. Kelber hopes to study nocturnal bees next: their compound eyes are optically less sensitive than hawkmoths鈥�, so having colour vision would be an even greater challenge for them. She also hopes to establish nocturnal colour vision in toads. Toads have two different kinds of rod, which some believe might be used for night colour vision, though it has never been consistently shown.
Deep colours
Perhaps the darkest habitat on Earth is the deep ocean. Even here, though, there may be creatures that see in colour. No behavioural experiments have yet been performed, but Partridge says that several deep-sea creatures have the cellular machinery necessary for colour vision, and that there are several possible reasons why colour vision might be useful to them. 鈥淭he spectrum in open seas varies with depth, especially in the first 500 metres or so, and an animal with colour vision might be able to tell its depth from the colour of the light,鈥� he says. 鈥淭here are also spectral clues associated with time of day, so perhaps colour vision could be used to entrain circadian clocks and control vertical migrations 鈥� the trips that some species make to feed near the surface each night.鈥�
In addition, many deep-sea species produce bioluminescence, which they use variously to communicate, attract mates, startle predators or camouflage their silhouettes against downwelling light. Partridge says that for perfect camouflage as seen from below these creatures need to match both brightness and the spectrum of the light 鈥� a match that is rarely perfect. Having colour vision would be a great way for a predator to break their prey鈥檚 camouflage.
Interestingly, Lydia M盲thger at the Marine Biological Laboratory in Woods Hole, Massachusetts, says that, except for some deep-sea species, most squid, octopuses and other cephalopods cannot see in colour, although their eyes are tremendously well developed. This is true even in species that camouflage themselves by changing their colour. 鈥淗ow colour change is accomplished in a colour-blind animal remains to be shown,鈥� she says. 鈥淥ne idea is that these animals may match their backgrounds by 鈥榠ntensity matching鈥�, for which wavelength information is not really necessary. We鈥檙e still trying to figure it out.鈥�
Another open question is about the sensitivity of cones in vertebrates other than ourselves. 鈥淲e know exactly when we humans lose colour vision when it gets dark,鈥� Kelber says. 鈥淲hen half a moon is up, or less, we do not see colour; with more than half a moon, we can see some faint colour. But this might well be different to what, for instance, a horse can do. It鈥檚 possible that horses can see colour under light conditions when we can鈥檛, and nobody has ever checked that.鈥�
Meanwhile, the knowledge that colour vision at night is theoretically possible is causing a great deal of excitement among researchers interested in the design of new and better gadgets. Eric Warrant, who works with Kelber at the University of Lund, is combining information about how nocturnal animals see in colour with mathematics to create algorithms that could allow cameras, microscopes and night-vision goggles to see colour in the dark. 鈥淭he hard part is to get the technology to do the job well in black and white. As soon as that鈥檚 solved, we can make it do the same thing over three colour channels, red, green and blue, and then combine them,鈥� he says. In the meantime, we鈥檒l just have to keep watching the experts in the animal kingdom for more clues on how to do it.
