Doppler-shift measurements with a remarkable precision of Δλ/λ=3×10⁻⁹, corresponding to velocities of 1 m s⁻¹, have been made repeatedly of 2500 stars located within 300 light years. The observed gravitational perturbations of the stars have revealed 250 orbiting planets, with 27 that cross in front of the host star, blocking a fraction of the starlight to allow measurement of the planet's mass, radius and density. Two new discoveries are the first good analog of Jupiter (HD 154345b) and the first system of five planets (55 Cancri). The predominantly eccentric orbits of exoplanets probably result from planet–planet gravitational interactions or angular momentum exchange by mean-motion resonances. The planet mass distribution ranges from ~15 MJUP to as low as ~5 MEarth and rises toward lower masses as dN/dM∞M⁻¹·¹. The distribution with orbital distance, a, rises (in logarithmic intervals) as dN/d log a∞a⁺⁰·⁴. Extrapolation and integration suggests that 19% of all Sun-like stars harbor a gas-giant planet within 20 AU, but there remains considerable incompleteness for large orbits. Beyond 20 AU, the occurrence of gas-giant planets may be less than a few per cent as protoplanetary disk material there has lower densities and is vulnerable to destruction. Jupiter-mass planets occur more commonly around more massive stars than low mass stars. The transit of the Neptune-mass planet, Gliese 436b, yields a density of 1.55 g cm⁻³ suggesting that its interior has an iron–silicate core surrounded by an envelope of water����ice and an outer H–He shell. Planets with masses as low as five Earth-masses may be commonly composed of iron–nickel, rock and water along with significant amounts of H and He, making the term 'super-Earth' misleading. The transiting planet HD147506b has high orbital eccentricity but no significant orbital inclination to the line of sight, presenting a puzzle about its history. Its orbit together with the mean motion resonances of 4 of the 22 multi-planet systems provides further evidence for the role of planet–planet interactions in shaping planetary architectures.