Douglas Adams, the founder of this site, famously dreamt up The Hitchhiker's Guide to the Galaxy while lying drunk in a field in Innsbruck, staring at the stars. He wasn't the first to have wondered whether the galaxy might be teeming with life, although he arguably was the first to have imagined it with such comic results.

Adams' ideas — the serious ones, anyway — are predated by at least two thousand years of speculation. The Greek philosophers Democritus, Epicurus and Leucippus also believed that the world was composed of atoms, but they had not a hope of proving either proposition.

Finding something huge is generally much easier than finding something tiny and invisible¹ but only if it's close at hand. Planet-sized objects can be even more difficult to observe than atoms if there are trillions of miles between you and them. This is the reason why scientists knew electrons existed within atoms many years before they knew that planets existed around other stars. Now that we have started to see these extrasolar planets, they are providing the first glimpse of a universe that is more weird than anyone could have imagined — even to those people blessed with remarkable creativity while lying intoxicated in a field in Innsbruck.

Twinkle, Twinkle...

The ease with which a planet is detected depends entirely upon the technology used. The planets that can be observed with the naked eye — Mercury, Venus, Mars, Jupiter and Saturn — have been known since ancient times, but it was the invention of the telescope that caused our conception of the Solar System to expand rapidly. The Astronomer Royal, William Herschel, observed Uranus in 1781, originally naming it, in a fit of sycophancy, Georgium Sidus (George's Star) after the then King. Neptune's existence was inferred indirectly in 1846 by its effect upon the orbit of Uranus, and then observed by James Challis in the same year, who found it exactly where expected. In contrast Pluto was observed directly by Clyde Tombaugh in 1930. Quaoar and Varuna are two recent additions to the Solar System, each about the size of Pluto's moon, Chiron. Sedna, almost the size of Pluto, orbits deep within the Kuiper Belt at three times the other planet's distance from the Sun. Yet another planet may exist, a huge gas giant, up to six times the mass of Jupiter and half a light year distant.

It's been difficult enough locating new planets orbiting our own Sun, even when equipped with the mightiest telescopes ever built. Planet-hunters looking at other stars have to rely upon indirect methods of inference. Detecting a planet orbiting a star many light years distant is therefore a demanding and difficult process that requires patience, guile, rigorous attention to detail and very effective technology.

Astrometry

The easiest way of detecting a distant planet is to look for any effects upon its parent star. As forces always exist in opposite pairs, the gravitational force between two objects is felt equally by both of them. A planet being whirled around a star by its gravity also causes the star to orbit a point inside its own volume, rather like an Olympic hammer thrower gyrating in his circle as he winds the hammer up for the big throw. The bigger the planet, the more pronounced the effect. The closer the planet, the shorter its year, and hence the faster the star 'wobbles' in the sky. Large planets orbiting close-in cause larger and more frequent wobbles than smaller planets orbiting further out.

Astrometry is essentially a process of identifying this wobble through 'successive subtraction'. All stars appear to move across the sky to some degree, but most of the movement is due to the effect of parallax as the earth orbits the sun. Find some way of factoring out the parallax, and you are left with the star's proper motion: the actual movement at a direction perpendicular to the line of vision. Part of the proper motion is due to the star's linear motion through the galaxy. Remove that, and what you see is the wobble caused by planets orbiting the star. Astrometry seeks to measure the absolute wobble of nearby stars against the cosmic background.

Measuring small differences in position against the backdrop of the night sky is fraught with hazards. One astronomer who came a cropper was Peter van der Kamp, who analysed wobbles in the motion of the famous red dwarf star Proxima Ophiuchi aka Barnard's Star². He claimed that a huge planet was in a 24-year orbit around the star. Subsequent discoveries of planets by van der Kamp showed that they too had a 24-year orbital period. The lens of the refracting telescope used to measure the wobble had, 24 years previous to the measurements, been replaced, and subtle changes in the refracting property displaced the image of the star relative to its background.

Van der Kamp might have been wrong about his extrasolar planets, but he was not wrong in believing that the Universe had many objects of mass somewhere in between Jupiter and that of red dwarf stars. These brown dwarfs³, as they are known, have now been observed and might actually be more numerous than stars themselves.

