Struggling to hear a conversation at a cocktail party, the last thing we humans want is the background racket to get louder. North American paddlefish (Polyodon spathula), would never be such party poopers. For, as new work in Nature1 shows, they are all for 'noisier' background hubbub – it helps them hunt fish.

Paddlefish graze on microscopic animals or 'zooplankton' called 'Daphnia'. As vision is not much use in the muddy waters of the rivers that they call home, paddlefish have evolved another way of finding their prey. They have a long antennae – like a radio aerial – in front of their mouths that detects low-frequency oscillating electric fields. These sensitive organs can pick up the tiny electric fields produced by other organisms. With its antennae a hungry paddlefish can sense Daphnia from a distance of four centimetres.

Common sense would suggest that the ability of these fish to detect dinner should decrease if, in addition to the electrical signals that the prey produce, the water is pervaded by a background electrical noise, akin to the hiss of radio 'static'. The noise would, you'd expect, obscure the signal.

But David Russell and colleagues from the University of Missouri at St Louis have been investigating a counter-intuitive effect of background noise called 'stochastic resonance', in which the noise actually helps an information-laden signal to be transmitted more clearly. In systems that show stochastic resonance, there is an optimal level of noise for signal transmission. Increasing or decreasing the noise level results in poorer transmission.

Stochastic resonance was first detected in geological records of past climate change, where random fluctuations in climate have contributed to the 100,000-year recurrence time of the ice ages. It was later demonstrated in electronic circuits, and led to speculations that the neural wiring of sensory systems in living organisms might have evolved to take advantage of stochastic resonance. Given that most natural environments are full of random fluctuations – or 'noise' – of all sorts, it seemed likely that natural selection would have guided organisms towards using the noise rather than fighting it.

Russell's colleagues Frank Moss and Lon Wilkens were part of a team that, in 1993, demonstrated this to be so (see Nature 365, 337; 1993). They showed that hair cells in crayfish, which sense directional flow in surrounding water, are more efficient detectors of motion when the water is disturbed by a low level of turbulent 'noise' than when there is none.

But this does not prove that crayfish actually make use of stochastic resonance, rather than just experiencing it. So Russell's group set out to see whether paddlefish, whose detectors sense electrical rather than mechanical signals, are better able to locate plankton during feeding with or without background noise.

They let the fish feed in a flowing stream of water pervaded by a random electric field. As the researchers stepped up the amplitude of the electrical noise (akin to turning up the volume of a untuned radio), the fish were able to detect Daphnia plankton from further afield until an optimal noise level was reached. Beyond this, the fish's feeding range declined as the noise truly started to interfere with signal detection.

Of course, fish aren't able to generate their own noisy electric fields to aid their feeding sprees. Conveniently, their hapless prey do it for them. While each individual plankton sends out an electrical signal that betrays its presence, a whole population produces a general buzz of background noise, of which paddlefish seem to have evolved to make the best use.