"Our holy grail in exoplanet science is to find an Earth twin," says Sara Seager of MIT.

An Earth twin would have three crucial characteristics: it would be a rocky planet the same size as Earth; it would orbit a sun-like star; and it would be located in its star's habitable zone, at the same distance from its star that the Earth is from the sun.

Such a planet would be a prime candidate for follow-up studies to determine whether life had gained a foothold there.

But finding Earths is hard. And figuring out whether or not they harbor life is even harder.

NASA's Kepler mission, which was launched last week, is designed to monitor 100,000 stars, searching for Earth-like planets in Earth-like orbits. If Earths are common, Kepler could find as many as 50 such planets around distant stars.

But Kepler will observe dim stars. So even if it finds Earth-like planets, it will be hard to do follow-up studies to determine whether or not they are inhabited. Our best shot at finding a living world outside our solar system, Seager says, is to study planets around bright stars.

"Brighter is always better," Seager says.

"More photons for us makes it easier to measure things." If an Earth twin is found orbiting a bright star, the extra light will make it easier to study the planet's atmosphere, and its atmosphere holds important clues about whether or not the planet harbors life. The oxygen in Earth's atmosphere, for example, is there because photosynthetic bacteria and plants produce it.

But because Kepler will look at thousands of stars simultaneously, "they can't put any bright stars in there...it would wreck a lot of the picture," Seager says.

It's "hard to measure really bright things and really faint things at the same time, with the same exposure." And unfortunately, there's no patch of sky where thousands of bright stars appear together. To see transits around bright stars, the stars have to be observed one at a time.

So Seager wants to build a tiny telescope, send it into orbit and have it monitor one star, just one bright star, hoping to find another Earth. That's in the proof-of-concept phase. If it works, she proposes to build a fleet of Earth-hunting telescopes, hundreds of them, and send the entire fleet into orbit, each one dedicated to watching a different bright star.

More than 300 planets have been discovered orbiting distant stars. Most of them have been found using the "radial-velocity" technique, which measures tiny shifts in the color of a star's light caused by a planet's back-and-forth gravitational tug on the star.

Radial-velocity planet searches have managed to find planets as small as Neptune, but they will be hard-pressed to find Earth-sized worlds, particularly if the planets orbit as far out from their stars as Earth does. Earth, literally, doesn't have enough pull.

So instead, Seager's ExoplanetSat telescope, like Kepler, would search for planets by monitoring stars for planetary "transits.

A transit occurs when a planet moves across the face of a star. A solar eclipse is a type of transit, during which the moon moves between the Earth and the sun, blotting out the sun's light. A transiting planet produces a much smaller effect, but the basic idea is the same: while the transit is in progress, the sun's light gets dimmer.

Transits of Mercury and Venus can be viewed from Earth on a predictable schedule. The last transit of Venus was in 2004; the next one will occur in 2012. A transit of Mercury or Venus appears as a small dark circle moving across the face of the sun.

But for distant stars, a transit won't look like a little black circle. Its only trace will be a slight dimming of the star's light. Really slight. About one-hundredth of one percent.

Imagine you're sitting in a dark room, no lights, the only illumination is an iPhone, its screen pointed away from you, all 154,000 of its pixels cranked up to maximum brightness. Now imagine that 13 pixels suddenly go dead. Just 13. Can you tell that the room got darker? Probably not. That's the brightness difference a telescope will need to detect to catch a transit of an Earth twin.

At this point, Seager's ExoplanetSat project is still only an idea. She has received an ASTID (Astrobiology Science and Technology for Instrument Development) grant from NASA to develop the concept.

A critical technical challenge for this design phase will be to devise a control system that can aim the telescope's CCD with pinpoint accuracy. The CCD, similar to the imaging device in a digital camera, can make useful measurements only if the image of the target star remains continuously on exactly the same pixels.

"Pixels behave differently from each other," Seager says.

"They respond very slightly differently to light." And that difference, if the image shifts, can be enough to throw off the experiment's results. A variety of forces - Earth's gravity, its magnetic field, its atmospheric drag - will conspire to push the telescope out of alignment. It will be the control system's job to push back, tilting its CCD, mere microns at a time, to keep it from drifting off-target.

If the technical hurdles can be overcome, and additional funds can be found, Seager hopes that the first telescope can be launched into orbit as early as 2012, piggybacked on the launch vehicle of another satellite.

To hitch that ride, the instrument will have to fit into an unusually small container: a rectangular cuboid box only 10 cm (4 inches) square and 30 cm (12 inches) long. That's about the size of the Manhattan phone book, cut in half vertically.

We're used to important astronomical discoveries being made by giant telescopes: the Hubble, or the Keck atop Mauna Kea.

But if Seager's work pans out, a telescope not much bigger than the one Galileo first used to explore the night sky could blaze a trail that leads to one of the most profound scientific discoveries of all time: life on another world.