Astrometry has yielded a few tantalising results, but nothing as dramatic as the next method of detection. Prospects of finding planets successfully using astrometry will improve when we can get astrometric telescopes beyond the Earth's atmosphere.

Radial Velocity Measurements

These make for a safer method of finding wobbles in a star's position. The principle works like this: imagine you have a whistle stuck at the end of a long rubber tube. You put the tube end in your mouth, whirl the tube around in a circle and blow hard. Observers standing to the side of you will hear the pitch of the whistle rise and fall as it alternately approaches and recedes from them. This is the Doppler effect.

The Doppler effect works on any kind of wave motion, including light. If a star rushes away from you, the wavelength lengthens and the light turns redder. If it comes towards you, the light becomes bluer as the wavelength shortens. If it wobbles around in a circle, these colour changes take on a rhythm. Passing the star's light through a spectroscope and looking for such characteristic periodic changes betrays the presence of a planet far more reliably than astrometry. Dark absorption lines in the star's light, called Fraunhofer lines, either move to the red or blue end of the spectrum depending upon the direction in which the star is travelling. The position of these lines is compared with those of a known, Earth-based standard, normally iodine vapour.

The main flaw in this method is that the planets it is most likely to detect are least likely to exist in the first place. Firstly, the wobble would need to be big to be detected easily, and this would require a heavy planet, most likely a gas giant. Moreover, the wobble would be most easy to spot if the planet's year was small, which would mean that the planet was close to its star. Big planets can't form too close to their stars, their mass being mostly volatile gases would boil off rapidly the closer they came, rather like Icarus's wings melting as he flew too close to the Sun.

Hot Jupiters

Contrary to all expectations, the first planets that were found were of exactly the wrong kind in exactly the wrong place. Two astronomers found the 'Planet from Hell' around a nearby star, 51 Pegasi, using radial velocity measurements in 1995. This planet seemed preposterous at first: it orbited at 1/20^th of the distance between Earth and the Sun, its year was 4.2 days long, but it had at least half the mass of Jupiter. This would imply that its atmosphere was incredibly hot (3000 K). The only plausible explanation was that the planet had formed much further out and then migrated closer to the star. Once it came close enough to the star to raise tides on its surface, this would transfer rotational energy from the star to the planet, and counteract any further decay in the orbit.

Unlikely as this explanation seemed, it was soon borne out by other observations. 18% of all planets found so far are 'hot Jupiters' (or 'roasters' as they are sometimes called), simply because the best detection methods can find them quickly and easily. Astrometry, on the other hand, is more sensitive to long period planets, so it complements the radial velocity method. Astrometry is more likely to find Earth-like planets. Moreover, radial velocity measurements work best when the planet's orbit is edge-on, and not at all when the orbit is perpendicular to our line of sight, and hence give a lower limit to the mass of the detected planet⁴. Astrometry works whatever the angle of the orbit and yields more precise mass estimations.

Transits

This simply means 'moving in front of something else'. When a planet moves in front of its star, the light reaching Earth drops. If Jupiter moved in front of the Sun it would be visible as a 2% light drop to a distant observer. Unfortunately, the chance of observing a direct transit is very small: in contrast, the two previous methods don't require a window of opportunity.

The transit method has been most effective when it vindicates other kinds of measurements. In 2000, two separate teams observed a star, HD 209458, thought to have had a hot Jupiter in orbit. They saw a well-defined drop in light intensity which allowed them to deduce the diameter of the planet as well as its mass. Also, the star flays the atmosphere from the planet, and the escaping gas filters the starlight, allowing the astronomers to deduce that the planet is mostly hydrogen and is probably less dense than Saturn.

As part of its swansong, the Hubble Telescope has been busy using transits to detect up to another 100 new planets. During a week-long period it pointed at the galactic bulge and monitored thousands of stars, seeing transit events in many of them. These events are only seen in planets in which the orbits are edge-on to us, so the actual figure is likely to be very much higher.

