Second author

The burst of radio waves emitted by a pulsar — a rapidly rotating neutron star — is like the beam of a lighthouse, sweeping briefly over anyone standing in its path. Andrew Lyne of the University of Manchester and a team of fellow astronomers surveyed the sky for these regular, or periodic, pulses of radio waves and found more than 750. But when they stepped back, they noticed that, in some data, there were transient bursts that repeated but were not regular like those from normal pulsars. Lyne and his team decided to focus on this phenomenon and look for more cases of it; they report their results on page 817. Lyne spoke with Nature about the search for the objects he and his colleagues now call RRATs, for rotating radio transients.

What made you focus on this rare type of event?

We didn't know what we would find. We thought it would be a good idea to open our eyes wider and see what other astronomical beasts we could spot.

How did you go about the search?

Rather than look at periodicity, we looked for bright flashes that had the characteristic signature of passage through interstellar space. We went through 150 days of telescope data (30 million samples) and saw such flashes in 11 of the 3,000 regions we studied. But however hard we looked, we couldn't see any periodicity. When we went back to them with our telescope we saw more flashes. There were things in the sky that were giving these very rare bursts; lasting 10 milliseconds or less, at intervals of between 4 minutes and 4 hours. Because they are so ephemeral, it took an enormous amount of time to establish them and make measurements.

How did you work out that these sources were different from ‘normal’ pulsars?

It took us three years or so to sort out what we had got. We measured the occurrence times of these flashes and found that the intervals between them were always a multiple of a well-defined, much smaller time, like a pulsar — but that 999 out of a thousand of the pulses were missing.

What's the significance of these gaps?

We don't know why these RRATs radiate only intermittently. There is evidence that they are related to magnetars, neutron stars with huge magnetic fields that emit only X-rays.

Any methodological lessons learned?

We really must keep our eyes open and not rely too much on instruments and software, which only find the sort of things you design them to.