Sinking through the inky ocean, it would seem that there is little light at depth: but you'd be wrong. 'In the mesopelagic realm [200 m] bioluminescence [light produced by animals] is very common', says Sonke Johnsen from Duke University, USA, explaining that many creatures are capable of producing light, yet rarely do so. But how much light do the inhabitants of the ocean floor (benthos) generate?

Explaining that some bioluminescence is generated when organisms collide, Johnsen says, 'In the benthos you have a current moving over complicated ground with all the things in the water banging into it, so one idea was that there would be a fair amount of bioluminescence.'

However, few people have visited this remote and inhospitable habitat. Intrigued by the animals that dwell there and the possibility that bioluminescent bacteria coating the ocean floor might glow faintly, long time collaborators Tamara Frank, Sonke Johnsen, Steven Haddock, Edith Widder and Charles Messing teamed up to find out just how much light is produced by seabed residents.

The team discovered that bioluminescent animals are relatively rare but blue-green bioluminescence produced when plankton collide with obstacles is relatively common. They also found that deep-sea predators have incredibly sensitive colour vision and they publish their discoveries in a pair of papers in the Journal of Experimental Biology at http:/jeb.biologists.org.

Descending to the bottom of the ocean near the Bahamas in Harbor Branch Oceanographic's Johnson- Sea-Link submersible, switching off all the lights and adapting to the darkness, the group were amazed to find themselves continually surrounded by tiny flashes of light as bioluminescent plankton collided with coral and boulders strewn across the floor.

However, there was no evidence of the all-pervasive glow produced by bioluminescent bacteria that the team had hoped to find. 'We weren't in regions where the currents were slow enough to allow for collection of detritus,' says Frank, adding, 'it's not that this phenomenon doesn't exist...we just weren't able to observe it on these dives.'

Next the submariners began searching for bioluminescent inhabitants, gently tapping coral, crabs and anything else they could reach with the submersible's robotic arm to see whether any of the organisms emitted light. The team found that only 20% of the species that they encountered produced bioluminescence (Johnsen et al. 2012).

Collecting specimens and returning to the surface, Johnsen and Haddock then photographed the animals' dim bluish glows - ranging from glowing corals and shrimp that literally vomit light (spewing out the chemicals that generate light where they mix in the surrounding currents) to the first bioluminescent anemone that has been discovered - and carefully measured their spectra.

The duo found that most of the species produced blue and blue-green spectra, peaking at wavelengths ranging from 455 to 495nm. However, a family of soft corals known as the pennatulaceans produced green light, with spectra peaking from 505 to 535nm.

'We were working at the absolute limits of what the equipment can do', remembers Johnsen, recalling the frustration of working in the cramped, pitch-dark conditions on the boat. 'It gives you respect for our vision, we can see the bioluminescence fine, but getting it recorded on an instrument or a camera is much harder', he adds. And as if that wasn't challenging enough, proving that anything living down there could even see the spectacular light display was even trickier.

Devising a strategy for collecting crustaceans ranging from crabs to isopods under dim red light - to protect their sensitive vision - by luring or gently sucking them into light-tight boxes, the submersible's crew then sealed the animals in boxes to protect their vision from harsh daylight when they reached the surface.

Back on the RV Seward Johnson, Frank painstakingly measured the weak electrical signals produced by the animals' eyes in response to dim flashes of light ranging from 370nm to over 600nm and found that the majority of the creatures were most sensitive to blue/green wavelengths, ranging from 470nm to 497nm (Frank at al. 2012). Most surprisingly, two of the animals were capable of detecting UV wavelengths.

Even though there is no UV left from the sun at this depth, Johnsen explains, 'Colour vision works by having two channels with different spectral sensitivities, and our best ability to discriminate colours is when you have light of wavelengths between the peak sensitivities of the two pigments.'

He suspects that combining the inputs from the blue and UV photoreceptors allows the crustaceans to pick out fine gradations in the blue-green spectrum that are beyond our perception, suggesting, 'These animals might be colour-coding their food': they may discard unpleasant-tasting green bioluminescent coral in favour of nutritious blue bioluminescent plankton.

Finally, after recording the crustacean's spectral sensitivity, Frank - from Nova Southeastern University, USA - measured how much light the animals' eyes had to collect before sending a signal to the brain (the flicker rate). She explains that there is a trade-off between the length of time that the eye collects light and the ability to track moving prey. Eyes that are sensitive to dim conditions lower the flicker rate to gather light for longer before sending the signal to the brain.

However, objects moving faster than the flicker rate become blurred and their direction of motion may not be clear. The crustaceans' flicker rates ranged from 10 to 24Hz (human vision, which is sensitive to bright light, has a flicker rate of 60Hz) and the team were amazed to find that one crustacean, the isopod Booralana tricarinata, had the slowest flicker rate ever recorded: 4Hz. According to Frank, the isopod would have problems tracking even the slowest-moving prey.

She suggests that as it is a scavenger, it is possible that it may be searching for pockets of glowing bacteria on rotting food and it might achieve the sensitivity required to see this dim bioluminescence with extremely slow vision.

Having shown that bioluminescent benthic species are scarce but the phenomenon itself is not, Johnsen is keen to return to the ocean floor to discover more about the exotic creatures that reside there. 'We would love to go back, get more basic data. We've only scratched the surface', he says, adding, 'When you are down there you are cramped and cold and stiff, but at the end of a dive I never want to come back up.'

Johnsen, S., Frank, T. M., Haddock, S. H. D., Widder, E. A. and Messing, C. G. (2012). Light and vision in the deep-sea benthos: I. Bioluminescence at 500m depth in the Bahamian Islands. J. Exp. Biol. 215, 3335-3343; Frank, T. M., Johnsen, S. and Cronin, T. W. (2012). Light and vision in the deep-sea benthos: II. Vision in deep-sea crustaceans. J. Exp. Biol. 215, 3344-3353.