The major advantage of the observation of transits over the other two methods is that it allows the diameter and orbital radius of the planet to be gauged. As the planet transits the star, the drop in light intensity gives the planet's size, whereas the length of time of that drop tells the astronomer how big the orbit is.

Direct Observation

It's almost impossible to image the disc of another star⁵, let alone see a planet orbiting it directly. The obstacles are formidable: the solar systems of interest are very far away, the stars very bright, the planets very dim. It's like trying to see a moth fluttering around an arc lamp miles away with a pair of binoculars. Observing the system in the infrared can help, as planets are comparatively brighter. However, it's still a hunt for a needle in a haystack, the feeble light from the planet being swamped by the brilliance of the star.

This hasn't stopped some people from trying. Wisely, they chose a very dim star (a brown dwarf) which was fairly close — only 230 light years away. They also used a very, very large telescope array which used advanced 'adaptive optics'. Despite the choice of a 'soft target' and a 'big gun', the result is still astounding: the tiny red blob next to its parent star is probably a gas giant planet. This may be the first planetary system beyond ours. The astronomers know it's a planet because it has water molecules in its atmosphere: stars chew up every chemical compound imaginable, including water.

Other Earths

So, after having compiled a bestiary of exotic planets, many the size of Jupiter and locked in a deadly embrace with their star, what are the chances of the planet-hunters identifying rather smaller, sedate rocks upon which life might actually get the chance to evolve? It's not as if these planets are likely to exist in minute numbers: current estimates border on there being 30 billion terrestrial planets in our Galaxy alone. The odds of finding such planets lengthen a lot when one considers that these planets would have longer years and cause much smaller wobbles in their star's position.

The odds shorten again the longer we look for these planets. Hot Jupiters tend to get found simply because the radial velocity method is most sensitive to their kind. It's only now that smaller planets are being found, although none of them are likely to harbour life. Adopt a different method of detection, and we might start to see terrestrial planets, instead of inferring their presence.

This is precisely what the Terrestrial Planet Finder (TPF) telescope is designed to do. This instrument comprises four space-based telescopes flying in formation. Their light is combined in such a way as to vastly increase resolution. One of the indicators that TPF will be looking for is the presence of elemental oxygen in a planet's atmosphere. All oxygen in our atmosphere is there because of photosynthetic organisms: plants and cyanobacteria. Oxygen is therefore a key signature of life⁶.

TPF is not due to fly until 2015 at the earliest. In the meantime, Earth-based telescopes will get bigger and better, and astronomers will be able to observe for longer wobbles than they currently can. Even if we can't see the little green men yet, we'll have a much better idea of where they might live.

Some Current Operational Facilities

• MOA: Microlensing Observations in Astrophysics is a Japan/New Zealand collaboration using the gravitational microlensing technique at the Mt John Observatory in New Zealand.
• OGLE: Optical Gravitational Lensing Experiment makes observations at the Las Campanas Observatory, Chile, using a second generation CCD 8kMOSAIC camera. OGLE regularly monitors 130 million stars in the galactic bulge of the Milky Way.
• RoboNet: Optimised robotic monitoring of galactic microlensing at a UK national facility, the two metre robotic telescope at the Telescope Management Centre at Liverpool JMU (John Moores University). Although mainly concerned with delivering school-age educational programmes, the technique is being utilised to assist in the search for rocky Earth-like extrasolar planets.
• Rocky Planet Search are searching for rocky worlds 1-20 times the size of Earth using the Automated Planet Finder Telescope at Lick Observatory on Mount Hamilton, North California, US. Fancy taking a virtual tour of the facilities?
• NASA's Spitzer Space Telescope has been searching the systems of stars similar to our own sun, as this is where some astronomers think the best chance of finding an Earth-like planet lies. Their observations have shown that between a fifth and a third of studied stars have the conditions which we think led to the Earth being formed. Alan Stern of NASA said: 'I expect that we will find a very large number of planets'.
• The Kepler mission search for extrasolar planets began in 2009. Planets are detected by slight dimming in a star's brightness as the planet transits the star. Over a thousand planetary candidates had been discovered by Kepler up to 2011.