We must follow-up the initial epoch of giant planet discoveries with an extensive dynamical, photometric, transit, and imaging exploration of main-sequence stars to determine the orbital characteristics and gross physical properties of their planets. A multi-pronged strategy of dynamical, photometric- transit, and imaging techniques should be pursued in series and in parallel. These should be implemented in three chronological phases.

In the first (reconnaissance) phase, astronomers must make a complete inventory of giant and Neptune-mass planets around all stars within 10 parsecs and around a statistically significant sample of more distant stars. Such a census, carried out with ground-based radial-velocity and astrometric techniques, will determine the abundance of planets and the correlation of stellar properties (such as mass, metallicity, and binarity) with giant planet properties (such as mass and orbital parameters). Importantly, giant planets dynamically constrain the orbits left available to terrestrial planets, influencing later searches for Earth-like worlds. In this sense, the study of giant planets is an important stepping stone to the more demanding study of the smaller terrestrial planets.

The above Doppler and astrometric surveys are challenging, requiring velocity precision of 1 m/s and astrometric precision of 20 microarcseconds (for example, the Keck Interferometer). Nonetheless, these efforts are relatively inexpensive and the technology is already relatively mature. Note that the planets detected in this first reconnaissance phase have intrinsic brightnesses of a millionth to a billionth that of the host star and many will be separated by an arcsecond or less.

A second phase employs the space-telescopes Kepler-- a new Discovery-class mission designed to photometrically search for terrestrial and giant planet transits around tens of thousands of nearby stars--and the Space Interferometry Mission (SIM), an interferometer with an astrometric precision for terrestrial and giant planet detection of 1-10 microarcseconds. Kepler will have a photometric precision of one part in 100,000 and should discover hundreds of terrestrial and giant planets, while SIM will discover and astrometrically measure planet masses down to a few Earth-masses. SIM will survey the youngest stars close to the Sun to study the formation and evolution of Jupiter-size planets. To obtain a secure mass for a terrestrial planet requires a dynamical technique such as only SIM will employ. The complementarity between the photometric- transit technique of Kepler and the astrometricinterferometric technique of SIM provides NASA with a powerful program for pioneering terrestrial planet discovery and preliminary terrestrial and giant planet characterization.

The first and second phases of dynamical and transit surveys must be followed by a third phase of direct spacebased detection of the reflection and/or intrinsic light of the planets themselves. For giant planets, the logical technological and scientific precursor to a Terrestrial Planet Finder (TPF) and the more difficult problem of direct terrestrial planet imaging and spectroscopy is a spacebased "giant planet finder." Using high-contrast imaging and low-resolution spectroscopy, such a mission would be capable of both discovery and analysis of the dynamically dominant and brighter components of planetary systems, while the later TPF will be able to observe at even larger star-planet flux contrasts the spectral features of the water, carbon dioxide, methane, and ammonia thought to reside in the atmospheres of the terrestrial planets. The technology, management structure, and discoveries of a giant planet finder program would provide NASA with valuable experience and guidance as it embarks upon the more challenging TPF initiative.

Though the direct photometric and spectroscopic detection of extrasolar giant planets will be a milestone in planetary research, the discovery and study of Earth-like planets that would be enabled by TPF is the ultimate goal of this first era of extrasolar planetary exploration.

Radial-velocity programs are unlikely to detect extrasolar planets with masses below a Uranus mass. Astrometric searches with an accuracy of 10 microarcseconds (KI) to 1 microarcsecond (SIM) can push the limit down to a few times the Earth's mass and survey a volume out to 5-10 parsecs. A space-based photometric-transit survey such as Kepler will extend to much larger volumes of space and provide an initial estimate of the frequency of terrestrial planets. However, direct imaging and spectroscopy of Earth-like planets will require TPF, an infrared interferometer or an optical coronagraph that can suppress the light of the central star to unprecedented levels, to reveal for the first time the atmospheres of planets like our own outside the solar system.